1002 lines
426 KiB
Plaintext
1002 lines
426 KiB
Plaintext
wandb: Detected [openai] in use.
|
||
wandb: Use W&B Weave for improved LLM call tracing. Install Weave with `pip install weave` then add `import weave` to the top of your script.
|
||
wandb: For more information, check out the docs at: https://weave-docs.wandb.ai/
|
||
Checkpoint tracker file does not exist: /mnt/mole/SDPO/L2T/checkpoints/datasets/sciknoweval/material/qwen3gen-material-RLSD_TR-Qwen-Qwen3-8B-mbs8-decay0-tr0.1-train32-rollout8-lr1e-6-vllm0.8/latest_checkpointed_iteration.txt
|
||
Training from scratch
|
||
Training Progress: 9%|▉ | 9/100 [06:35<1:02:54, 41.48s/it]
|
||
step:1 - global_seqlen/min:18040 - global_seqlen/max:28894 - global_seqlen/minmax_diff:10854 - global_seqlen/balanced_min:23555 - global_seqlen/balanced_max:23558 - global_seqlen/mean:23556.875 - actor/entropy:0.13252580165863037 - perf/mfu/actor_infer:0 - training/rollout_probs_diff_valid:1 - training/rollout_probs_diff_max:0.380796879529953 - training/rollout_probs_diff_mean:0.0026935043279081583 - training/rollout_probs_diff_std:0.010040820576250553 - training/rollout_actor_probs_pearson_corr:0.9984611868858337 - self_distillation/success_group_fraction:0.6875 - self_distillation/success_sample_fraction:0.6796875 - self_distillation/correct_sample_fraction:0.53125 - self_distillation/feedback_available_fraction:0.0 - self_distillation/feedback_used_fraction:0.0 - self_distillation/reprompt_sample_fraction:0.6796875 - rollout_corr/rollout_is_mean:0.999983549118042 - rollout_corr/rollout_is_ratio_fraction_high:1.0524653589527588e-05 - rollout_corr/rollout_is_ratio_fraction_low:3.157396349706687e-05 - rollout_corr/rollout_is_max:2.075308322906494 - rollout_corr/rollout_is_min:0.2384062260389328 - rollout_corr/rollout_is_std:0.028271863237023354 - rollout_corr/rollout_is_eff_sample_size:0.9992012210621306 - rollout_corr/rollout_is_seq_mean:0.999904990196228 - rollout_corr/rollout_is_seq_std:0.0016567112179473042 - rollout_corr/rollout_is_seq_max:1.0049364566802979 - rollout_corr/rollout_is_seq_min:0.9954314827919006 - rollout_corr/rollout_is_seq_max_deviation:0.0049364566802978516 - rollout_corr/rollout_is_seq_fraction_high:0.0 - rollout_corr/rollout_is_seq_fraction_low:0.0 - rollout_corr/training_ppl:np.float64(1.1430769278667867) - rollout_corr/training_log_ppl:np.float64(0.13030084642377915) - rollout_corr/kl:np.float64(0.0005215430860481263) - rollout_corr/k3_kl:np.float64(0.0004265131673406586) - rollout_corr/rollout_ppl:np.float64(1.1424769102595747) - rollout_corr/rollout_log_ppl:np.float64(0.12977930237684632) - rollout_corr/log_ppl_diff:np.float64(0.0005215440469328314) - rollout_corr/log_ppl_abs_diff:np.float64(0.0013395921414485201) - rollout_corr/log_ppl_diff_max:np.float64(0.004913784563541412) - rollout_corr/log_ppl_diff_min:np.float64(-0.004234239459037781) - rollout_corr/ppl_ratio:np.float64(1.0005230973474681) - rollout_corr/chi2_token:np.float64(0.0006653782911598682) - rollout_corr/chi2_seq:np.float64(0.3484567936975509) - actor/pg_loss:np.float64(0.0012517671566456556) - actor/kl_loss:np.float64(0.0) - actor/pg_clipfrac:np.float64(0.0) - actor/ppo_kl:np.float64(0.0) - actor/pg_clipfrac_lower:np.float64(0.0) - self_distillation/token_reweight_lambda:np.float64(0.5) - self_distillation/token_reweight_eps_w:np.float64(0.20000000000000004) - self_distillation/token_reweight_active_token_fraction:np.float64(0.6796875) - self_distillation/token_reweight_context_sample_fraction:np.float64(0.6796875) - self_distillation/token_reweight_correct_sample_fraction:np.float64(0.53125) - self_distillation/token_reweight_w_mean:np.float64(70357.71696712193) - self_distillation/token_reweight_w_std:np.float64(1319602.365722549) - self_distillation/token_reweight_w_clipped_mean:np.float64(1.0009874245151877) - self_distillation/token_reweight_w_clip_frac:np.float64(0.039357453788397834) - self_distillation/empty_target_batch:np.float64(0.3203125) - actor/grad_norm:np.float64(0.20237117260694504) - perf/mfu/actor:np.float64(0.039861969436277) - perf/max_memory_allocated_gb:np.float64(130.98325157165527) - perf/max_memory_reserved_gb:np.float64(140.861328125) - perf/cpu_memory_used_gb:np.float64(215.84062671661377) - actor/lr:np.float64(0.0) - training/global_step:1 - training/epoch:0 - critic/score/mean:0.53125 - critic/score/max:1.0 - critic/score/min:0.0 - critic/rewards/mean:0.53125 - critic/rewards/max:1.0 - critic/rewards/min:0.0 - critic/advantages/mean:0.004840025212615728 - critic/advantages/max:0.875 - critic/advantages/min:-0.875 - critic/returns/mean:0.004840025212615728 - critic/returns/max:0.875 - critic/returns/min:-0.875 - response_length/mean:371.15234375 - response_length/max:931.0 - response_length/min:1
|
||
step:2 - global_seqlen/min:19701 - global_seqlen/max:27799 - global_seqlen/minmax_diff:8098 - global_seqlen/balanced_min:23392 - global_seqlen/balanced_max:23397 - global_seqlen/mean:23395.0 - actor/entropy:0.1490602344274521 - perf/mfu/actor_infer:0 - training/rollout_probs_diff_valid:1 - training/rollout_probs_diff_max:0.3122412860393524 - training/rollout_probs_diff_mean:0.0028302131686359644 - training/rollout_probs_diff_std:0.010099596343934536 - training/rollout_actor_probs_pearson_corr:0.9986505508422852 - self_distillation/success_group_fraction:0.59375 - self_distillation/success_sample_fraction:0.58203125 - self_distillation/correct_sample_fraction:0.38671875 - self_distillation/feedback_available_fraction:0.0 - self_distillation/feedback_used_fraction:0.0 - self_distillation/reprompt_sample_fraction:0.58203125 - rollout_corr/rollout_is_mean:1.0000975131988525 - rollout_corr/rollout_is_ratio_fraction_high:0.0 - rollout_corr/rollout_is_ratio_fraction_low:3.368364195921458e-05 - rollout_corr/rollout_is_max:1.9516174793243408 - rollout_corr/rollout_is_min:0.02673615701496601 - rollout_corr/rollout_is_std:0.029674889519810677 - rollout_corr/rollout_is_eff_sample_size:0.9991202947500916 - rollout_corr/rollout_is_seq_mean:1.0001499652862549 - rollout_corr/rollout_is_seq_std:0.002038458827883005 - rollout_corr/rollout_is_seq_max:1.0074812173843384 - rollout_corr/rollout_is_seq_min:0.9940746426582336 - rollout_corr/rollout_is_seq_max_deviation:0.007481217384338379 - rollout_corr/rollout_is_seq_fraction_high:0.0 - rollout_corr/rollout_is_seq_fraction_low:0.0 - rollout_corr/training_ppl:np.float64(1.163191289640963) - rollout_corr/training_log_ppl:np.float64(0.14718691588495858) - rollout_corr/kl:np.float64(0.00040047891468475427) - rollout_corr/k3_kl:np.float64(0.0004855660700684439) - rollout_corr/rollout_ppl:np.float64(1.1627161586657166) - rollout_corr/rollout_log_ppl:np.float64(0.1467864363003173) - rollout_corr/log_ppl_diff:np.float64(0.00040047958464128897) - rollout_corr/log_ppl_abs_diff:np.float64(0.001473499280109536) - rollout_corr/log_ppl_diff_max:np.float64(0.012188106775283813) - rollout_corr/log_ppl_diff_min:np.float64(-0.006110548973083496) - rollout_corr/ppl_ratio:np.float64(1.0004026449751109) - rollout_corr/chi2_token:np.float64(0.00108897197060287) - rollout_corr/chi2_seq:np.float64(0.3472360894083977) - actor/pg_loss:np.float64(0.0025250519393011928) - actor/kl_loss:np.float64(0.0) - actor/pg_clipfrac:np.float64(0.00019680887908180011) - actor/ppo_kl:np.float64(6.269452830842681e-05) - actor/pg_clipfrac_lower:np.float64(0.0) - self_distillation/token_reweight_lambda:np.float64(0.5) - self_distillation/token_reweight_eps_w:np.float64(0.20000000000000004) - self_distillation/token_reweight_active_token_fraction:np.float64(0.58203125) - self_distillation/token_reweight_context_sample_fraction:np.float64(0.58203125) - self_distillation/token_reweight_correct_sample_fraction:np.float64(0.38671875) - self_distillation/token_reweight_w_mean:np.float64(74700.38124252623) - self_distillation/token_reweight_w_std:np.float64(1210201.8953123423) - self_distillation/token_reweight_w_clipped_mean:np.float64(1.0012454255484045) - self_distillation/token_reweight_w_clip_frac:np.float64(0.04983127454761416) - self_distillation/empty_target_batch:np.float64(0.41796875) - actor/grad_norm:np.float64(0.3075130842626095) - perf/mfu/actor:np.float64(0.05284626358520091) - perf/max_memory_allocated_gb:np.float64(130.98325157165527) - perf/max_memory_reserved_gb:np.float64(140.86328125) - perf/cpu_memory_used_gb:np.float64(209.64998292922974) - actor/lr:np.float64(1e-07) - training/global_step:2 - training/epoch:0 - critic/score/mean:0.38671875 - critic/score/max:1.0 - critic/score/min:0.0 - critic/rewards/mean:0.38671875 - critic/rewards/max:1.0 - critic/rewards/min:0.0 - critic/advantages/mean:0.0014736594166606665 - critic/advantages/max:0.875 - critic/advantages/min:-0.875 - critic/returns/mean:0.0014736594166606665 - critic/returns/max:0.875 - critic/returns/min:-0.875 - response_length/mean:347.90625 - response_length/
|
||
step:3 - global_seqlen/min:19078 - global_seqlen/max:26721 - global_seqlen/minmax_diff:7643 - global_seqlen/balanced_min:23358 - global_seqlen/balanced_max:23362 - global_seqlen/mean:23359.625 - actor/entropy:0.15590029954910278 - perf/mfu/actor_infer:0 - training/rollout_probs_diff_valid:1 - training/rollout_probs_diff_max:0.30163779854774475 - training/rollout_probs_diff_mean:0.0029317946173250675 - training/rollout_probs_diff_std:0.01029909122735262 - training/rollout_actor_probs_pearson_corr:0.9986982345581055 - self_distillation/success_group_fraction:0.6875 - self_distillation/success_sample_fraction:0.6796875 - self_distillation/correct_sample_fraction:0.49609375 - self_distillation/feedback_available_fraction:0.0 - self_distillation/feedback_used_fraction:0.0 - self_distillation/reprompt_sample_fraction:0.6796875 - rollout_corr/rollout_is_mean:1.0001177787780762 - rollout_corr/rollout_is_ratio_fraction_high:3.3867308957269415e-05 - rollout_corr/rollout_is_ratio_fraction_low:2.2578205971512944e-05 - rollout_corr/rollout_is_max:2.3356175422668457 - rollout_corr/rollout_is_min:0.3792489171028137 - rollout_corr/rollout_is_std:0.030862310901284218 - rollout_corr/rollout_is_eff_sample_size:0.9990486621006893 - rollout_corr/rollout_is_seq_mean:1.000056505203247 - rollout_corr/rollout_is_seq_std:0.0021426607854664326 - rollout_corr/rollout_is_seq_max:1.0069150924682617 - rollout_corr/rollout_is_seq_min:0.9846767783164978 - rollout_corr/rollout_is_seq_max_deviation:0.015323221683502197 - rollout_corr/rollout_is_seq_fraction_high:0.0 - rollout_corr/rollout_is_seq_fraction_low:0.0 - rollout_corr/training_ppl:np.float64(1.1627097805030644) - rollout_corr/training_log_ppl:np.float64(0.14735600428684847) - rollout_corr/kl:np.float64(0.0004571706433154077) - rollout_corr/k3_kl:np.float64(0.0005028695968576358) - rollout_corr/rollout_ppl:np.float64(1.1621715258806944) - rollout_corr/rollout_log_ppl:np.float64(0.14689883225946687) - rollout_corr/log_ppl_diff:np.float64(0.0004571720273816027) - rollout_corr/log_ppl_abs_diff:np.float64(0.0015406559177790768) - rollout_corr/log_ppl_diff_max:np.float64(0.020770907402038574) - rollout_corr/log_ppl_diff_min:np.float64(-0.004069015383720398) - rollout_corr/ppl_ratio:np.float64(1.0004598621744663) - rollout_corr/chi2_token:np.float64(0.0011095223017036915) - rollout_corr/chi2_seq:np.float64(0.6445373445749283) - actor/pg_loss:np.float64(0.002107005100697279) - actor/kl_loss:np.float64(0.0) - actor/pg_clipfrac:np.float64(0.00036592048172678915) - actor/ppo_kl:np.float64(1.7318454858639143e-05) - actor/pg_clipfrac_lower:np.float64(0.0) - self_distillation/token_reweight_lambda:np.float64(0.5) - self_distillation/token_reweight_eps_w:np.float64(0.20000000000000004) - self_distillation/token_reweight_active_token_fraction:np.float64(0.6796875) - self_distillation/token_reweight_context_sample_fraction:np.float64(0.6796875) - self_distillation/token_reweight_correct_sample_fraction:np.float64(0.49609375) - self_distillation/token_reweight_w_mean:np.float64(112490.0611913898) - self_distillation/token_reweight_w_std:np.float64(1413144.9096413348) - self_distillation/token_reweight_w_clipped_mean:np.float64(1.0011537729296833) - self_distillation/token_reweight_w_clip_frac:np.float64(0.0391525321174413) - self_distillation/empty_target_batch:np.float64(0.3203125) - actor/grad_norm:np.float64(0.34674258530139923) - perf/mfu/actor:np.float64(0.05266991472643874) - perf/max_memory_allocated_gb:np.float64(130.98325157165527) - perf/max_memory_reserved_gb:np.float64(140.86328125) - perf/cpu_memory_used_gb:np.float64(202.34539413452148) - actor/lr:np.float64(2e-07) - training/global_step:3 - training/epoch:0 - critic/score/mean:0.49609375 - critic/score/max:1.0 - critic/score/min:0.0 - critic/rewards/mean:0.49609375 - critic/rewards/max:1.0 - critic/rewards/min:0.0 - critic/advantages/mean:0.0016721982974559069 - critic/advantages/max:0.875 - critic/advantages/min:-0.875 - critic/returns/mean:0.0016721982974559069 - critic/returns/max:0.875 - critic/returns/min:-0.875 - response_length/mean:346.01953
|
||
step:4 - global_seqlen/min:18660 - global_seqlen/max:24954 - global_seqlen/minmax_diff:6294 - global_seqlen/balanced_min:22434 - global_seqlen/balanced_max:22441 - global_seqlen/mean:22438.625 - actor/entropy:0.16062773764133453 - perf/mfu/actor_infer:0 - training/rollout_probs_diff_valid:1 - training/rollout_probs_diff_max:0.21522796154022217 - training/rollout_probs_diff_mean:0.0029284933116286993 - training/rollout_probs_diff_std:0.009929878637194633 - training/rollout_actor_probs_pearson_corr:0.9988178610801697 - self_distillation/success_group_fraction:0.75 - self_distillation/success_sample_fraction:0.7421875 - self_distillation/correct_sample_fraction:0.51953125 - self_distillation/feedback_available_fraction:0.0 - self_distillation/feedback_used_fraction:0.0 - self_distillation/reprompt_sample_fraction:0.7421875 - rollout_corr/rollout_is_mean:0.99990314245224 - rollout_corr/rollout_is_ratio_fraction_high:1.217655972141074e-05 - rollout_corr/rollout_is_ratio_fraction_low:0.0 - rollout_corr/rollout_is_max:4.432721138000488 - rollout_corr/rollout_is_min:0.5506865978240967 - rollout_corr/rollout_is_std:0.029347693547606468 - rollout_corr/rollout_is_eff_sample_size:0.9991390970476317 - rollout_corr/rollout_is_seq_mean:0.9999251365661621 - rollout_corr/rollout_is_seq_std:0.001967586111277342 - rollout_corr/rollout_is_seq_max:1.0079396963119507 - rollout_corr/rollout_is_seq_min:0.9941493272781372 - rollout_corr/rollout_is_seq_max_deviation:0.007939696311950684 - rollout_corr/rollout_is_seq_fraction_high:0.0 - rollout_corr/rollout_is_seq_fraction_low:0.0 - rollout_corr/training_ppl:np.float64(1.1754684075713158) - rollout_corr/training_log_ppl:np.float64(0.1580598432192346) - rollout_corr/kl:np.float64(0.0005949336379469372) - rollout_corr/k3_kl:np.float64(0.0004836697877621532) - rollout_corr/rollout_ppl:np.float64(1.1747515690512955) - rollout_corr/rollout_log_ppl:np.float64(0.15746490829769755) - rollout_corr/log_ppl_diff:np.float64(0.0005949349215370603) - rollout_corr/log_ppl_abs_diff:np.float64(0.0015271891097654589) - rollout_corr/log_ppl_diff_max:np.float64(0.006942823529243469) - rollout_corr/log_ppl_diff_min:np.float64(-0.007561936974525452) - rollout_corr/ppl_ratio:np.float64(1.0005969791673124) - rollout_corr/chi2_token:np.float64(0.0007792506366968155) - rollout_corr/chi2_seq:np.float64(0.16005740803666413) - actor/pg_loss:np.float64(0.003112674574367702) - actor/kl_loss:np.float64(0.0) - actor/pg_clipfrac:np.float64(0.000264650520421128) - actor/ppo_kl:np.float64(5.038039862448329e-05) - actor/pg_clipfrac_lower:np.float64(0.0) - self_distillation/token_reweight_lambda:np.float64(0.5) - self_distillation/token_reweight_eps_w:np.float64(0.20000000000000004) - self_distillation/token_reweight_active_token_fraction:np.float64(0.7421875) - self_distillation/token_reweight_context_sample_fraction:np.float64(0.7421875) - self_distillation/token_reweight_correct_sample_fraction:np.float64(0.51953125) - self_distillation/token_reweight_w_mean:np.float64(132160.3446191419) - self_distillation/token_reweight_w_std:np.float64(1558573.065292833) - self_distillation/token_reweight_w_clipped_mean:np.float64(1.0021516405977309) - self_distillation/token_reweight_w_clip_frac:np.float64(0.057209294056519866) - self_distillation/empty_target_batch:np.float64(0.2578125) - actor/grad_norm:np.float64(0.2746843136847019) - perf/mfu/actor:np.float64(0.05098593083269999) - perf/max_memory_allocated_gb:np.float64(130.98325157165527) - perf/max_memory_reserved_gb:np.float64(140.86328125) - perf/cpu_memory_used_gb:np.float64(199.53969955444336) - actor/lr:np.float64(3e-07) - training/global_step:4 - training/epoch:0 - critic/score/mean:0.51953125 - critic/score/max:1.0 - critic/score/min:0.0 - critic/rewards/mean:0.51953125 - critic/rewards/max:1.0 - critic/rewards/min:0.0 - critic/advantages/mean:0.007302891928702593 - critic/advantages/max:0.875 - critic/advantages/min:-0.875 - critic/returns/mean:0.007302891928702593 - critic/returns/max:0.875 - critic/returns/min:-0.875 - response_length/mean:320.80078125 - response_length/max:708
|
||
step:5 - global_seqlen/min:20545 - global_seqlen/max:27821 - global_seqlen/minmax_diff:7276 - global_seqlen/balanced_min:24970 - global_seqlen/balanced_max:24979 - global_seqlen/mean:24975.625 - actor/entropy:0.15091656148433685 - perf/mfu/actor_infer:0 - training/rollout_probs_diff_valid:1 - training/rollout_probs_diff_max:0.7706084847450256 - training/rollout_probs_diff_mean:0.0028691221959888935 - training/rollout_probs_diff_std:0.011015417985618114 - training/rollout_actor_probs_pearson_corr:0.998382568359375 - self_distillation/success_group_fraction:0.75 - self_distillation/success_sample_fraction:0.734375 - self_distillation/correct_sample_fraction:0.55078125 - self_distillation/feedback_available_fraction:0.0 - self_distillation/feedback_used_fraction:0.0 - self_distillation/reprompt_sample_fraction:0.734375 - rollout_corr/rollout_is_mean:1.0000356435775757 - rollout_corr/rollout_is_ratio_fraction_high:3.040222145500593e-05 - rollout_corr/rollout_is_ratio_fraction_low:8.107259054668248e-05 - rollout_corr/rollout_is_max:4.460296630859375 - rollout_corr/rollout_is_min:0.10832203179597855 - rollout_corr/rollout_is_std:0.029624631628394127 - rollout_corr/rollout_is_eff_sample_size:0.9991231507497049 - rollout_corr/rollout_is_seq_mean:1.000041127204895 - rollout_corr/rollout_is_seq_std:0.0017892636824399233 - rollout_corr/rollout_is_seq_max:1.0100066661834717 - rollout_corr/rollout_is_seq_min:0.9941058158874512 - rollout_corr/rollout_is_seq_max_deviation:0.01000666618347168 - rollout_corr/rollout_is_seq_fraction_high:0.0 - rollout_corr/rollout_is_seq_fraction_low:0.0 - rollout_corr/training_ppl:np.float64(1.1704301312565804) - rollout_corr/training_log_ppl:np.float64(0.1531289697086322) - rollout_corr/kl:np.float64(0.0006430324443513769) - rollout_corr/k3_kl:np.float64(0.0004778158909459762) - rollout_corr/rollout_ppl:np.float64(1.1696790526621044) - rollout_corr/rollout_log_ppl:np.float64(0.15248593678552425) - rollout_corr/log_ppl_diff:np.float64(0.0006430329231079668) - rollout_corr/log_ppl_abs_diff:np.float64(0.0013642137928400189) - rollout_corr/log_ppl_diff_max:np.float64(0.006972640752792358) - rollout_corr/log_ppl_diff_min:np.float64(-0.003345698118209839) - rollout_corr/ppl_ratio:np.float64(1.000644703162834) - rollout_corr/chi2_token:np.float64(0.0006134768482297659) - rollout_corr/chi2_seq:np.float64(0.15191638213582337) - actor/pg_loss:np.float64(0.0021661146311089396) - actor/kl_loss:np.float64(0.0) - actor/pg_clipfrac:np.float64(0.00027980470531474566) - actor/ppo_kl:np.float64(0.0001787012683429623) - actor/pg_clipfrac_lower:np.float64(0.0) - self_distillation/token_reweight_lambda:np.float64(0.5) - self_distillation/token_reweight_eps_w:np.float64(0.20000000000000004) - self_distillation/token_reweight_active_token_fraction:np.float64(0.734375) - self_distillation/token_reweight_context_sample_fraction:np.float64(0.734375) - self_distillation/token_reweight_correct_sample_fraction:np.float64(0.55078125) - self_distillation/token_reweight_w_mean:np.float64(29962.478135516634) - self_distillation/token_reweight_w_std:np.float64(466126.0920916958) - self_distillation/token_reweight_w_clipped_mean:np.float64(1.0019697740208358) - self_distillation/token_reweight_w_clip_frac:np.float64(0.04917439886776265) - self_distillation/empty_target_batch:np.float64(0.265625) - actor/grad_norm:np.float64(0.25831466168165207) - perf/mfu/actor:np.float64(0.055663197405998247) - perf/max_memory_allocated_gb:np.float64(130.98325157165527) - perf/max_memory_reserved_gb:np.float64(140.86328125) - perf/cpu_memory_used_gb:np.float64(197.2271270751953) - actor/lr:np.float64(4e-07) - training/global_step:5 - training/epoch:0 - critic/score/mean:0.55078125 - critic/score/max:1.0 - critic/score/min:0.0 - critic/rewards/mean:0.55078125 - critic/rewards/max:1.0 - critic/rewards/min:0.0 - critic/advantages/mean:0.008793842978775501 - critic/advantages/max:0.875 - critic/advantages/min:-0.875 - critic/returns/mean:0.008793842978775501 - critic/returns/max:0.875 - critic/returns/min:-0.875 - response_length/mean:385.45703125 - respons
|
||
step:6 - global_seqlen/min:21570 - global_seqlen/max:26877 - global_seqlen/minmax_diff:5307 - global_seqlen/balanced_min:24460 - global_seqlen/balanced_max:24474 - global_seqlen/mean:24470.0 - actor/entropy:0.1679408699274063 - perf/mfu/actor_infer:0 - training/rollout_probs_diff_valid:1 - training/rollout_probs_diff_max:0.2932349443435669 - training/rollout_probs_diff_mean:0.0028370977379381657 - training/rollout_probs_diff_std:0.009747914038598537 - training/rollout_actor_probs_pearson_corr:0.9988725781440735 - self_distillation/success_group_fraction:0.625 - self_distillation/success_sample_fraction:0.62109375 - self_distillation/correct_sample_fraction:0.46484375 - self_distillation/feedback_available_fraction:0.0 - self_distillation/feedback_used_fraction:0.0 - self_distillation/reprompt_sample_fraction:0.62109375 - rollout_corr/rollout_is_mean:0.9999483227729797 - rollout_corr/rollout_is_ratio_fraction_high:1.0613007361826021e-05 - rollout_corr/rollout_is_ratio_fraction_low:2.1226014723652042e-05 - rollout_corr/rollout_is_max:2.0842576026916504 - rollout_corr/rollout_is_min:0.2971738278865814 - rollout_corr/rollout_is_std:0.02851119264960289 - rollout_corr/rollout_is_eff_sample_size:0.9991877721049804 - rollout_corr/rollout_is_seq_mean:0.9999916553497314 - rollout_corr/rollout_is_seq_std:0.00165208475664258 - rollout_corr/rollout_is_seq_max:1.0048378705978394 - rollout_corr/rollout_is_seq_min:0.9935877323150635 - rollout_corr/rollout_is_seq_max_deviation:0.0064122676849365234 - rollout_corr/rollout_is_seq_fraction_high:0.0 - rollout_corr/rollout_is_seq_fraction_low:0.0 - rollout_corr/training_ppl:np.float64(1.1880658236332238) - rollout_corr/training_log_ppl:np.float64(0.16674391722335713) - rollout_corr/kl:np.float64(0.0004197123074280995) - rollout_corr/k3_kl:np.float64(0.00044735126996897634) - rollout_corr/rollout_ppl:np.float64(1.1875555021688342) - rollout_corr/rollout_log_ppl:np.float64(0.16632420469250064) - rollout_corr/log_ppl_diff:np.float64(0.0004197125308564864) - rollout_corr/log_ppl_abs_diff:np.float64(0.0013543068562285043) - rollout_corr/log_ppl_diff_max:np.float64(0.007323503494262695) - rollout_corr/log_ppl_diff_min:np.float64(-0.006066650152206421) - rollout_corr/ppl_ratio:np.float64(1.0004213638603687) - rollout_corr/chi2_token:np.float64(0.0009538920130580664) - rollout_corr/chi2_seq:np.float64(0.34668579092249274) - actor/pg_loss:np.float64(0.0021740320371463895) - actor/kl_loss:np.float64(0.0) - actor/pg_clipfrac:np.float64(0.00024408937542830245) - actor/ppo_kl:np.float64(-1.3061117352108909e-05) - actor/pg_clipfrac_lower:np.float64(0.0) - self_distillation/token_reweight_lambda:np.float64(0.5) - self_distillation/token_reweight_eps_w:np.float64(0.20000000000000004) - self_distillation/token_reweight_active_token_fraction:np.float64(0.62109375) - self_distillation/token_reweight_context_sample_fraction:np.float64(0.62109375) - self_distillation/token_reweight_correct_sample_fraction:np.float64(0.46484375) - self_distillation/token_reweight_w_mean:np.float64(12125.803428596817) - self_distillation/token_reweight_w_std:np.float64(228323.23556002305) - self_distillation/token_reweight_w_clipped_mean:np.float64(1.0006039331201464) - self_distillation/token_reweight_w_clip_frac:np.float64(0.04103416547877714) - self_distillation/empty_target_batch:np.float64(0.37890625) - actor/grad_norm:np.float64(0.23871183395385742) - perf/mfu/actor:np.float64(0.055154717740877504) - perf/max_memory_allocated_gb:np.float64(130.98325157165527) - perf/max_memory_reserved_gb:np.float64(140.86328125) - perf/cpu_memory_used_gb:np.float64(195.44304656982422) - actor/lr:np.float64(5e-07) - training/global_step:6 - training/epoch:0 - critic/score/mean:0.46484375 - critic/score/max:1.0 - critic/score/min:0.0 - critic/rewards/mean:0.46484375 - critic/rewards/max:1.0 - critic/rewards/min:0.0 - critic/advantages/mean:-0.001598584232851863 - critic/advantages/max:0.875 - critic/advantages/min:-0.875 - critic/returns/mean:-0.001598584232851863 - critic/returns/max:0.875 - critic/returns/min:-0.875 - response_length/mean:36
|
||
step:7 - global_seqlen/min:20602 - global_seqlen/max:25172 - global_seqlen/minmax_diff:4570 - global_seqlen/balanced_min:23067 - global_seqlen/balanced_max:23073 - global_seqlen/mean:23070.625 - actor/entropy:0.15623463690280914 - perf/mfu/actor_infer:0 - training/rollout_probs_diff_valid:1 - training/rollout_probs_diff_max:0.2152363657951355 - training/rollout_probs_diff_mean:0.002890103030949831 - training/rollout_probs_diff_std:0.009971817955374718 - training/rollout_actor_probs_pearson_corr:0.998791515827179 - self_distillation/success_group_fraction:0.78125 - self_distillation/success_sample_fraction:0.765625 - self_distillation/correct_sample_fraction:0.5703125 - self_distillation/feedback_available_fraction:0.0 - self_distillation/feedback_used_fraction:0.0 - self_distillation/reprompt_sample_fraction:0.765625 - rollout_corr/rollout_is_mean:0.9998694062232971 - rollout_corr/rollout_is_ratio_fraction_high:0.0 - rollout_corr/rollout_is_ratio_fraction_low:0.0 - rollout_corr/rollout_is_max:1.6217707395553589 - rollout_corr/rollout_is_min:0.5242272615432739 - rollout_corr/rollout_is_std:0.028752677142620087 - rollout_corr/rollout_is_eff_sample_size:0.9991738474600336 - rollout_corr/rollout_is_seq_mean:0.9997782707214355 - rollout_corr/rollout_is_seq_std:0.0018057251581922174 - rollout_corr/rollout_is_seq_max:1.0071547031402588 - rollout_corr/rollout_is_seq_min:0.9923622608184814 - rollout_corr/rollout_is_seq_max_deviation:0.007637739181518555 - rollout_corr/rollout_is_seq_fraction_high:0.0 - rollout_corr/rollout_is_seq_fraction_low:0.0 - rollout_corr/training_ppl:np.float64(1.1638093632645905) - rollout_corr/training_log_ppl:np.float64(0.14910100147972116) - rollout_corr/kl:np.float64(0.0005835918657055572) - rollout_corr/k3_kl:np.float64(0.0005304985300256249) - rollout_corr/rollout_ppl:np.float64(1.1631218031980097) - rollout_corr/rollout_log_ppl:np.float64(0.1485174078406999) - rollout_corr/log_ppl_diff:np.float64(0.0005835936390212737) - rollout_corr/log_ppl_abs_diff:np.float64(0.0014776486423215829) - rollout_corr/log_ppl_diff_max:np.float64(0.008957058191299438) - rollout_corr/log_ppl_diff_min:np.float64(-0.006038248538970947) - rollout_corr/ppl_ratio:np.float64(1.000585664762184) - rollout_corr/chi2_token:np.float64(0.0010103294625878334) - rollout_corr/chi2_seq:np.float64(0.5633869711309671) - actor/pg_loss:np.float64(0.0023104893043637276) - actor/kl_loss:np.float64(0.0) - actor/pg_clipfrac:np.float64(0.0005257784177956637) - actor/ppo_kl:np.float64(-7.02262496155015e-05) - actor/pg_clipfrac_lower:np.float64(1.2480031728046015e-05) - self_distillation/token_reweight_lambda:np.float64(0.5) - self_distillation/token_reweight_eps_w:np.float64(0.20000000000000004) - self_distillation/token_reweight_active_token_fraction:np.float64(0.765625) - self_distillation/token_reweight_context_sample_fraction:np.float64(0.765625) - self_distillation/token_reweight_correct_sample_fraction:np.float64(0.5703125) - self_distillation/token_reweight_w_mean:np.float64(38489.73988903733) - self_distillation/token_reweight_w_std:np.float64(680399.7013303294) - self_distillation/token_reweight_w_clipped_mean:np.float64(1.0020484644919634) - self_distillation/token_reweight_w_clip_frac:np.float64(0.05133856544853188) - self_distillation/empty_target_batch:np.float64(0.234375) - actor/grad_norm:np.float64(0.20752820372581482) - perf/mfu/actor:np.float64(0.051895715182718787) - perf/max_memory_allocated_gb:np.float64(130.98325157165527) - perf/max_memory_reserved_gb:np.float64(140.86328125) - perf/cpu_memory_used_gb:np.float64(194.3308835029602) - actor/lr:np.float64(6e-07) - training/global_step:7 - training/epoch:0 - critic/score/mean:0.5703125 - critic/score/max:1.0 - critic/score/min:0.0 - critic/rewards/mean:0.5703125 - critic/rewards/max:1.0 - critic/rewards/min:0.0 - critic/advantages/mean:0.0017678930889815092 - critic/advantages/max:0.875 - critic/advantages/min:-0.875 - critic/returns/mean:0.0017678930889815092 - critic/returns/max:0.875 - critic/returns/min:-0.875 - response_length/mean:328.39453125 - response_length/max:708.0
|
||
step:8 - global_seqlen/min:24061 - global_seqlen/max:28999 - global_seqlen/minmax_diff:4938 - global_seqlen/balanced_min:25584 - global_seqlen/balanced_max:25591 - global_seqlen/mean:25588.375 - actor/entropy:0.13975059986114502 - perf/mfu/actor_infer:0 - training/rollout_probs_diff_valid:1 - training/rollout_probs_diff_max:0.3394762873649597 - training/rollout_probs_diff_mean:0.0025367196649312973 - training/rollout_probs_diff_std:0.009486767463386059 - training/rollout_actor_probs_pearson_corr:0.9987502694129944 - self_distillation/success_group_fraction:0.71875 - self_distillation/success_sample_fraction:0.70703125 - self_distillation/correct_sample_fraction:0.54296875 - self_distillation/feedback_available_fraction:0.0 - self_distillation/feedback_used_fraction:0.0 - self_distillation/reprompt_sample_fraction:0.70703125 - rollout_corr/rollout_is_mean:1.000047206878662 - rollout_corr/rollout_is_ratio_fraction_high:0.0 - rollout_corr/rollout_is_ratio_fraction_low:1.0149090485356282e-05 - rollout_corr/rollout_is_max:1.7584669589996338 - rollout_corr/rollout_is_min:0.48319488763809204 - rollout_corr/rollout_is_std:0.026686256751418114 - rollout_corr/rollout_is_eff_sample_size:0.9992883505098884 - rollout_corr/rollout_is_seq_mean:1.000035285949707 - rollout_corr/rollout_is_seq_std:0.0015413133660331368 - rollout_corr/rollout_is_seq_max:1.0042535066604614 - rollout_corr/rollout_is_seq_min:0.9951448440551758 - rollout_corr/rollout_is_seq_max_deviation:0.004855155944824219 - rollout_corr/rollout_is_seq_fraction_high:0.0 - rollout_corr/rollout_is_seq_fraction_low:0.0 - rollout_corr/training_ppl:np.float64(1.1515559651888907) - rollout_corr/training_log_ppl:np.float64(0.1378133037214866) - rollout_corr/kl:np.float64(0.0004199405625131547) - rollout_corr/k3_kl:np.float64(0.0004286105829791609) - rollout_corr/rollout_ppl:np.float64(1.151046257931739) - rollout_corr/rollout_log_ppl:np.float64(0.13739336132857716) - rollout_corr/log_ppl_diff:np.float64(0.0004199423929094337) - rollout_corr/log_ppl_abs_diff:np.float64(0.0012693901589955203) - rollout_corr/log_ppl_diff_max:np.float64(0.0062560513615608215) - rollout_corr/log_ppl_diff_min:np.float64(-0.007002726197242737) - rollout_corr/ppl_ratio:np.float64(1.0004214763175696) - rollout_corr/chi2_token:np.float64(0.0008909308817237616) - rollout_corr/chi2_seq:np.float64(0.4807258325163275) - actor/pg_loss:np.float64(0.0015849395422264934) - actor/kl_loss:np.float64(0.0) - actor/pg_clipfrac:np.float64(0.00031393380186273134) - actor/ppo_kl:np.float64(8.258763223256693e-05) - actor/pg_clipfrac_lower:np.float64(0.0) - self_distillation/token_reweight_lambda:np.float64(0.5) - self_distillation/token_reweight_eps_w:np.float64(0.20000000000000004) - self_distillation/token_reweight_active_token_fraction:np.float64(0.70703125) - self_distillation/token_reweight_context_sample_fraction:np.float64(0.70703125) - self_distillation/token_reweight_correct_sample_fraction:np.float64(0.54296875) - self_distillation/token_reweight_w_mean:np.float64(70652.89189844136) - self_distillation/token_reweight_w_std:np.float64(1263024.7517924902) - self_distillation/token_reweight_w_clipped_mean:np.float64(1.0011306994128972) - self_distillation/token_reweight_w_clip_frac:np.float64(0.03974175913754152) - self_distillation/empty_target_batch:np.float64(0.29296875) - actor/grad_norm:np.float64(0.20534029975533485) - perf/mfu/actor:np.float64(0.05763979131692125) - perf/max_memory_allocated_gb:np.float64(130.98325157165527) - perf/max_memory_reserved_gb:np.float64(140.86328125) - perf/cpu_memory_used_gb:np.float64(194.22815704345703) - actor/lr:np.float64(7e-07) - training/global_step:8 - training/epoch:0 - critic/score/mean:0.54296875 - critic/score/max:1.0 - critic/score/min:0.0 - critic/rewards/mean:0.54296875 - critic/rewards/max:1.0 - critic/rewards/min:0.0 - critic/advantages/mean:-0.0018788502784445882 - critic/advantages/max:0.875 - critic/advantages/min:-0.875 - critic/returns/mean:-0.0018788502784445882 - critic/returns/max:0.875 - critic/returns/min:-0.875 - response_length/mean:384.88671875 - respo
|
||
step:9 - global_seqlen/min:21639 - global_seqlen/max:27603 - global_seqlen/minmax_diff:5964 - global_seqlen/balanced_min:25512 - global_seqlen/balanced_max:25516 - global_seqlen/mean:25514.25 - actor/entropy:0.13293637335300446 - perf/mfu/actor_infer:0 - training/rollout_probs_diff_valid:1 - training/rollout_probs_diff_max:0.35835716128349304 - training/rollout_probs_diff_mean:0.0024616296868771315 - training/rollout_probs_diff_std:0.00920971017330885 - training/rollout_actor_probs_pearson_corr:0.998745858669281 - self_distillation/success_group_fraction:0.71875 - self_distillation/success_sample_fraction:0.71484375 - self_distillation/correct_sample_fraction:0.57421875 - self_distillation/feedback_available_fraction:0.0 - self_distillation/feedback_used_fraction:0.0 - self_distillation/reprompt_sample_fraction:0.71484375 - rollout_corr/rollout_is_mean:1.0000070333480835 - rollout_corr/rollout_is_ratio_fraction_high:2.0587569451890886e-05 - rollout_corr/rollout_is_ratio_fraction_low:2.0587569451890886e-05 - rollout_corr/rollout_is_max:2.1169991493225098 - rollout_corr/rollout_is_min:0.3592906892299652 - rollout_corr/rollout_is_std:0.02719292603433132 - rollout_corr/rollout_is_eff_sample_size:0.9992610911669125 - rollout_corr/rollout_is_seq_mean:1.000044822692871 - rollout_corr/rollout_is_seq_std:0.0015514242695644498 - rollout_corr/rollout_is_seq_max:1.006229281425476 - rollout_corr/rollout_is_seq_min:0.9937273859977722 - rollout_corr/rollout_is_seq_max_deviation:0.006272614002227783 - rollout_corr/rollout_is_seq_fraction_high:0.0 - rollout_corr/rollout_is_seq_fraction_low:0.0 - rollout_corr/training_ppl:np.float64(1.1473851264454424) - rollout_corr/training_log_ppl:np.float64(0.1347677028097678) - rollout_corr/kl:np.float64(0.0003151960298435341) - rollout_corr/k3_kl:np.float64(0.0004782045983944272) - rollout_corr/rollout_ppl:np.float64(1.1470249816775322) - rollout_corr/rollout_log_ppl:np.float64(0.13445250532822683) - rollout_corr/log_ppl_diff:np.float64(0.00031519748154096305) - rollout_corr/log_ppl_abs_diff:np.float64(0.0013751465157838538) - rollout_corr/log_ppl_diff_max:np.float64(0.006894737482070923) - rollout_corr/log_ppl_diff_min:np.float64(-0.008644089102745056) - rollout_corr/ppl_ratio:np.float64(1.000316975172609) - rollout_corr/chi2_token:np.float64(0.0013149096630513668) - rollout_corr/chi2_seq:np.float64(1.3669278046581894) - actor/pg_loss:np.float64(0.0021283028181642294) - actor/kl_loss:np.float64(0.0) - actor/pg_clipfrac:np.float64(0.0005161492645129329) - actor/ppo_kl:np.float64(-2.8284332655559297e-05) - actor/pg_clipfrac_lower:np.float64(0.0) - self_distillation/token_reweight_lambda:np.float64(0.5) - self_distillation/token_reweight_eps_w:np.float64(0.20000000000000004) - self_distillation/token_reweight_active_token_fraction:np.float64(0.71484375) - self_distillation/token_reweight_context_sample_fraction:np.float64(0.71484375) - self_distillation/token_reweight_correct_sample_fraction:np.float64(0.57421875) - self_distillation/token_reweight_w_mean:np.float64(61966.87940568989) - self_distillation/token_reweight_w_std:np.float64(1195643.6628316608) - self_distillation/token_reweight_w_clipped_mean:np.float64(1.0005232871044427) - self_distillation/token_reweight_w_clip_frac:np.float64(0.0459767281281529) - self_distillation/empty_target_batch:np.float64(0.28515625) - actor/grad_norm:np.float64(0.23176702111959457) - perf/mfu/actor:np.float64(0.05717264714519469) - perf/max_memory_allocated_gb:np.float64(130.98325157165527) - perf/max_memory_reserved_gb:np.float64(140.86328125) - perf/cpu_memory_used_gb:np.float64(194.3613052368164) - actor/lr:np.float64(8e-07) - training/global_step:9 - training/epoch:0 - critic/score/mean:0.57421875 - critic/score/max:1.0 - critic/score/min:0.0 - critic/rewards/mean:0.57421875 - critic/rewards/max:1.0 - critic/rewards/min:0.0 - critic/advantages/mean:-0.0036658740136772394 - critic/advantages/max:0.875 - critic/advantages/min:-0.875 - critic/returns/mean:-0.0036658740136772394 - critic/returns/max:0.875 - critic/returns/min:-0.875 - response_length/mean:379.47
|
||
test_gen_batch meta info: {'eos_token_id': 151645, 'pad_token_id': 151643, 'recompute_log_prob': False, 'do_sample': True, 'validate': True, 'global_steps': 10}
|
||
validation generation end
|
||
[prompt] system
|
||
|
||
Given a question and four options, please select the right answer. Respond in the following format:
|
||
<reasoning>
|
||
...
|
||
</reasoning>
|
||
<answer>
|
||
...
|
||
</answer>
|
||
|
||
For the answer, only output the letter corresponding to the correct option (A, B, C, or D), and nothing else. Do not restate the answer text. For example, if the answer is "A", just output:
|
||
<answer>
|
||
A
|
||
</answer>
|
||
|
||
user
|
||
Given the following crystal structure parameters for the material RbLa9(IrO6)4 (Material ID: mp-560657), calculate the volume of the unit cell (in ų).
|
||
|
||
Lattice Parameters:
|
||
|
||
a = 7.82 Å
|
||
b = 7.82 Å
|
||
c = 17.88 Å
|
||
α = 90.00º
|
||
β = 90.00º
|
||
γ = 90.00º
|
||
|
||
|
||
|
||
A: 1025.67 ų
|
||
|
||
B: 1094.31 ų
|
||
|
||
C: 1200.45 ų
|
||
|
||
D: 1150.78 ų
|
||
|
||
Please reason step by step.
|
||
assistant
|
||
<think>
|
||
|
||
</think>
|
||
|
||
|
||
[response] <reasoning>
|
||
The given crystal structure has lattice parameters a = 7.82 Å, b = 7.82 Å, c = 17.88 Å, and all angles (α, β, γ) are 90.00º, indicating a cubic crystal system. The volume $ V $ of a unit cell in a cubic system is calculated using the formula:
|
||
|
||
$$
|
||
V = a \times b \times c
|
||
$$
|
||
|
||
Substituting the given values:
|
||
|
||
$$
|
||
V = 7.82 \times 7.82 \times 17.88
|
||
$$
|
||
|
||
First, calculate $ 7.82 \times 7.82 $:
|
||
|
||
$$
|
||
7.82 \times 7.82 = 61.1524
|
||
$$
|
||
|
||
Then multiply by 17.88:
|
||
|
||
$$
|
||
61.1524 \times 17.88 = 1094.31
|
||
$$
|
||
|
||
Thus, the volume of the unit cell is 1094.31 ų.
|
||
</reasoning>
|
||
<answer>
|
||
B
|
||
</answer>
|
||
[ground_truth] B
|
||
[score] 1.0
|
||
[acc] 1.0
|
||
[pred] B
|
||
[incorrect_format] 1
|
||
[feedback]
|
||
local_global_step_folder: /mnt/mole/SDPO/L2T/checkpoints/datasets/sciknoweval/material/qwen3gen-material-RLSD_TR-Qwen-Qwen3-8B-mbs8-decay0-tr0.1-train32-rollout8-lr1e-6-vllm0.8/global_step_10
|
||
step:10 - global_seqlen/min:21584 - global_seqlen/max:29632 - global_seqlen/minmax_diff:8048 - global_seqlen/balanced_min:25339 - global_seqlen/balanced_max:25344 - global_seqlen/mean:25342.25 - actor/entropy:0.17639316618442535 - perf/mfu/actor_infer:0 - training/rollout_probs_diff_valid:1 - training/rollout_probs_diff_max:0.5623742341995239 - training/rollout_probs_diff_mean:0.00297564291395247 - training/rollout_probs_diff_std:0.010022003203630447 - training/rollout_actor_probs_pearson_corr:0.9988273978233337 - self_distillation/success_group_fraction:0.8125 - self_distillation/success_sample_fraction:0.8125 - self_distillation/correct_sample_fraction:0.61328125 - self_distillation/feedback_available_fraction:0.0 - self_distillation/feedback_used_fraction:0.0 - self_distillation/reprompt_sample_fraction:0.8125 - rollout_corr/rollout_is_mean:1.0002245903015137 - rollout_corr/rollout_is_ratio_fraction_high:3.9999198634177446e-05 - rollout_corr/rollout_is_ratio_fraction_low:1.9999599317088723e-05 - rollout_corr/rollout_is_max:3.417344331741333 - rollout_corr/rollout_is_min:0.29307252168655396 - rollout_corr/rollout_is_std:0.0296165831387043 - rollout_corr/rollout_is_eff_sample_size:0.9991238647521595 - rollout_corr/rollout_is_seq_mean:1.0002385377883911 - rollout_corr/rollout_is_seq_std:0.0017602697480469942 - rollout_corr/rollout_is_seq_max:1.0098330974578857 - rollout_corr/rollout_is_seq_min:0.99287348985672 - rollout_corr/rollout_is_seq_max_deviation:0.009833097457885742 - rollout_corr/rollout_is_seq_fraction_high:0.0 - rollout_corr/rollout_is_seq_fraction_low:0.0 - rollout_corr/training_ppl:np.float64(1.1970365992747247) - rollout_corr/training_log_ppl:np.float64(0.17606113059446216) - rollout_corr/kl:np.float64(0.00041335755928439966) - rollout_corr/k3_kl:np.float64(0.0006021688509321166) - rollout_corr/rollout_ppl:np.float64(1.1965299416333437) - rollout_corr/rollout_log_ppl:np.float64(0.17564777103689266) - rollout_corr/log_ppl_diff:np.float64(0.0004133595575694926) - rollout_corr/log_ppl_abs_diff:np.float64(0.001399658045556862) - rollout_corr/log_ppl_diff_max:np.float64(0.007768571376800537) - rollout_corr/log_ppl_diff_min:np.float64(-0.0074193477630615234) - rollout_corr/ppl_ratio:np.float64(1.000415229005739) - rollout_corr/chi2_token:np.float64(0.0016666329465806484) - rollout_corr/chi2_seq:np.float64(0.5307988335844129) - actor/pg_loss:np.float64(0.0035515406634658575) - actor/kl_loss:np.float64(0.0) - actor/pg_clipfrac:np.float64(0.00041831631688182824) - actor/ppo_kl:np.float64(0.00018670088022787468) - actor/pg_clipfrac_lower:np.float64(0.0) - self_distillation/token_reweight_lambda:np.float64(0.5) - self_distillation/token_reweight_eps_w:np.float64(0.20000000000000004) - self_distillation/token_reweight_active_token_fraction:np.float64(0.8125) - self_distillation/token_reweight_context_sample_fraction:np.float64(0.8125) - self_distillation/token_reweight_correct_sample_fraction:np.float64(0.61328125) - self_distillation/token_reweight_w_mean:np.float64(93607.4896354326) - self_distillation/token_reweight_w_std:np.float64(1211600.9292068654) - self_distillation/token_reweight_w_clipped_mean:np.float64(1.0003265906125307) - self_distillation/token_reweight_w_clip_frac:np.float64(0.0656057171872817) - self_distillation/empty_target_batch:np.float64(0.1875) - actor/grad_norm:np.float64(0.3258971497416496) - perf/mfu/actor:np.float64(0.057390654797647184) - perf/max_memory_allocated_gb:np.float64(130.98325157165527) - perf/max_memory_reserved_gb:np.float64(140.86328125) - perf/cpu_memory_used_gb:np.float64(194.34308242797852) - actor/lr:np.float64(9e-07) - val-aux/sciknoweval/reward/mean@16:np.float64(0.6555851063829787) - val-aux/sciknoweval/reward/std@16:np.float64(0.1378521080587476) - val-aux/sciknoweval/reward/best@2/mean:np.float64(0.7133085106382979) - val-aux/sciknoweval/reward/best@2/std:np.float64(0.09345261532381192) - val-aux/sciknoweval/reward/worst@2/mean:np.float64(0.5977872340425531) - val-aux/sciknoweval/reward/worst@2/std:np.float64(0.14093939606080857) - val-aux/sciknoweval/reward/maj@2/mea
|
||
step:11 - global_seqlen/min:23133 - global_seqlen/max:28243 - global_seqlen/minmax_diff:5110 - global_seqlen/balanced_min:25525 - global_seqlen/balanced_max:25529 - global_seqlen/mean:25526.75 - actor/entropy:0.13923069834709167 - perf/mfu/actor_infer:0 - training/rollout_probs_diff_valid:1 - training/rollout_probs_diff_max:0.3595540523529053 - training/rollout_probs_diff_mean:0.002506059827283025 - training/rollout_probs_diff_std:0.009211650118231773 - training/rollout_actor_probs_pearson_corr:0.998887836933136 - self_distillation/success_group_fraction:0.625 - self_distillation/success_sample_fraction:0.62109375 - self_distillation/correct_sample_fraction:0.5390625 - self_distillation/feedback_available_fraction:0.0 - self_distillation/feedback_used_fraction:0.0 - self_distillation/reprompt_sample_fraction:0.62109375 - rollout_corr/rollout_is_mean:1.0000337362289429 - rollout_corr/rollout_is_ratio_fraction_high:0.0 - rollout_corr/rollout_is_ratio_fraction_low:0.0 - rollout_corr/rollout_is_max:1.6297050714492798 - rollout_corr/rollout_is_min:0.5049138069152832 - rollout_corr/rollout_is_std:0.026764316484332085 - rollout_corr/rollout_is_eff_sample_size:0.9992841841384783 - rollout_corr/rollout_is_seq_mean:1.0000133514404297 - rollout_corr/rollout_is_seq_std:0.0015445039607584476 - rollout_corr/rollout_is_seq_max:1.0065466165542603 - rollout_corr/rollout_is_seq_min:0.9930241703987122 - rollout_corr/rollout_is_seq_max_deviation:0.006975829601287842 - rollout_corr/rollout_is_seq_fraction_high:0.0 - rollout_corr/rollout_is_seq_fraction_low:0.0 - rollout_corr/training_ppl:np.float64(1.1522930269129574) - rollout_corr/training_log_ppl:np.float64(0.13917134414077736) - rollout_corr/kl:np.float64(0.0005147107860317135) - rollout_corr/k3_kl:np.float64(0.0004735012372378833) - rollout_corr/rollout_ppl:np.float64(1.1516877599060535) - rollout_corr/rollout_log_ppl:np.float64(0.1386566325527383) - rollout_corr/log_ppl_diff:np.float64(0.0005147115880390629) - rollout_corr/log_ppl_abs_diff:np.float64(0.001257772310054861) - rollout_corr/log_ppl_diff_max:np.float64(0.007037058472633362) - rollout_corr/log_ppl_diff_min:np.float64(-0.005003094673156738) - rollout_corr/ppl_ratio:np.float64(1.0005161934532225) - rollout_corr/chi2_token:np.float64(0.0008689719252288342) - rollout_corr/chi2_seq:np.float64(0.1595524800941348) - actor/pg_loss:np.float64(0.0011308909161016345) - actor/kl_loss:np.float64(0.0) - actor/pg_clipfrac:np.float64(0.00022216506886252319) - actor/ppo_kl:np.float64(0.00015927736775145362) - actor/pg_clipfrac_lower:np.float64(0.0) - self_distillation/token_reweight_lambda:np.float64(0.5) - self_distillation/token_reweight_eps_w:np.float64(0.20000000000000004) - self_distillation/token_reweight_active_token_fraction:np.float64(0.62109375) - self_distillation/token_reweight_context_sample_fraction:np.float64(0.62109375) - self_distillation/token_reweight_correct_sample_fraction:np.float64(0.5390625) - self_distillation/token_reweight_w_mean:np.float64(16985.92218018137) - self_distillation/token_reweight_w_std:np.float64(351130.6301018601) - self_distillation/token_reweight_w_clipped_mean:np.float64(1.0002653282135725) - self_distillation/token_reweight_w_clip_frac:np.float64(0.021051821066066623) - self_distillation/empty_target_batch:np.float64(0.37890625) - actor/grad_norm:np.float64(0.25365722738206387) - perf/mfu/actor:np.float64(0.05824596645962646) - perf/max_memory_allocated_gb:np.float64(130.98325157165527) - perf/max_memory_reserved_gb:np.float64(140.86328125) - perf/cpu_memory_used_gb:np.float64(196.24867153167725) - actor/lr:np.float64(1e-06) - training/global_step:11 - training/epoch:0 - critic/score/mean:0.5390625 - critic/score/max:1.0 - critic/score/min:0.0 - critic/rewards/mean:0.5390625 - critic/rewards/max:1.0 - critic/rewards/min:0.0 - critic/advantages/mean:0.000572855758946389 - critic/advantages/max:0.875 - critic/advantages/min:-0.875 - critic/returns/mean:0.000572855758946389 - critic/returns/max:0.875 - critic/returns/min:-0.875 - response_length/mean:407.4296875 - response_length/max:690.0 - respons
|
||
step:12 - global_seqlen/min:20942 - global_seqlen/max:26854 - global_seqlen/minmax_diff:5912 - global_seqlen/balanced_min:24478 - global_seqlen/balanced_max:24482 - global_seqlen/mean:24480.5 - actor/entropy:0.16560876369476318 - perf/mfu/actor_infer:0 - training/rollout_probs_diff_valid:1 - training/rollout_probs_diff_max:0.35140830278396606 - training/rollout_probs_diff_mean:0.002876439830288291 - training/rollout_probs_diff_std:0.009866539388895035 - training/rollout_actor_probs_pearson_corr:0.9988297820091248 - self_distillation/success_group_fraction:0.78125 - self_distillation/success_sample_fraction:0.77734375 - self_distillation/correct_sample_fraction:0.62890625 - self_distillation/feedback_available_fraction:0.0 - self_distillation/feedback_used_fraction:0.0 - self_distillation/reprompt_sample_fraction:0.77734375 - rollout_corr/rollout_is_mean:1.000069260597229 - rollout_corr/rollout_is_ratio_fraction_high:1.034211709338706e-05 - rollout_corr/rollout_is_ratio_fraction_low:3.102635309915058e-05 - rollout_corr/rollout_is_max:4.401683330535889 - rollout_corr/rollout_is_min:0.27226293087005615 - rollout_corr/rollout_is_std:0.028915980830788612 - rollout_corr/rollout_is_eff_sample_size:0.9991645645790151 - rollout_corr/rollout_is_seq_mean:1.0001089572906494 - rollout_corr/rollout_is_seq_std:0.0018035214161500335 - rollout_corr/rollout_is_seq_max:1.0091018676757812 - rollout_corr/rollout_is_seq_min:0.9952834248542786 - rollout_corr/rollout_is_seq_max_deviation:0.00910186767578125 - rollout_corr/rollout_is_seq_fraction_high:0.0 - rollout_corr/rollout_is_seq_fraction_low:0.0 - rollout_corr/training_ppl:np.float64(1.176113523542881) - rollout_corr/training_log_ppl:np.float64(0.15972186841099756) - rollout_corr/kl:np.float64(0.0005557158914193394) - rollout_corr/k3_kl:np.float64(0.0005383371243823376) - rollout_corr/rollout_ppl:np.float64(1.1754575315862894) - rollout_corr/rollout_log_ppl:np.float64(0.159166151235695) - rollout_corr/log_ppl_diff:np.float64(0.0005557171753025614) - rollout_corr/log_ppl_abs_diff:np.float64(0.0015308187503251247) - rollout_corr/log_ppl_diff_max:np.float64(0.006714761257171631) - rollout_corr/log_ppl_diff_min:np.float64(-0.005140185356140137) - rollout_corr/ppl_ratio:np.float64(1.000557643827051) - rollout_corr/chi2_token:np.float64(0.001142165856435895) - rollout_corr/chi2_seq:np.float64(0.8536760269198567) - actor/pg_loss:np.float64(0.00240803649649024) - actor/kl_loss:np.float64(0.0) - actor/pg_clipfrac:np.float64(0.00033394563706679037) - actor/ppo_kl:np.float64(0.0002117141521171817) - actor/pg_clipfrac_lower:np.float64(0.0) - self_distillation/token_reweight_lambda:np.float64(0.5) - self_distillation/token_reweight_eps_w:np.float64(0.20000000000000004) - self_distillation/token_reweight_active_token_fraction:np.float64(0.77734375) - self_distillation/token_reweight_context_sample_fraction:np.float64(0.77734375) - self_distillation/token_reweight_correct_sample_fraction:np.float64(0.62890625) - self_distillation/token_reweight_w_mean:np.float64(61640.22243928304) - self_distillation/token_reweight_w_std:np.float64(870763.4397258498) - self_distillation/token_reweight_w_clipped_mean:np.float64(0.9999482547864318) - self_distillation/token_reweight_w_clip_frac:np.float64(0.04829584649996832) - self_distillation/empty_target_batch:np.float64(0.22265625) - actor/grad_norm:np.float64(0.23177143931388855) - perf/mfu/actor:np.float64(0.05500504927244522) - perf/max_memory_allocated_gb:np.float64(130.98325157165527) - perf/max_memory_reserved_gb:np.float64(140.86328125) - perf/cpu_memory_used_gb:np.float64(195.91437911987305) - actor/lr:np.float64(1e-06) - training/global_step:12 - training/epoch:0 - critic/score/mean:0.62890625 - critic/score/max:1.0 - critic/score/min:0.0 - critic/rewards/mean:0.62890625 - critic/rewards/max:1.0 - critic/rewards/min:0.0 - critic/advantages/mean:-0.0031685661524534225 - critic/advantages/max:0.875 - critic/advantages/min:-0.875 - critic/returns/mean:-0.0031685661524534225 - critic/returns/max:0.875 - critic/returns/min:-0.875 - response_length/mean:377.703125
|
||
step:13 - global_seqlen/min:22843 - global_seqlen/max:27080 - global_seqlen/minmax_diff:4237 - global_seqlen/balanced_min:24516 - global_seqlen/balanced_max:24518 - global_seqlen/mean:24517.25 - actor/entropy:0.15799753367900848 - perf/mfu/actor_infer:0 - training/rollout_probs_diff_valid:1 - training/rollout_probs_diff_max:0.999995231628418 - training/rollout_probs_diff_mean:0.0029094803612679243 - training/rollout_probs_diff_std:0.010655753314495087 - training/rollout_actor_probs_pearson_corr:0.9985682368278503 - self_distillation/success_group_fraction:0.625 - self_distillation/success_sample_fraction:0.625 - self_distillation/correct_sample_fraction:0.5546875 - self_distillation/feedback_available_fraction:0.0 - self_distillation/feedback_used_fraction:0.0 - self_distillation/reprompt_sample_fraction:0.625 - rollout_corr/rollout_is_mean:0.9997681975364685 - rollout_corr/rollout_is_ratio_fraction_high:0.0 - rollout_corr/rollout_is_ratio_fraction_low:2.2301019271253608e-05 - rollout_corr/rollout_is_max:1.7753705978393555 - rollout_corr/rollout_is_min:5.715029374186997e-07 - rollout_corr/rollout_is_std:0.028576970100402832 - rollout_corr/rollout_is_eff_sample_size:0.9991836065722173 - rollout_corr/rollout_is_seq_mean:0.9997411370277405 - rollout_corr/rollout_is_seq_std:0.0016131839947775006 - rollout_corr/rollout_is_seq_max:1.0061041116714478 - rollout_corr/rollout_is_seq_min:0.9933863282203674 - rollout_corr/rollout_is_seq_max_deviation:0.006613671779632568 - rollout_corr/rollout_is_seq_fraction_high:0.0 - rollout_corr/rollout_is_seq_fraction_low:0.0 - rollout_corr/training_ppl:np.float64(1.1687511443160474) - rollout_corr/training_log_ppl:np.float64(0.15300021482107695) - rollout_corr/kl:np.float64(0.0008388460865909053) - rollout_corr/k3_kl:np.float64(0.0006182161795891261) - rollout_corr/rollout_ppl:np.float64(1.1677398411557078) - rollout_corr/rollout_log_ppl:np.float64(0.15216136802337132) - rollout_corr/log_ppl_diff:np.float64(0.000838846797705628) - rollout_corr/log_ppl_abs_diff:np.float64(0.0016679515683790669) - rollout_corr/log_ppl_diff_max:np.float64(0.02418047934770584) - rollout_corr/log_ppl_diff_min:np.float64(-0.004306040704250336) - rollout_corr/ppl_ratio:np.float64(1.0008421933744103) - rollout_corr/chi2_token:np.float64(0.000714912312105298) - rollout_corr/chi2_seq:np.float64(0.22370657115243375) - actor/pg_loss:np.float64(0.0012121170293539762) - actor/kl_loss:np.float64(0.0) - actor/pg_clipfrac:np.float64(0.00016096123908937443) - actor/ppo_kl:np.float64(-4.754266964823728e-05) - actor/pg_clipfrac_lower:np.float64(0.0) - self_distillation/token_reweight_lambda:np.float64(0.5) - self_distillation/token_reweight_eps_w:np.float64(0.20000000000000004) - self_distillation/token_reweight_active_token_fraction:np.float64(0.625) - self_distillation/token_reweight_context_sample_fraction:np.float64(0.625) - self_distillation/token_reweight_correct_sample_fraction:np.float64(0.5546875) - self_distillation/token_reweight_w_mean:np.float64(26895.081849052105) - self_distillation/token_reweight_w_std:np.float64(552510.5327741585) - self_distillation/token_reweight_w_clipped_mean:np.float64(0.99996522301808) - self_distillation/token_reweight_w_clip_frac:np.float64(0.02378630399471149) - self_distillation/empty_target_batch:np.float64(0.375) - actor/grad_norm:np.float64(0.14329994656145573) - perf/mfu/actor:np.float64(0.055844523004072354) - perf/max_memory_allocated_gb:np.float64(130.98325157165527) - perf/max_memory_reserved_gb:np.float64(140.86328125) - perf/cpu_memory_used_gb:np.float64(195.88736724853516) - actor/lr:np.float64(1e-06) - training/global_step:13 - training/epoch:0 - critic/score/mean:0.5546875 - critic/score/max:1.0 - critic/score/min:0.0 - critic/rewards/mean:0.5546875 - critic/rewards/max:1.0 - critic/rewards/min:0.0 - critic/advantages/mean:0.004566133487969637 - critic/advantages/max:0.75 - critic/advantages/min:-0.875 - critic/returns/mean:0.004566133487969637 - critic/returns/max:0.75 - critic/returns/min:-0.875 - response_length/mean:350.3203125 - response_length/max:628.0 - response_len
|
||
step:14 - global_seqlen/min:19037 - global_seqlen/max:25125 - global_seqlen/minmax_diff:6088 - global_seqlen/balanced_min:22643 - global_seqlen/balanced_max:22645 - global_seqlen/mean:22644.25 - actor/entropy:0.14876748621463776 - perf/mfu/actor_infer:0 - training/rollout_probs_diff_valid:1 - training/rollout_probs_diff_max:0.4311178922653198 - training/rollout_probs_diff_mean:0.0028640730306506157 - training/rollout_probs_diff_std:0.010166647844016552 - training/rollout_actor_probs_pearson_corr:0.9985819458961487 - self_distillation/success_group_fraction:0.8125 - self_distillation/success_sample_fraction:0.8125 - self_distillation/correct_sample_fraction:0.71875 - self_distillation/feedback_available_fraction:0.0 - self_distillation/feedback_used_fraction:0.0 - self_distillation/reprompt_sample_fraction:0.8125 - rollout_corr/rollout_is_mean:0.9999682307243347 - rollout_corr/rollout_is_ratio_fraction_high:0.0 - rollout_corr/rollout_is_ratio_fraction_low:1.2039778994221706e-05 - rollout_corr/rollout_is_max:1.7173131704330444 - rollout_corr/rollout_is_min:0.015287053771317005 - rollout_corr/rollout_is_std:0.02836867794394493 - rollout_corr/rollout_is_eff_sample_size:0.9991958652393542 - rollout_corr/rollout_is_seq_mean:0.9999401569366455 - rollout_corr/rollout_is_seq_std:0.0017750123515725136 - rollout_corr/rollout_is_seq_max:1.008059024810791 - rollout_corr/rollout_is_seq_min:0.9954343438148499 - rollout_corr/rollout_is_seq_max_deviation:0.008059024810791016 - rollout_corr/rollout_is_seq_fraction_high:0.0 - rollout_corr/rollout_is_seq_fraction_low:0.0 - rollout_corr/training_ppl:np.float64(1.156547600403428) - rollout_corr/training_log_ppl:np.float64(0.14360159163334174) - rollout_corr/kl:np.float64(0.0006690974439980124) - rollout_corr/k3_kl:np.float64(0.0005123109068421172) - rollout_corr/rollout_ppl:np.float64(1.1557744019664824) - rollout_corr/rollout_log_ppl:np.float64(0.14293249235197436) - rollout_corr/log_ppl_diff:np.float64(0.0006690992813673802) - rollout_corr/log_ppl_abs_diff:np.float64(0.0015188633187790401) - rollout_corr/log_ppl_diff_max:np.float64(0.01147855818271637) - rollout_corr/log_ppl_diff_min:np.float64(-0.0071881115436553955) - rollout_corr/ppl_ratio:np.float64(1.0006714367773384) - rollout_corr/chi2_token:np.float64(0.0006734333001077175) - rollout_corr/chi2_seq:np.float64(0.2402285693679005) - actor/pg_loss:np.float64(0.0010356203420087695) - actor/kl_loss:np.float64(0.0) - actor/pg_clipfrac:np.float64(0.0002713244098231371) - actor/ppo_kl:np.float64(0.00016327230558488282) - actor/pg_clipfrac_lower:np.float64(0.0) - self_distillation/token_reweight_lambda:np.float64(0.5) - self_distillation/token_reweight_eps_w:np.float64(0.20000000000000004) - self_distillation/token_reweight_active_token_fraction:np.float64(0.8125) - self_distillation/token_reweight_context_sample_fraction:np.float64(0.8125) - self_distillation/token_reweight_correct_sample_fraction:np.float64(0.71875) - self_distillation/token_reweight_w_mean:np.float64(94791.48625119124) - self_distillation/token_reweight_w_std:np.float64(1204151.7757004234) - self_distillation/token_reweight_w_clipped_mean:np.float64(1.0003668137360364) - self_distillation/token_reweight_w_clip_frac:np.float64(0.02809872478246689) - self_distillation/empty_target_batch:np.float64(0.1875) - actor/grad_norm:np.float64(0.19146711006760597) - perf/mfu/actor:np.float64(0.05182206348917821) - perf/max_memory_allocated_gb:np.float64(130.98325157165527) - perf/max_memory_reserved_gb:np.float64(140.86328125) - perf/cpu_memory_used_gb:np.float64(195.85970306396484) - actor/lr:np.float64(1e-06) - training/global_step:14 - training/epoch:0 - critic/score/mean:0.71875 - critic/score/max:1.0 - critic/score/min:0.0 - critic/rewards/mean:0.71875 - critic/rewards/max:1.0 - critic/rewards/min:0.0 - critic/advantages/mean:-0.0017532928613945842 - critic/advantages/max:0.75 - critic/advantages/min:-0.875 - critic/returns/mean:-0.0017532928613945842 - critic/returns/max:0.75 - critic/returns/min:-0.875 - response_length/mean:324.4453125 - response_length/max:755.0 - response_le
|
||
step:15 - global_seqlen/min:20357 - global_seqlen/max:26318 - global_seqlen/minmax_diff:5961 - global_seqlen/balanced_min:23560 - global_seqlen/balanced_max:23564 - global_seqlen/mean:23562.5 - actor/entropy:0.14631496369838715 - perf/mfu/actor_infer:0 - training/rollout_probs_diff_valid:1 - training/rollout_probs_diff_max:0.3016349673271179 - training/rollout_probs_diff_mean:0.002760756528005004 - training/rollout_probs_diff_std:0.009806844405829906 - training/rollout_actor_probs_pearson_corr:0.9987110495567322 - self_distillation/success_group_fraction:0.78125 - self_distillation/success_sample_fraction:0.7734375 - self_distillation/correct_sample_fraction:0.61328125 - self_distillation/feedback_available_fraction:0.0 - self_distillation/feedback_used_fraction:0.0 - self_distillation/reprompt_sample_fraction:0.7734375 - rollout_corr/rollout_is_mean:1.0000892877578735 - rollout_corr/rollout_is_ratio_fraction_high:2.2541815269505605e-05 - rollout_corr/rollout_is_ratio_fraction_low:0.0 - rollout_corr/rollout_is_max:4.229223728179932 - rollout_corr/rollout_is_min:0.5123857259750366 - rollout_corr/rollout_is_std:0.029010646045207977 - rollout_corr/rollout_is_eff_sample_size:0.9991590901402961 - rollout_corr/rollout_is_seq_mean:1.000056266784668 - rollout_corr/rollout_is_seq_std:0.0018069344805553555 - rollout_corr/rollout_is_seq_max:1.0083919763565063 - rollout_corr/rollout_is_seq_min:0.9939211010932922 - rollout_corr/rollout_is_seq_max_deviation:0.008391976356506348 - rollout_corr/rollout_is_seq_fraction_high:0.0 - rollout_corr/rollout_is_seq_fraction_low:0.0 - rollout_corr/training_ppl:np.float64(1.1575869643129408) - rollout_corr/training_log_ppl:np.float64(0.14444846322294325) - rollout_corr/kl:np.float64(0.0006515819939068024) - rollout_corr/k3_kl:np.float64(0.0005389036207645859) - rollout_corr/rollout_ppl:np.float64(1.1568320151418447) - rollout_corr/rollout_log_ppl:np.float64(0.14379688013286795) - rollout_corr/log_ppl_diff:np.float64(0.0006515830900752917) - rollout_corr/log_ppl_abs_diff:np.float64(0.0015404976875288412) - rollout_corr/log_ppl_diff_max:np.float64(0.00785207748413086) - rollout_corr/log_ppl_diff_min:np.float64(-0.008581697940826416) - rollout_corr/ppl_ratio:np.float64(1.0006536967121065) - rollout_corr/chi2_token:np.float64(0.0009729461744427681) - rollout_corr/chi2_seq:np.float64(0.3421388922724873) - actor/pg_loss:np.float64(0.0018546859500929713) - actor/kl_loss:np.float64(0.0) - actor/pg_clipfrac:np.float64(0.0004201540004942217) - actor/ppo_kl:np.float64(0.0002769002403564258) - actor/pg_clipfrac_lower:np.float64(0.0) - self_distillation/token_reweight_lambda:np.float64(0.5) - self_distillation/token_reweight_eps_w:np.float64(0.20000000000000004) - self_distillation/token_reweight_active_token_fraction:np.float64(0.7734375) - self_distillation/token_reweight_context_sample_fraction:np.float64(0.7734375) - self_distillation/token_reweight_correct_sample_fraction:np.float64(0.61328125) - self_distillation/token_reweight_w_mean:np.float64(111500.54993374902) - self_distillation/token_reweight_w_std:np.float64(1809348.6934464541) - self_distillation/token_reweight_w_clipped_mean:np.float64(1.0001659584231675) - self_distillation/token_reweight_w_clip_frac:np.float64(0.044835288790636696) - self_distillation/empty_target_batch:np.float64(0.2265625) - actor/grad_norm:np.float64(0.28271641582250595) - perf/mfu/actor:np.float64(0.05300478413685146) - perf/max_memory_allocated_gb:np.float64(130.98325157165527) - perf/max_memory_reserved_gb:np.float64(140.86328125) - perf/cpu_memory_used_gb:np.float64(195.90932846069336) - actor/lr:np.float64(1e-06) - training/global_step:15 - training/epoch:0 - critic/score/mean:0.61328125 - critic/score/max:1.0 - critic/score/min:0.0 - critic/rewards/mean:0.61328125 - critic/rewards/max:1.0 - critic/rewards/min:0.0 - critic/advantages/mean:-0.0009932487737387419 - critic/advantages/max:0.875 - critic/advantages/min:-0.875 - critic/returns/mean:-0.0009932487737387419 - critic/returns/max:0.875 - critic/returns/min:-0.875 - response_length/mean:346.578125 - response_leng
|
||
step:16 - global_seqlen/min:21151 - global_seqlen/max:26600 - global_seqlen/minmax_diff:5449 - global_seqlen/balanced_min:24328 - global_seqlen/balanced_max:24340 - global_seqlen/mean:24336.75 - actor/entropy:0.1754712462425232 - perf/mfu/actor_infer:0 - training/rollout_probs_diff_valid:1 - training/rollout_probs_diff_max:0.43991324305534363 - training/rollout_probs_diff_mean:0.002930275397375226 - training/rollout_probs_diff_std:0.009878823533654213 - training/rollout_actor_probs_pearson_corr:0.9988336563110352 - self_distillation/success_group_fraction:0.75 - self_distillation/success_sample_fraction:0.74609375 - self_distillation/correct_sample_fraction:0.66015625 - self_distillation/feedback_available_fraction:0.0 - self_distillation/feedback_used_fraction:0.0 - self_distillation/reprompt_sample_fraction:0.74609375 - rollout_corr/rollout_is_mean:1.000189185142517 - rollout_corr/rollout_is_ratio_fraction_high:2.2015277863829397e-05 - rollout_corr/rollout_is_ratio_fraction_low:1.1007638931914698e-05 - rollout_corr/rollout_is_max:3.340711832046509 - rollout_corr/rollout_is_min:0.12017349898815155 - rollout_corr/rollout_is_std:0.02910090796649456 - rollout_corr/rollout_is_eff_sample_size:0.9991542107998265 - rollout_corr/rollout_is_seq_mean:1.000267505645752 - rollout_corr/rollout_is_seq_std:0.0017066416330635548 - rollout_corr/rollout_is_seq_max:1.0081307888031006 - rollout_corr/rollout_is_seq_min:0.9957355260848999 - rollout_corr/rollout_is_seq_max_deviation:0.008130788803100586 - rollout_corr/rollout_is_seq_fraction_high:0.0 - rollout_corr/rollout_is_seq_fraction_low:0.0 - rollout_corr/training_ppl:np.float64(1.1941090752370656) - rollout_corr/training_log_ppl:np.float64(0.17241633862431627) - rollout_corr/kl:np.float64(0.00014409434729500958) - rollout_corr/k3_kl:np.float64(0.0005055610327104887) - rollout_corr/rollout_ppl:np.float64(1.1939395442605019) - rollout_corr/rollout_log_ppl:np.float64(0.17227224242378725) - rollout_corr/log_ppl_diff:np.float64(0.00014409620052902028) - rollout_corr/log_ppl_abs_diff:np.float64(0.0012714471304207109) - rollout_corr/log_ppl_diff_max:np.float64(0.007993444800376892) - rollout_corr/log_ppl_diff_min:np.float64(-0.004106059670448303) - rollout_corr/ppl_ratio:np.float64(1.0001455154269934) - rollout_corr/chi2_token:np.float64(0.0017365424428135157) - rollout_corr/chi2_seq:np.float64(0.48078759643249214) - actor/pg_loss:np.float64(0.0014060178073123097) - actor/kl_loss:np.float64(0.0) - actor/pg_clipfrac:np.float64(0.0002084290190396132) - actor/ppo_kl:np.float64(-5.135476640383274e-05) - actor/pg_clipfrac_lower:np.float64(1.1388484381313901e-05) - self_distillation/token_reweight_lambda:np.float64(0.5) - self_distillation/token_reweight_eps_w:np.float64(0.20000000000000004) - self_distillation/token_reweight_active_token_fraction:np.float64(0.74609375) - self_distillation/token_reweight_context_sample_fraction:np.float64(0.74609375) - self_distillation/token_reweight_correct_sample_fraction:np.float64(0.66015625) - self_distillation/token_reweight_w_mean:np.float64(46786.27347472962) - self_distillation/token_reweight_w_std:np.float64(763521.2723203925) - self_distillation/token_reweight_w_clipped_mean:np.float64(0.9998286094050854) - self_distillation/token_reweight_w_clip_frac:np.float64(0.03509276293334551) - self_distillation/empty_target_batch:np.float64(0.25390625) - actor/grad_norm:np.float64(0.1977279968559742) - perf/mfu/actor:np.float64(0.05489509643176197) - perf/max_memory_allocated_gb:np.float64(130.98325157165527) - perf/max_memory_reserved_gb:np.float64(140.86328125) - perf/cpu_memory_used_gb:np.float64(195.75335693359375) - actor/lr:np.float64(1e-06) - training/global_step:16 - training/epoch:0 - critic/score/mean:0.66015625 - critic/score/max:1.0 - critic/score/min:0.0 - critic/rewards/mean:0.66015625 - critic/rewards/max:1.0 - critic/rewards/min:0.0 - critic/advantages/mean:0.001325044548138976 - critic/advantages/max:0.875 - critic/advantages/min:-0.875 - critic/returns/mean:0.001325044548138976 - critic/returns/max:0.875 - critic/returns/min:-0.875 - response_l
|
||
step:17 - global_seqlen/min:20062 - global_seqlen/max:26482 - global_seqlen/minmax_diff:6420 - global_seqlen/balanced_min:23020 - global_seqlen/balanced_max:23024 - global_seqlen/mean:23022.25 - actor/entropy:0.16389913856983185 - perf/mfu/actor_infer:0 - training/rollout_probs_diff_valid:1 - training/rollout_probs_diff_max:0.19498690962791443 - training/rollout_probs_diff_mean:0.00290701468475163 - training/rollout_probs_diff_std:0.009640076197683811 - training/rollout_actor_probs_pearson_corr:0.9987958073616028 - self_distillation/success_group_fraction:0.65625 - self_distillation/success_sample_fraction:0.6484375 - self_distillation/correct_sample_fraction:0.55078125 - self_distillation/feedback_available_fraction:0.0 - self_distillation/feedback_used_fraction:0.0 - self_distillation/reprompt_sample_fraction:0.6484375 - rollout_corr/rollout_is_mean:0.9999469518661499 - rollout_corr/rollout_is_ratio_fraction_high:1.214034273289144e-05 - rollout_corr/rollout_is_ratio_fraction_low:1.214034273289144e-05 - rollout_corr/rollout_is_max:2.1111412048339844 - rollout_corr/rollout_is_min:0.48321589827537537 - rollout_corr/rollout_is_std:0.028326625004410744 - rollout_corr/rollout_is_eff_sample_size:0.9991978885434327 - rollout_corr/rollout_is_seq_mean:0.9999359846115112 - rollout_corr/rollout_is_seq_std:0.0017924453131854534 - rollout_corr/rollout_is_seq_max:1.0070422887802124 - rollout_corr/rollout_is_seq_min:0.9944337010383606 - rollout_corr/rollout_is_seq_max_deviation:0.007042288780212402 - rollout_corr/rollout_is_seq_fraction_high:0.0 - rollout_corr/rollout_is_seq_fraction_low:0.0 - rollout_corr/training_ppl:np.float64(1.1799702616408467) - rollout_corr/training_log_ppl:np.float64(0.16088108645635657) - rollout_corr/kl:np.float64(0.0005829943140627236) - rollout_corr/k3_kl:np.float64(0.0004866512913253018) - rollout_corr/rollout_ppl:np.float64(1.1792868948541582) - rollout_corr/rollout_log_ppl:np.float64(0.16029809114843374) - rollout_corr/log_ppl_diff:np.float64(0.0005829953079228289) - rollout_corr/log_ppl_abs_diff:np.float64(0.001544620368804317) - rollout_corr/log_ppl_diff_max:np.float64(0.009654134511947632) - rollout_corr/log_ppl_diff_min:np.float64(-0.006283402442932129) - rollout_corr/ppl_ratio:np.float64(1.0005852151662111) - rollout_corr/chi2_token:np.float64(0.0007389525417238474) - rollout_corr/chi2_seq:np.float64(0.31677811266854405) - actor/pg_loss:np.float64(0.0011101557174697518) - actor/kl_loss:np.float64(0.0) - actor/pg_clipfrac:np.float64(5.436478249976062e-05) - actor/ppo_kl:np.float64(0.00010106923795805134) - actor/pg_clipfrac_lower:np.float64(0.0) - self_distillation/token_reweight_lambda:np.float64(0.5) - self_distillation/token_reweight_eps_w:np.float64(0.20000000000000004) - self_distillation/token_reweight_active_token_fraction:np.float64(0.6484375) - self_distillation/token_reweight_context_sample_fraction:np.float64(0.6484375) - self_distillation/token_reweight_correct_sample_fraction:np.float64(0.55078125) - self_distillation/token_reweight_w_mean:np.float64(50939.54964332422) - self_distillation/token_reweight_w_std:np.float64(836517.6223717686) - self_distillation/token_reweight_w_clipped_mean:np.float64(1.0006192203145474) - self_distillation/token_reweight_w_clip_frac:np.float64(0.029739680663624313) - self_distillation/empty_target_batch:np.float64(0.3515625) - actor/grad_norm:np.float64(0.17823269963264465) - perf/mfu/actor:np.float64(0.05247941166656937) - perf/max_memory_allocated_gb:np.float64(130.98325157165527) - perf/max_memory_reserved_gb:np.float64(140.86328125) - perf/cpu_memory_used_gb:np.float64(195.76335906982422) - actor/lr:np.float64(1e-06) - training/global_step:17 - training/epoch:0 - critic/score/mean:0.55078125 - critic/score/max:1.0 - critic/score/min:0.0 - critic/rewards/mean:0.55078125 - critic/rewards/max:1.0 - critic/rewards/min:0.0 - critic/advantages/mean:-0.00027619278989732265 - critic/advantages/max:0.875 - critic/advantages/min:-0.875 - critic/returns/mean:-0.00027619278989732265 - critic/returns/max:0.875 - critic/returns/min:-0.875 - response_length/mean:321
|
||
step:18 - global_seqlen/min:20589 - global_seqlen/max:24439 - global_seqlen/minmax_diff:3850 - global_seqlen/balanced_min:22796 - global_seqlen/balanced_max:22808 - global_seqlen/mean:22804.25 - actor/entropy:0.14592979848384857 - perf/mfu/actor_infer:0 - training/rollout_probs_diff_valid:1 - training/rollout_probs_diff_max:0.43096956610679626 - training/rollout_probs_diff_mean:0.0027572556864470243 - training/rollout_probs_diff_std:0.009901202283799648 - training/rollout_actor_probs_pearson_corr:0.9987658262252808 - self_distillation/success_group_fraction:0.6875 - self_distillation/success_sample_fraction:0.68359375 - self_distillation/correct_sample_fraction:0.62890625 - self_distillation/feedback_available_fraction:0.0 - self_distillation/feedback_used_fraction:0.0 - self_distillation/reprompt_sample_fraction:0.68359375 - rollout_corr/rollout_is_mean:1.000087857246399 - rollout_corr/rollout_is_ratio_fraction_high:0.0 - rollout_corr/rollout_is_ratio_fraction_low:1.2085176422260702e-05 - rollout_corr/rollout_is_max:1.678128957748413 - rollout_corr/rollout_is_min:0.2909236252307892 - rollout_corr/rollout_is_std:0.028930407017469406 - rollout_corr/rollout_is_eff_sample_size:0.9991639695284223 - rollout_corr/rollout_is_seq_mean:1.0000859498977661 - rollout_corr/rollout_is_seq_std:0.0018274225294589996 - rollout_corr/rollout_is_seq_max:1.006705641746521 - rollout_corr/rollout_is_seq_min:0.9936803579330444 - rollout_corr/rollout_is_seq_max_deviation:0.006705641746520996 - rollout_corr/rollout_is_seq_fraction_high:0.0 - rollout_corr/rollout_is_seq_fraction_low:0.0 - rollout_corr/training_ppl:np.float64(1.1651787175796926) - rollout_corr/training_log_ppl:np.float64(0.14963296141650062) - rollout_corr/kl:np.float64(0.0005520073682241389) - rollout_corr/k3_kl:np.float64(0.000504673919003551) - rollout_corr/rollout_ppl:np.float64(1.1645195069722831) - rollout_corr/rollout_log_ppl:np.float64(0.14908095342980232) - rollout_corr/log_ppl_diff:np.float64(0.0005520079866982996) - rollout_corr/log_ppl_abs_diff:np.float64(0.001438016741303727) - rollout_corr/log_ppl_diff_max:np.float64(0.008772194385528564) - rollout_corr/log_ppl_diff_min:np.float64(-0.004790559411048889) - rollout_corr/ppl_ratio:np.float64(1.0005540039855987) - rollout_corr/chi2_token:np.float64(0.0008889026939868927) - rollout_corr/chi2_seq:np.float64(0.19275876367464662) - actor/pg_loss:np.float64(0.0009842339204624295) - actor/kl_loss:np.float64(0.0) - actor/pg_clipfrac:np.float64(0.00016771803802839713) - actor/ppo_kl:np.float64(0.00019636925287880302) - actor/pg_clipfrac_lower:np.float64(0.0) - self_distillation/token_reweight_lambda:np.float64(0.5) - self_distillation/token_reweight_eps_w:np.float64(0.20000000000000004) - self_distillation/token_reweight_active_token_fraction:np.float64(0.68359375) - self_distillation/token_reweight_context_sample_fraction:np.float64(0.68359375) - self_distillation/token_reweight_correct_sample_fraction:np.float64(0.62890625) - self_distillation/token_reweight_w_mean:np.float64(41495.434180613374) - self_distillation/token_reweight_w_std:np.float64(678382.9856499105) - self_distillation/token_reweight_w_clipped_mean:np.float64(0.9997132562566549) - self_distillation/token_reweight_w_clip_frac:np.float64(0.030070643435465172) - self_distillation/empty_target_batch:np.float64(0.31640625) - actor/grad_norm:np.float64(0.1628562556579709) - perf/mfu/actor:np.float64(0.051830712186790084) - perf/max_memory_allocated_gb:np.float64(130.98325157165527) - perf/max_memory_reserved_gb:np.float64(140.86328125) - perf/cpu_memory_used_gb:np.float64(195.74004364013672) - actor/lr:np.float64(1e-06) - training/global_step:18 - training/epoch:0 - critic/score/mean:0.62890625 - critic/score/max:1.0 - critic/score/min:0.0 - critic/rewards/mean:0.62890625 - critic/rewards/max:1.0 - critic/rewards/min:0.0 - critic/advantages/mean:0.0018233510199934244 - critic/advantages/max:0.875 - critic/advantages/min:-0.875 - critic/returns/mean:0.0018233510199934244 - critic/returns/max:0.875 - critic/returns/min:-0.875 - response_length/mean:323.2265625 - respons
|
||
step:19 - global_seqlen/min:20728 - global_seqlen/max:27101 - global_seqlen/minmax_diff:6373 - global_seqlen/balanced_min:23735 - global_seqlen/balanced_max:23739 - global_seqlen/mean:23737.25 - actor/entropy:0.150180384516716 - perf/mfu/actor_infer:0 - training/rollout_probs_diff_valid:1 - training/rollout_probs_diff_max:0.22949278354644775 - training/rollout_probs_diff_mean:0.002771648345515132 - training/rollout_probs_diff_std:0.00962701253592968 - training/rollout_actor_probs_pearson_corr:0.9988031387329102 - self_distillation/success_group_fraction:0.75 - self_distillation/success_sample_fraction:0.75 - self_distillation/correct_sample_fraction:0.60546875 - self_distillation/feedback_available_fraction:0.0 - self_distillation/feedback_used_fraction:0.0 - self_distillation/reprompt_sample_fraction:0.75 - rollout_corr/rollout_is_mean:1.000072956085205 - rollout_corr/rollout_is_ratio_fraction_high:0.0 - rollout_corr/rollout_is_ratio_fraction_low:0.0 - rollout_corr/rollout_is_max:1.6179542541503906 - rollout_corr/rollout_is_min:0.5627838969230652 - rollout_corr/rollout_is_std:0.028339248150587082 - rollout_corr/rollout_is_eff_sample_size:0.9991976505072335 - rollout_corr/rollout_is_seq_mean:1.000105381011963 - rollout_corr/rollout_is_seq_std:0.0016573466127738357 - rollout_corr/rollout_is_seq_max:1.006208896636963 - rollout_corr/rollout_is_seq_min:0.994246244430542 - rollout_corr/rollout_is_seq_max_deviation:0.006208896636962891 - rollout_corr/rollout_is_seq_fraction_high:0.0 - rollout_corr/rollout_is_seq_fraction_low:0.0 - rollout_corr/training_ppl:np.float64(1.1647716057486832) - rollout_corr/training_log_ppl:np.float64(0.15018300905649085) - rollout_corr/kl:np.float64(0.00027759616153133493) - rollout_corr/k3_kl:np.float64(0.00045629992244755613) - rollout_corr/rollout_ppl:np.float64(1.1644534692168236) - rollout_corr/rollout_log_ppl:np.float64(0.14990541196311824) - rollout_corr/log_ppl_diff:np.float64(0.00027759709337260574) - rollout_corr/log_ppl_abs_diff:np.float64(0.0013563748070737347) - rollout_corr/log_ppl_diff_max:np.float64(0.006419584155082703) - rollout_corr/log_ppl_diff_min:np.float64(-0.0054564327001571655) - rollout_corr/ppl_ratio:np.float64(1.0002791315782815) - rollout_corr/chi2_token:np.float64(0.001294158399105072) - rollout_corr/chi2_seq:np.float64(0.5072512249462306) - actor/pg_loss:np.float64(0.0019059174228459597) - actor/kl_loss:np.float64(0.0) - actor/pg_clipfrac:np.float64(0.00024170369306375505) - actor/ppo_kl:np.float64(-3.959948672971336e-05) - actor/pg_clipfrac_lower:np.float64(0.0) - self_distillation/token_reweight_lambda:np.float64(0.5) - self_distillation/token_reweight_eps_w:np.float64(0.20000000000000004) - self_distillation/token_reweight_active_token_fraction:np.float64(0.75) - self_distillation/token_reweight_context_sample_fraction:np.float64(0.75) - self_distillation/token_reweight_correct_sample_fraction:np.float64(0.60546875) - self_distillation/token_reweight_w_mean:np.float64(12786.887558560818) - self_distillation/token_reweight_w_std:np.float64(239622.63726934168) - self_distillation/token_reweight_w_clipped_mean:np.float64(1.0005574417300522) - self_distillation/token_reweight_w_clip_frac:np.float64(0.03681229500216432) - self_distillation/empty_target_batch:np.float64(0.25) - actor/grad_norm:np.float64(0.2085462026298046) - perf/mfu/actor:np.float64(0.05353441975860699) - perf/max_memory_allocated_gb:np.float64(130.98325157165527) - perf/max_memory_reserved_gb:np.float64(140.86328125) - perf/cpu_memory_used_gb:np.float64(195.7970199584961) - actor/lr:np.float64(1e-06) - training/global_step:19 - training/epoch:0 - critic/score/mean:0.60546875 - critic/score/max:1.0 - critic/score/min:0.0 - critic/rewards/mean:0.60546875 - critic/rewards/max:1.0 - critic/rewards/min:0.0 - critic/advantages/mean:0.0037669024895876646 - critic/advantages/max:0.75 - critic/advantages/min:-0.875 - critic/returns/mean:0.0037669024895876646 - critic/returns/max:0.75 - critic/returns/min:-0.875 - response_length/mean:357.6328125 - response_length/max:672.0 - response_length/min:142.0 - respons
|
||
test_gen_batch meta info: {'eos_token_id': 151645, 'pad_token_id': 151643, 'recompute_log_prob': False, 'do_sample': True, 'validate': True, 'global_steps': 20}
|
||
validation generation end
|
||
[prompt] system
|
||
|
||
Given a question and four options, please select the right answer. Respond in the following format:
|
||
<reasoning>
|
||
...
|
||
</reasoning>
|
||
<answer>
|
||
...
|
||
</answer>
|
||
|
||
For the answer, only output the letter corresponding to the correct option (A, B, C, or D), and nothing else. Do not restate the answer text. For example, if the answer is "A", just output:
|
||
<answer>
|
||
A
|
||
</answer>
|
||
|
||
user
|
||
Given the following crystal structure parameters for the material RbLa9(IrO6)4 (Material ID: mp-560657), calculate the volume of the unit cell (in ų).
|
||
|
||
Lattice Parameters:
|
||
|
||
a = 7.82 Å
|
||
b = 7.82 Å
|
||
c = 17.88 Å
|
||
α = 90.00º
|
||
β = 90.00º
|
||
γ = 90.00º
|
||
|
||
|
||
|
||
A: 1025.67 ų
|
||
|
||
B: 1094.31 ų
|
||
|
||
C: 1200.45 ų
|
||
|
||
D: 1150.78 ų
|
||
|
||
Please reason step by step.
|
||
assistant
|
||
<think>
|
||
|
||
</think>
|
||
|
||
|
||
[response] <reasoning>
|
||
The given crystal structure is orthorhombic, as all angles (α, β, γ) are 90.00º. The volume $ V $ of a unit cell in an orthorhombic system is calculated using the formula:
|
||
|
||
$$
|
||
V = a \times b \times c
|
||
$$
|
||
|
||
Substituting the given lattice parameters:
|
||
|
||
$$
|
||
V = 7.82 \, \text{Å} \times 7.82 \, \text{Å} \times 17.88 \, \text{Å}
|
||
$$
|
||
|
||
First, calculate $ 7.82 \times 7.82 $:
|
||
|
||
$$
|
||
7.82 \times 7.82 = 61.1524
|
||
$$
|
||
|
||
Then multiply by 17.88:
|
||
|
||
$$
|
||
61.1524 \times 17.88 = 1094.31 \, \text{Å}^3
|
||
$$
|
||
|
||
Thus, the correct volume of the unit cell is 1094.31 ų.
|
||
</reasoning>
|
||
<answer>
|
||
B
|
||
</answer>
|
||
[ground_truth] B
|
||
[score] 1.0
|
||
[acc] 1.0
|
||
[pred] B
|
||
[incorrect_format] 1
|
||
[feedback]
|
||
local_global_step_folder: /mnt/mole/SDPO/L2T/checkpoints/datasets/sciknoweval/material/qwen3gen-material-RLSD_TR-Qwen-Qwen3-8B-mbs8-decay0-tr0.1-train32-rollout8-lr1e-6-vllm0.8/global_step_20
|
||
step:20 - global_seqlen/min:20727 - global_seqlen/max:27737 - global_seqlen/minmax_diff:7010 - global_seqlen/balanced_min:23695 - global_seqlen/balanced_max:23702 - global_seqlen/mean:23699.875 - actor/entropy:0.1525348573923111 - perf/mfu/actor_infer:0 - training/rollout_probs_diff_valid:1 - training/rollout_probs_diff_max:0.31862038373947144 - training/rollout_probs_diff_mean:0.0029412272851914167 - training/rollout_probs_diff_std:0.010256594978272915 - training/rollout_actor_probs_pearson_corr:0.9986908435821533 - self_distillation/success_group_fraction:0.71875 - self_distillation/success_sample_fraction:0.71484375 - self_distillation/correct_sample_fraction:0.62109375 - self_distillation/feedback_available_fraction:0.0 - self_distillation/feedback_used_fraction:0.0 - self_distillation/reprompt_sample_fraction:0.71484375 - rollout_corr/rollout_is_mean:0.999910295009613 - rollout_corr/rollout_is_ratio_fraction_high:0.0 - rollout_corr/rollout_is_ratio_fraction_low:2.3007811250863597e-05 - rollout_corr/rollout_is_max:1.9659780263900757 - rollout_corr/rollout_is_min:0.27226293087005615 - rollout_corr/rollout_is_std:0.029111146926879883 - rollout_corr/rollout_is_eff_sample_size:0.9991532587389499 - rollout_corr/rollout_is_seq_mean:0.999940037727356 - rollout_corr/rollout_is_seq_std:0.0016126699047163129 - rollout_corr/rollout_is_seq_max:1.0052599906921387 - rollout_corr/rollout_is_seq_min:0.9946086406707764 - rollout_corr/rollout_is_seq_max_deviation:0.005391359329223633 - rollout_corr/rollout_is_seq_fraction_high:0.0 - rollout_corr/rollout_is_seq_fraction_low:0.0 - rollout_corr/training_ppl:np.float64(1.162003489676863) - rollout_corr/training_log_ppl:np.float64(0.14800938246480655) - rollout_corr/kl:np.float64(0.0004912164907806549) - rollout_corr/k3_kl:np.float64(0.00046629576226564495) - rollout_corr/rollout_ppl:np.float64(1.1614080583676696) - rollout_corr/rollout_log_ppl:np.float64(0.14751816556963604) - rollout_corr/log_ppl_diff:np.float64(0.0004912168951705098) - rollout_corr/log_ppl_abs_diff:np.float64(0.0013770487275905907) - rollout_corr/log_ppl_diff_max:np.float64(0.006052792072296143) - rollout_corr/log_ppl_diff_min:np.float64(-0.003998763859272003) - rollout_corr/ppl_ratio:np.float64(1.0004927921108902) - rollout_corr/chi2_token:np.float64(0.0008720688056200743) - rollout_corr/chi2_seq:np.float64(0.20135757699608803) - actor/pg_loss:np.float64(0.0011970872292295098) - actor/kl_loss:np.float64(0.0) - actor/pg_clipfrac:np.float64(0.00019925688775401795) - actor/ppo_kl:np.float64(-1.4209182749524985e-05) - actor/pg_clipfrac_lower:np.float64(0.0) - self_distillation/token_reweight_lambda:np.float64(0.5) - self_distillation/token_reweight_eps_w:np.float64(0.20000000000000004) - self_distillation/token_reweight_active_token_fraction:np.float64(0.71484375) - self_distillation/token_reweight_context_sample_fraction:np.float64(0.71484375) - self_distillation/token_reweight_correct_sample_fraction:np.float64(0.62109375) - self_distillation/token_reweight_w_mean:np.float64(7654.9028110676445) - self_distillation/token_reweight_w_std:np.float64(124788.77614233922) - self_distillation/token_reweight_w_clipped_mean:np.float64(0.9999270187690854) - self_distillation/token_reweight_w_clip_frac:np.float64(0.030343565143994056) - self_distillation/empty_target_batch:np.float64(0.28515625) - actor/grad_norm:np.float64(0.14196937903761864) - perf/mfu/actor:np.float64(0.05396931844995495) - perf/max_memory_allocated_gb:np.float64(130.98325157165527) - perf/max_memory_reserved_gb:np.float64(140.86328125) - perf/cpu_memory_used_gb:np.float64(195.8521385192871) - actor/lr:np.float64(1e-06) - val-aux/sciknoweval/reward/mean@16:np.float64(0.6888297872340425) - val-aux/sciknoweval/reward/std@16:np.float64(0.11079258764978504) - val-aux/sciknoweval/reward/best@2/mean:np.float64(0.7310212765957446) - val-aux/sciknoweval/reward/best@2/std:np.float64(0.06887196928165733) - val-aux/sciknoweval/reward/worst@2/mean:np.float64(0.6454787234042554) - val-aux/sciknoweval/reward/worst@2/std:np.float64(0.11678760945598075) - val-aux/sciknoweval/re
|
||
step:21 - global_seqlen/min:20660 - global_seqlen/max:25901 - global_seqlen/minmax_diff:5241 - global_seqlen/balanced_min:23335 - global_seqlen/balanced_max:23343 - global_seqlen/mean:23339.75 - actor/entropy:0.1605978012084961 - perf/mfu/actor_infer:0 - training/rollout_probs_diff_valid:1 - training/rollout_probs_diff_max:0.2772907614707947 - training/rollout_probs_diff_mean:0.0030818735249340534 - training/rollout_probs_diff_std:0.010238205082714558 - training/rollout_actor_probs_pearson_corr:0.9987055659294128 - self_distillation/success_group_fraction:0.625 - self_distillation/success_sample_fraction:0.62109375 - self_distillation/correct_sample_fraction:0.54296875 - self_distillation/feedback_available_fraction:0.0 - self_distillation/feedback_used_fraction:0.0 - self_distillation/reprompt_sample_fraction:0.62109375 - rollout_corr/rollout_is_mean:1.0001020431518555 - rollout_corr/rollout_is_ratio_fraction_high:0.0 - rollout_corr/rollout_is_ratio_fraction_low:2.33062201004941e-05 - rollout_corr/rollout_is_max:1.8625247478485107 - rollout_corr/rollout_is_min:0.13879206776618958 - rollout_corr/rollout_is_std:0.029451090842485428 - rollout_corr/rollout_is_eff_sample_size:0.9991337418908153 - rollout_corr/rollout_is_seq_mean:1.000085711479187 - rollout_corr/rollout_is_seq_std:0.001832032692618668 - rollout_corr/rollout_is_seq_max:1.0050110816955566 - rollout_corr/rollout_is_seq_min:0.9940919280052185 - rollout_corr/rollout_is_seq_max_deviation:0.005908071994781494 - rollout_corr/rollout_is_seq_fraction_high:0.0 - rollout_corr/rollout_is_seq_fraction_low:0.0 - rollout_corr/training_ppl:np.float64(1.1757359504699707) - rollout_corr/training_log_ppl:np.float64(0.15899174506193958) - rollout_corr/kl:np.float64(0.0005214925452143859) - rollout_corr/k3_kl:np.float64(0.0005145951237324198) - rollout_corr/rollout_ppl:np.float64(1.175128533039242) - rollout_corr/rollout_log_ppl:np.float64(0.15847025081893662) - rollout_corr/log_ppl_diff:np.float64(0.0005214942430029623) - rollout_corr/log_ppl_abs_diff:np.float64(0.001540094926895108) - rollout_corr/log_ppl_diff_max:np.float64(0.01268022507429123) - rollout_corr/log_ppl_diff_min:np.float64(-0.004476115107536316) - rollout_corr/ppl_ratio:np.float64(1.0005237653385848) - rollout_corr/chi2_token:np.float64(0.000967380590736866) - rollout_corr/chi2_seq:np.float64(0.49231302621774375) - actor/pg_loss:np.float64(0.0008376834448426962) - actor/kl_loss:np.float64(0.0) - actor/pg_clipfrac:np.float64(0.00014917148973836447) - actor/ppo_kl:np.float64(0.00014587882370187266) - actor/pg_clipfrac_lower:np.float64(0.0) - self_distillation/token_reweight_lambda:np.float64(0.5) - self_distillation/token_reweight_eps_w:np.float64(0.20000000000000004) - self_distillation/token_reweight_active_token_fraction:np.float64(0.62109375) - self_distillation/token_reweight_context_sample_fraction:np.float64(0.62109375) - self_distillation/token_reweight_correct_sample_fraction:np.float64(0.54296875) - self_distillation/token_reweight_w_mean:np.float64(12218.91463384754) - self_distillation/token_reweight_w_std:np.float64(170807.23045986396) - self_distillation/token_reweight_w_clipped_mean:np.float64(1.000466012628749) - self_distillation/token_reweight_w_clip_frac:np.float64(0.025998536788392812) - self_distillation/empty_target_batch:np.float64(0.37890625) - actor/grad_norm:np.float64(0.1443084478378296) - perf/mfu/actor:np.float64(0.05338657488221517) - perf/max_memory_allocated_gb:np.float64(130.98325157165527) - perf/max_memory_reserved_gb:np.float64(140.86328125) - perf/cpu_memory_used_gb:np.float64(197.68404006958008) - actor/lr:np.float64(1e-06) - training/global_step:21 - training/epoch:0 - critic/score/mean:0.54296875 - critic/score/max:1.0 - critic/score/min:0.0 - critic/rewards/mean:0.54296875 - critic/rewards/max:1.0 - critic/rewards/min:0.0 - critic/advantages/mean:0.002923473948612809 - critic/advantages/max:0.875 - critic/advantages/min:-0.875 - critic/returns/mean:0.002923473948612809 - critic/returns/max:0.875 - critic/returns/min:-0.875 - response_length/mean:335.2109375 - response_length/
|
||
step:22 - global_seqlen/min:17725 - global_seqlen/max:27066 - global_seqlen/minmax_diff:9341 - global_seqlen/balanced_min:22584 - global_seqlen/balanced_max:22586 - global_seqlen/mean:22584.875 - actor/entropy:0.15070903301239014 - perf/mfu/actor_infer:0 - training/rollout_probs_diff_valid:1 - training/rollout_probs_diff_max:0.8226406574249268 - training/rollout_probs_diff_mean:0.002898238133639097 - training/rollout_probs_diff_std:0.010579418390989304 - training/rollout_actor_probs_pearson_corr:0.9985734224319458 - self_distillation/success_group_fraction:0.78125 - self_distillation/success_sample_fraction:0.78125 - self_distillation/correct_sample_fraction:0.69140625 - self_distillation/feedback_available_fraction:0.0 - self_distillation/feedback_used_fraction:0.0 - self_distillation/reprompt_sample_fraction:0.78125 - rollout_corr/rollout_is_mean:0.9998201727867126 - rollout_corr/rollout_is_ratio_fraction_high:0.0 - rollout_corr/rollout_is_ratio_fraction_low:2.3862077796366066e-05 - rollout_corr/rollout_is_max:1.9132329225540161 - rollout_corr/rollout_is_min:0.034405920654535294 - rollout_corr/rollout_is_std:0.02868003025650978 - rollout_corr/rollout_is_eff_sample_size:0.9991780129114268 - rollout_corr/rollout_is_seq_mean:0.9998583793640137 - rollout_corr/rollout_is_seq_std:0.0018176808953285217 - rollout_corr/rollout_is_seq_max:1.0082428455352783 - rollout_corr/rollout_is_seq_min:0.9950010180473328 - rollout_corr/rollout_is_seq_max_deviation:0.00824284553527832 - rollout_corr/rollout_is_seq_fraction_high:0.0 - rollout_corr/rollout_is_seq_fraction_low:0.0 - rollout_corr/training_ppl:np.float64(1.1599832442589104) - rollout_corr/training_log_ppl:np.float64(0.14644497753761243) - rollout_corr/kl:np.float64(0.000640428488739353) - rollout_corr/k3_kl:np.float64(0.0004915901672006839) - rollout_corr/rollout_ppl:np.float64(1.1592276142910123) - rollout_corr/rollout_log_ppl:np.float64(0.14580454777023988) - rollout_corr/log_ppl_diff:np.float64(0.0006404297673725523) - rollout_corr/log_ppl_abs_diff:np.float64(0.0014647771386080422) - rollout_corr/log_ppl_diff_max:np.float64(0.006041541695594788) - rollout_corr/log_ppl_diff_min:np.float64(-0.005858853459358215) - rollout_corr/ppl_ratio:np.float64(1.0006423159502447) - rollout_corr/chi2_token:np.float64(0.0006925887428224087) - rollout_corr/chi2_seq:np.float64(0.03699774341657758) - actor/pg_loss:np.float64(0.0015359516255557537) - actor/kl_loss:np.float64(0.0) - actor/pg_clipfrac:np.float64(0.00022245777108764742) - actor/ppo_kl:np.float64(2.9938944005714063e-05) - actor/pg_clipfrac_lower:np.float64(1.2361550943751354e-05) - self_distillation/token_reweight_lambda:np.float64(0.5) - self_distillation/token_reweight_eps_w:np.float64(0.20000000000000004) - self_distillation/token_reweight_active_token_fraction:np.float64(0.78125) - self_distillation/token_reweight_context_sample_fraction:np.float64(0.78125) - self_distillation/token_reweight_correct_sample_fraction:np.float64(0.69140625) - self_distillation/token_reweight_w_mean:np.float64(12695.067828800296) - self_distillation/token_reweight_w_std:np.float64(233788.51185772906) - self_distillation/token_reweight_w_clipped_mean:np.float64(1.000253927661106) - self_distillation/token_reweight_w_clip_frac:np.float64(0.02838251144567039) - self_distillation/empty_target_batch:np.float64(0.21875) - actor/grad_norm:np.float64(0.19917604327201843) - perf/mfu/actor:np.float64(0.05172575909415755) - perf/max_memory_allocated_gb:np.float64(130.98325157165527) - perf/max_memory_reserved_gb:np.float64(140.86328125) - perf/cpu_memory_used_gb:np.float64(197.19318389892578) - actor/lr:np.float64(1e-06) - training/global_step:22 - training/epoch:0 - critic/score/mean:0.69140625 - critic/score/max:1.0 - critic/score/min:0.0 - critic/rewards/mean:0.69140625 - critic/rewards/max:1.0 - critic/rewards/min:0.0 - critic/advantages/mean:-0.0006770864129066467 - critic/advantages/max:0.75 - critic/advantages/min:-0.875 - critic/returns/mean:-0.0006770864129066467 - critic/returns/max:0.75 - critic/returns/min:-0.875 - response_length/mean:327.40234375 -
|
||
step:23 - global_seqlen/min:16066 - global_seqlen/max:27288 - global_seqlen/minmax_diff:11222 - global_seqlen/balanced_min:22127 - global_seqlen/balanced_max:22131 - global_seqlen/mean:22129.625 - actor/entropy:0.1543504297733307 - perf/mfu/actor_infer:0 - training/rollout_probs_diff_valid:1 - training/rollout_probs_diff_max:0.4668269157409668 - training/rollout_probs_diff_mean:0.0029893529135733843 - training/rollout_probs_diff_std:0.010312956757843494 - training/rollout_actor_probs_pearson_corr:0.9986623525619507 - self_distillation/success_group_fraction:0.84375 - self_distillation/success_sample_fraction:0.8359375 - self_distillation/correct_sample_fraction:0.73828125 - self_distillation/feedback_available_fraction:0.0 - self_distillation/feedback_used_fraction:0.0 - self_distillation/reprompt_sample_fraction:0.8359375 - rollout_corr/rollout_is_mean:1.0000005960464478 - rollout_corr/rollout_is_ratio_fraction_high:0.0 - rollout_corr/rollout_is_ratio_fraction_low:1.268826235900633e-05 - rollout_corr/rollout_is_max:1.6697542667388916 - rollout_corr/rollout_is_min:0.48396989703178406 - rollout_corr/rollout_is_std:0.029727065935730934 - rollout_corr/rollout_is_eff_sample_size:0.9991170817700432 - rollout_corr/rollout_is_seq_mean:0.9999738335609436 - rollout_corr/rollout_is_seq_std:0.0019343866733834147 - rollout_corr/rollout_is_seq_max:1.0052357912063599 - rollout_corr/rollout_is_seq_min:0.9925729632377625 - rollout_corr/rollout_is_seq_max_deviation:0.007427036762237549 - rollout_corr/rollout_is_seq_fraction_high:0.0 - rollout_corr/rollout_is_seq_fraction_low:0.0 - rollout_corr/training_ppl:np.float64(1.1613912233151495) - rollout_corr/training_log_ppl:np.float64(0.1482296362228226) - rollout_corr/kl:np.float64(0.0005620803344044134) - rollout_corr/k3_kl:np.float64(0.00048597054205856693) - rollout_corr/rollout_ppl:np.float64(1.1607246096245944) - rollout_corr/rollout_log_ppl:np.float64(0.14766755513846874) - rollout_corr/log_ppl_diff:np.float64(0.0005620810843538493) - rollout_corr/log_ppl_abs_diff:np.float64(0.0015226107789203525) - rollout_corr/log_ppl_diff_max:np.float64(0.009226351976394653) - rollout_corr/log_ppl_diff_min:np.float64(-0.004461884498596191) - rollout_corr/ppl_ratio:np.float64(1.0005641290917993) - rollout_corr/chi2_token:np.float64(0.0008328419644385576) - rollout_corr/chi2_seq:np.float64(0.15795218478888273) - actor/pg_loss:np.float64(0.0010325053008273244) - actor/kl_loss:np.float64(0.0) - actor/pg_clipfrac:np.float64(0.00012042900743836071) - actor/ppo_kl:np.float64(8.446178508592084e-05) - actor/pg_clipfrac_lower:np.float64(0.0) - self_distillation/token_reweight_lambda:np.float64(0.5) - self_distillation/token_reweight_eps_w:np.float64(0.20000000000000004) - self_distillation/token_reweight_active_token_fraction:np.float64(0.8359375) - self_distillation/token_reweight_context_sample_fraction:np.float64(0.8359375) - self_distillation/token_reweight_correct_sample_fraction:np.float64(0.73828125) - self_distillation/token_reweight_w_mean:np.float64(40982.084487203276) - self_distillation/token_reweight_w_std:np.float64(662032.4441022928) - self_distillation/token_reweight_w_clipped_mean:np.float64(1.0005126828327775) - self_distillation/token_reweight_w_clip_frac:np.float64(0.025580635032383725) - self_distillation/empty_target_batch:np.float64(0.1640625) - actor/grad_norm:np.float64(0.1630999743938446) - perf/mfu/actor:np.float64(0.050710404347214785) - perf/max_memory_allocated_gb:np.float64(130.98325157165527) - perf/max_memory_reserved_gb:np.float64(140.86328125) - perf/cpu_memory_used_gb:np.float64(197.17858123779297) - actor/lr:np.float64(1e-06) - training/global_step:23 - training/epoch:0 - critic/score/mean:0.73828125 - critic/score/max:1.0 - critic/score/min:0.0 - critic/rewards/mean:0.73828125 - critic/rewards/max:1.0 - critic/rewards/min:0.0 - critic/advantages/mean:0.002686739433556795 - critic/advantages/max:0.875 - critic/advantages/min:-0.875 - critic/returns/mean:0.002686739433556795 - critic/returns/max:0.875 - critic/returns/min:-0.875 - response_length/mean:307.86328125 - response
|
||
step:24 - global_seqlen/min:19871 - global_seqlen/max:26117 - global_seqlen/minmax_diff:6246 - global_seqlen/balanced_min:22997 - global_seqlen/balanced_max:23000 - global_seqlen/mean:22999.125 - actor/entropy:0.13676773011684418 - perf/mfu/actor_infer:0 - training/rollout_probs_diff_valid:1 - training/rollout_probs_diff_max:0.22949346899986267 - training/rollout_probs_diff_mean:0.00260890438221395 - training/rollout_probs_diff_std:0.009457413107156754 - training/rollout_actor_probs_pearson_corr:0.9987369179725647 - self_distillation/success_group_fraction:0.90625 - self_distillation/success_sample_fraction:0.90625 - self_distillation/correct_sample_fraction:0.8515625 - self_distillation/feedback_available_fraction:0.0 - self_distillation/feedback_used_fraction:0.0 - self_distillation/reprompt_sample_fraction:0.90625 - rollout_corr/rollout_is_mean:0.9999614357948303 - rollout_corr/rollout_is_ratio_fraction_high:1.1675559107970912e-05 - rollout_corr/rollout_is_ratio_fraction_low:3.502667823340744e-05 - rollout_corr/rollout_is_max:2.3356175422668457 - rollout_corr/rollout_is_min:0.3921355903148651 - rollout_corr/rollout_is_std:0.027214836329221725 - rollout_corr/rollout_is_eff_sample_size:0.9992600198693956 - rollout_corr/rollout_is_seq_mean:0.999939501285553 - rollout_corr/rollout_is_seq_std:0.0015718950890004635 - rollout_corr/rollout_is_seq_max:1.0046483278274536 - rollout_corr/rollout_is_seq_min:0.9949087500572205 - rollout_corr/rollout_is_seq_max_deviation:0.005091249942779541 - rollout_corr/rollout_is_seq_fraction_high:0.0 - rollout_corr/rollout_is_seq_fraction_low:0.0 - rollout_corr/training_ppl:np.float64(1.1496808505617082) - rollout_corr/training_log_ppl:np.float64(0.13687511737953173) - rollout_corr/kl:np.float64(0.0004387259168705526) - rollout_corr/k3_kl:np.float64(0.00045451685279829235) - rollout_corr/rollout_ppl:np.float64(1.1491743493825197) - rollout_corr/rollout_log_ppl:np.float64(0.13643639061774593) - rollout_corr/log_ppl_diff:np.float64(0.0004387267617858015) - rollout_corr/log_ppl_abs_diff:np.float64(0.00128690085693961) - rollout_corr/log_ppl_diff_max:np.float64(0.005763456225395203) - rollout_corr/log_ppl_diff_min:np.float64(-0.004293650388717651) - rollout_corr/ppl_ratio:np.float64(1.0004402522463351) - rollout_corr/chi2_token:np.float64(0.0009429841302335262) - rollout_corr/chi2_seq:np.float64(0.39535231585614383) - actor/pg_loss:np.float64(0.0010843543568626046) - actor/kl_loss:np.float64(0.0) - actor/pg_clipfrac:np.float64(8.718408753338736e-05) - actor/ppo_kl:np.float64(-7.393197986260702e-06) - actor/pg_clipfrac_lower:np.float64(0.0) - self_distillation/token_reweight_lambda:np.float64(0.5) - self_distillation/token_reweight_eps_w:np.float64(0.20000000000000004) - self_distillation/token_reweight_active_token_fraction:np.float64(0.90625) - self_distillation/token_reweight_context_sample_fraction:np.float64(0.90625) - self_distillation/token_reweight_correct_sample_fraction:np.float64(0.8515625) - self_distillation/token_reweight_w_mean:np.float64(22220.938996580895) - self_distillation/token_reweight_w_std:np.float64(409145.10946293676) - self_distillation/token_reweight_w_clipped_mean:np.float64(0.9995599235408008) - self_distillation/token_reweight_w_clip_frac:np.float64(0.022219942387891933) - self_distillation/empty_target_batch:np.float64(0.09375) - actor/grad_norm:np.float64(0.15542156994342804) - perf/mfu/actor:np.float64(0.052601399516352225) - perf/max_memory_allocated_gb:np.float64(130.98325157165527) - perf/max_memory_reserved_gb:np.float64(140.86328125) - perf/cpu_memory_used_gb:np.float64(197.20500946044922) - actor/lr:np.float64(1e-06) - training/global_step:24 - training/epoch:0 - critic/score/mean:0.8515625 - critic/score/max:1.0 - critic/score/min:0.0 - critic/rewards/mean:0.8515625 - critic/rewards/max:1.0 - critic/rewards/min:0.0 - critic/advantages/mean:0.001507606590166688 - critic/advantages/max:0.5 - critic/advantages/min:-0.875 - critic/returns/mean:0.001507606590166688 - critic/returns/max:0.5 - critic/returns/min:-0.875 - response_length/mean:334.56640625 - response_
|
||
step:25 - global_seqlen/min:18321 - global_seqlen/max:25199 - global_seqlen/minmax_diff:6878 - global_seqlen/balanced_min:22271 - global_seqlen/balanced_max:22278 - global_seqlen/mean:22275.5 - actor/entropy:0.14500775933265686 - perf/mfu/actor_infer:0 - training/rollout_probs_diff_valid:1 - training/rollout_probs_diff_max:0.21093684434890747 - training/rollout_probs_diff_mean:0.0027230808045715094 - training/rollout_probs_diff_std:0.009582200087606907 - training/rollout_actor_probs_pearson_corr:0.9987487196922302 - self_distillation/success_group_fraction:0.78125 - self_distillation/success_sample_fraction:0.7734375 - self_distillation/correct_sample_fraction:0.625 - self_distillation/feedback_available_fraction:0.0 - self_distillation/feedback_used_fraction:0.0 - self_distillation/reprompt_sample_fraction:0.7734375 - rollout_corr/rollout_is_mean:0.9998183846473694 - rollout_corr/rollout_is_ratio_fraction_high:0.0 - rollout_corr/rollout_is_ratio_fraction_low:2.5558452762197703e-05 - rollout_corr/rollout_is_max:1.5865436792373657 - rollout_corr/rollout_is_min:0.32034292817115784 - rollout_corr/rollout_is_std:0.027499165385961533 - rollout_corr/rollout_is_eff_sample_size:0.9992440697114005 - rollout_corr/rollout_is_seq_mean:0.9998300671577454 - rollout_corr/rollout_is_seq_std:0.0015994401182979345 - rollout_corr/rollout_is_seq_max:1.005348563194275 - rollout_corr/rollout_is_seq_min:0.9952322840690613 - rollout_corr/rollout_is_seq_max_deviation:0.005348563194274902 - rollout_corr/rollout_is_seq_fraction_high:0.0 - rollout_corr/rollout_is_seq_fraction_low:0.0 - rollout_corr/training_ppl:np.float64(1.1642502681352198) - rollout_corr/training_log_ppl:np.float64(0.14770634190062992) - rollout_corr/kl:np.float64(0.0004993382486020437) - rollout_corr/k3_kl:np.float64(0.0004678560410127375) - rollout_corr/rollout_ppl:np.float64(1.163647570181638) - rollout_corr/rollout_log_ppl:np.float64(0.14720700281031895) - rollout_corr/log_ppl_diff:np.float64(0.0004993390903109685) - rollout_corr/log_ppl_abs_diff:np.float64(0.0014981203275965527) - rollout_corr/log_ppl_diff_max:np.float64(0.007449612021446228) - rollout_corr/log_ppl_diff_min:np.float64(-0.004783540964126587) - rollout_corr/ppl_ratio:np.float64(1.0005014017224312) - rollout_corr/chi2_token:np.float64(0.0008534942753612995) - rollout_corr/chi2_seq:np.float64(0.24121627327986062) - actor/pg_loss:np.float64(0.0011007965076714754) - actor/kl_loss:np.float64(0.0) - actor/pg_clipfrac:np.float64(0.00024173177780539845) - actor/ppo_kl:np.float64(-6.909645383146312e-05) - actor/pg_clipfrac_lower:np.float64(0.0) - self_distillation/token_reweight_lambda:np.float64(0.5) - self_distillation/token_reweight_eps_w:np.float64(0.20000000000000004) - self_distillation/token_reweight_active_token_fraction:np.float64(0.7734375) - self_distillation/token_reweight_context_sample_fraction:np.float64(0.7734375) - self_distillation/token_reweight_correct_sample_fraction:np.float64(0.625) - self_distillation/token_reweight_w_mean:np.float64(120953.98852239526) - self_distillation/token_reweight_w_std:np.float64(1569361.12798946) - self_distillation/token_reweight_w_clipped_mean:np.float64(1.0008900640532374) - self_distillation/token_reweight_w_clip_frac:np.float64(0.031809512540348805) - self_distillation/empty_target_batch:np.float64(0.2265625) - actor/grad_norm:np.float64(0.23915600031614304) - perf/mfu/actor:np.float64(0.05094053058519299) - perf/max_memory_allocated_gb:np.float64(130.98325157165527) - perf/max_memory_reserved_gb:np.float64(140.86328125) - perf/cpu_memory_used_gb:np.float64(197.14062881469727) - actor/lr:np.float64(1e-06) - training/global_step:25 - training/epoch:0 - critic/score/mean:0.625 - critic/score/max:1.0 - critic/score/min:0.0 - critic/rewards/mean:0.625 - critic/rewards/max:1.0 - critic/rewards/min:0.0 - critic/advantages/mean:0.005090923979878426 - critic/advantages/max:0.875 - critic/advantages/min:-0.875 - critic/returns/mean:0.005090923979878426 - critic/returns/max:0.875 - critic/returns/min:-0.875 - response_length/mean:305.671875 - response_length/max:580.0 - resp
|
||
step:26 - global_seqlen/min:21549 - global_seqlen/max:26767 - global_seqlen/minmax_diff:5218 - global_seqlen/balanced_min:23515 - global_seqlen/balanced_max:23518 - global_seqlen/mean:23516.625 - actor/entropy:0.13138793408870697 - perf/mfu/actor_infer:0 - training/rollout_probs_diff_valid:1 - training/rollout_probs_diff_max:0.635709285736084 - training/rollout_probs_diff_mean:0.002796715358272195 - training/rollout_probs_diff_std:0.01032547652721405 - training/rollout_actor_probs_pearson_corr:0.9984015822410583 - self_distillation/success_group_fraction:0.78125 - self_distillation/success_sample_fraction:0.77734375 - self_distillation/correct_sample_fraction:0.66796875 - self_distillation/feedback_available_fraction:0.0 - self_distillation/feedback_used_fraction:0.0 - self_distillation/reprompt_sample_fraction:0.77734375 - rollout_corr/rollout_is_mean:0.9999243021011353 - rollout_corr/rollout_is_ratio_fraction_high:0.0 - rollout_corr/rollout_is_ratio_fraction_low:3.467847272986546e-05 - rollout_corr/rollout_is_max:1.7744193077087402 - rollout_corr/rollout_is_min:0.29728731513023376 - rollout_corr/rollout_is_std:0.028742309659719467 - rollout_corr/rollout_is_eff_sample_size:0.999174442522392 - rollout_corr/rollout_is_seq_mean:0.9999179840087891 - rollout_corr/rollout_is_seq_std:0.0018690851284191012 - rollout_corr/rollout_is_seq_max:1.0080863237380981 - rollout_corr/rollout_is_seq_min:0.9940353631973267 - rollout_corr/rollout_is_seq_max_deviation:0.008086323738098145 - rollout_corr/rollout_is_seq_fraction_high:0.0 - rollout_corr/rollout_is_seq_fraction_low:0.0 - rollout_corr/training_ppl:np.float64(1.1414004624821246) - rollout_corr/training_log_ppl:np.float64(0.13071745182242012) - rollout_corr/kl:np.float64(0.0004450980413324146) - rollout_corr/k3_kl:np.float64(0.0005135989932227858) - rollout_corr/rollout_ppl:np.float64(1.140865498688072) - rollout_corr/rollout_log_ppl:np.float64(0.13027235258050496) - rollout_corr/log_ppl_diff:np.float64(0.00044509924191515893) - rollout_corr/log_ppl_abs_diff:np.float64(0.0015409416664624587) - rollout_corr/log_ppl_diff_max:np.float64(0.008853912353515625) - rollout_corr/log_ppl_diff_min:np.float64(-0.005702786147594452) - rollout_corr/ppl_ratio:np.float64(1.000447248108685) - rollout_corr/chi2_token:np.float64(0.0011436883360147476) - rollout_corr/chi2_seq:np.float64(0.6314138330053538) - actor/pg_loss:np.float64(0.0011609047651290894) - actor/kl_loss:np.float64(0.0) - actor/pg_clipfrac:np.float64(0.0002807997716445243) - actor/ppo_kl:np.float64(-7.160629040470212e-05) - actor/pg_clipfrac_lower:np.float64(0.0) - self_distillation/token_reweight_lambda:np.float64(0.5) - self_distillation/token_reweight_eps_w:np.float64(0.20000000000000004) - self_distillation/token_reweight_active_token_fraction:np.float64(0.77734375) - self_distillation/token_reweight_context_sample_fraction:np.float64(0.77734375) - self_distillation/token_reweight_correct_sample_fraction:np.float64(0.66796875) - self_distillation/token_reweight_w_mean:np.float64(97124.35309204645) - self_distillation/token_reweight_w_std:np.float64(1541826.973082643) - self_distillation/token_reweight_w_clipped_mean:np.float64(1.000492023769766) - self_distillation/token_reweight_w_clip_frac:np.float64(0.02323406282812357) - self_distillation/empty_target_batch:np.float64(0.22265625) - actor/grad_norm:np.float64(0.14892314560711384) - perf/mfu/actor:np.float64(0.05380290758664994) - perf/max_memory_allocated_gb:np.float64(130.98325157165527) - perf/max_memory_reserved_gb:np.float64(140.86328125) - perf/cpu_memory_used_gb:np.float64(197.13983535766602) - actor/lr:np.float64(1e-06) - training/global_step:26 - training/epoch:0 - critic/score/mean:0.66796875 - critic/score/max:1.0 - critic/score/min:0.0 - critic/rewards/mean:0.66796875 - critic/rewards/max:1.0 - critic/rewards/min:0.0 - critic/advantages/mean:0.0030618200544267893 - critic/advantages/max:0.875 - critic/advantages/min:-0.875 - critic/returns/mean:0.0030618200544267893 - critic/returns/max:0.875 - critic/returns/min:-0.875 - response_length/mean:337.92578125 - response_l
|
||
step:27 - global_seqlen/min:19791 - global_seqlen/max:24654 - global_seqlen/minmax_diff:4863 - global_seqlen/balanced_min:22366 - global_seqlen/balanced_max:22371 - global_seqlen/mean:22369.25 - actor/entropy:0.10306167602539062 - perf/mfu/actor_infer:0 - training/rollout_probs_diff_valid:1 - training/rollout_probs_diff_max:0.31862038373947144 - training/rollout_probs_diff_mean:0.002335259458050132 - training/rollout_probs_diff_std:0.009463846683502197 - training/rollout_actor_probs_pearson_corr:0.9983808398246765 - self_distillation/success_group_fraction:0.9375 - self_distillation/success_sample_fraction:0.93359375 - self_distillation/correct_sample_fraction:0.81640625 - self_distillation/feedback_available_fraction:0.0 - self_distillation/feedback_used_fraction:0.0 - self_distillation/reprompt_sample_fraction:0.93359375 - rollout_corr/rollout_is_mean:0.999907910823822 - rollout_corr/rollout_is_ratio_fraction_high:2.5938317776308395e-05 - rollout_corr/rollout_is_ratio_fraction_low:0.0 - rollout_corr/rollout_is_max:2.11165452003479 - rollout_corr/rollout_is_min:0.589073896408081 - rollout_corr/rollout_is_std:0.026092225685715675 - rollout_corr/rollout_is_eff_sample_size:0.999319539879347 - rollout_corr/rollout_is_seq_mean:0.999942421913147 - rollout_corr/rollout_is_seq_std:0.001715592690743506 - rollout_corr/rollout_is_seq_max:1.0085691213607788 - rollout_corr/rollout_is_seq_min:0.9957650899887085 - rollout_corr/rollout_is_seq_max_deviation:0.008569121360778809 - rollout_corr/rollout_is_seq_fraction_high:0.0 - rollout_corr/rollout_is_seq_fraction_low:0.0 - rollout_corr/training_ppl:np.float64(1.1074565374292433) - rollout_corr/training_log_ppl:np.float64(0.10104348854656564) - rollout_corr/kl:np.float64(0.0006335088800497601) - rollout_corr/k3_kl:np.float64(0.0005041445960429769) - rollout_corr/rollout_ppl:np.float64(1.1067317719571292) - rollout_corr/rollout_log_ppl:np.float64(0.10040997985197464) - rollout_corr/log_ppl_diff:np.float64(0.0006335086945910007) - rollout_corr/log_ppl_abs_diff:np.float64(0.0015751363680465147) - rollout_corr/log_ppl_diff_max:np.float64(0.00826522707939148) - rollout_corr/log_ppl_diff_min:np.float64(-0.005233548581600189) - rollout_corr/ppl_ratio:np.float64(1.0006358111277223) - rollout_corr/chi2_token:np.float64(0.0006916692946106195) - rollout_corr/chi2_seq:np.float64(0.19716444029472768) - actor/pg_loss:np.float64(0.0012738716322928667) - actor/kl_loss:np.float64(0.0) - actor/pg_clipfrac:np.float64(0.0003194190621798043) - actor/ppo_kl:np.float64(0.00022773302247180993) - actor/pg_clipfrac_lower:np.float64(0.0) - self_distillation/token_reweight_lambda:np.float64(0.5) - self_distillation/token_reweight_eps_w:np.float64(0.20000000000000004) - self_distillation/token_reweight_active_token_fraction:np.float64(0.93359375) - self_distillation/token_reweight_context_sample_fraction:np.float64(0.93359375) - self_distillation/token_reweight_correct_sample_fraction:np.float64(0.81640625) - self_distillation/token_reweight_w_mean:np.float64(77265.22965213796) - self_distillation/token_reweight_w_std:np.float64(1212805.6968070571) - self_distillation/token_reweight_w_clipped_mean:np.float64(0.9997106643859297) - self_distillation/token_reweight_w_clip_frac:np.float64(0.031144319364102557) - self_distillation/empty_target_batch:np.float64(0.06640625) - actor/grad_norm:np.float64(0.21108173206448555) - perf/mfu/actor:np.float64(0.05175513238257462) - perf/max_memory_allocated_gb:np.float64(130.98325157165527) - perf/max_memory_reserved_gb:np.float64(140.86328125) - perf/cpu_memory_used_gb:np.float64(198.09855270385742) - actor/lr:np.float64(1e-06) - training/global_step:27 - training/epoch:1 - critic/score/mean:0.81640625 - critic/score/max:1.0 - critic/score/min:0.0 - critic/rewards/mean:0.81640625 - critic/rewards/max:1.0 - critic/rewards/min:0.0 - critic/advantages/mean:-0.0017638056306168437 - critic/advantages/max:0.875 - critic/advantages/min:-0.875 - critic/returns/mean:-0.0017638056306168437 - critic/returns/max:0.875 - critic/returns/min:-0.875 - response_length/mean:301.1953125 - response_l
|
||
step:28 - global_seqlen/min:18667 - global_seqlen/max:26411 - global_seqlen/minmax_diff:7744 - global_seqlen/balanced_min:22005 - global_seqlen/balanced_max:22010 - global_seqlen/mean:22008.25 - actor/entropy:0.1240900456905365 - perf/mfu/actor_infer:0 - training/rollout_probs_diff_valid:1 - training/rollout_probs_diff_max:0.5311316251754761 - training/rollout_probs_diff_mean:0.0027813054621219635 - training/rollout_probs_diff_std:0.01046876423060894 - training/rollout_actor_probs_pearson_corr:0.9983556866645813 - self_distillation/success_group_fraction:0.625 - self_distillation/success_sample_fraction:0.625 - self_distillation/correct_sample_fraction:0.5859375 - self_distillation/feedback_available_fraction:0.0 - self_distillation/feedback_used_fraction:0.0 - self_distillation/reprompt_sample_fraction:0.625 - rollout_corr/rollout_is_mean:0.9999943971633911 - rollout_corr/rollout_is_ratio_fraction_high:1.2547366168291774e-05 - rollout_corr/rollout_is_ratio_fraction_low:5.0189464673167095e-05 - rollout_corr/rollout_is_max:3.121652126312256 - rollout_corr/rollout_is_min:0.003787645371630788 - rollout_corr/rollout_is_std:0.028779612854123116 - rollout_corr/rollout_is_eff_sample_size:0.9991721812892008 - rollout_corr/rollout_is_seq_mean:0.9999889135360718 - rollout_corr/rollout_is_seq_std:0.0019287803443148732 - rollout_corr/rollout_is_seq_max:1.0120155811309814 - rollout_corr/rollout_is_seq_min:0.9934726357460022 - rollout_corr/rollout_is_seq_max_deviation:0.012015581130981445 - rollout_corr/rollout_is_seq_fraction_high:0.0 - rollout_corr/rollout_is_seq_fraction_low:0.0 - rollout_corr/training_ppl:np.float64(1.1382743194699287) - rollout_corr/training_log_ppl:np.float64(0.12592740579566453) - rollout_corr/kl:np.float64(0.0009248649318678304) - rollout_corr/k3_kl:np.float64(0.0007315356932053874) - rollout_corr/rollout_ppl:np.float64(1.1372004742734134) - rollout_corr/rollout_log_ppl:np.float64(0.12500254029873759) - rollout_corr/log_ppl_diff:np.float64(0.000924865496926941) - rollout_corr/log_ppl_abs_diff:np.float64(0.0019478512549540028) - rollout_corr/log_ppl_diff_max:np.float64(0.01921156421303749) - rollout_corr/log_ppl_diff_min:np.float64(-0.006299503147602081) - rollout_corr/ppl_ratio:np.float64(1.0009291872847825) - rollout_corr/chi2_token:np.float64(0.0008862421382218599) - rollout_corr/chi2_seq:np.float64(0.23664455325342715) - actor/pg_loss:np.float64(0.0004893041914328933) - actor/kl_loss:np.float64(0.0) - actor/pg_clipfrac:np.float64(8.747176980250515e-05) - actor/ppo_kl:np.float64(0.0004110109568475906) - actor/pg_clipfrac_lower:np.float64(0.0) - self_distillation/token_reweight_lambda:np.float64(0.5) - self_distillation/token_reweight_eps_w:np.float64(0.20000000000000004) - self_distillation/token_reweight_active_token_fraction:np.float64(0.625) - self_distillation/token_reweight_context_sample_fraction:np.float64(0.625) - self_distillation/token_reweight_correct_sample_fraction:np.float64(0.5859375) - self_distillation/token_reweight_w_mean:np.float64(3621.874775248114) - self_distillation/token_reweight_w_std:np.float64(62977.78215696255) - self_distillation/token_reweight_w_clipped_mean:np.float64(1.0000168199185282) - self_distillation/token_reweight_w_clip_frac:np.float64(0.007946641184389591) - self_distillation/empty_target_batch:np.float64(0.375) - actor/grad_norm:np.float64(0.12617958337068558) - perf/mfu/actor:np.float64(0.05071472735477538) - perf/max_memory_allocated_gb:np.float64(130.98325157165527) - perf/max_memory_reserved_gb:np.float64(140.86328125) - perf/cpu_memory_used_gb:np.float64(198.16430282592773) - actor/lr:np.float64(1e-06) - training/global_step:28 - training/epoch:1 - critic/score/mean:0.5859375 - critic/score/max:1.0 - critic/score/min:0.0 - critic/rewards/mean:0.5859375 - critic/rewards/max:1.0 - critic/rewards/min:0.0 - critic/advantages/mean:0.002507904777303338 - critic/advantages/max:0.625 - critic/advantages/min:-0.875 - critic/returns/mean:0.002507904777303338 - critic/returns/max:0.625 - critic/returns/min:-0.875 - response_length/mean:311.3203125 - response_length/max:64
|
||
step:29 - global_seqlen/min:18052 - global_seqlen/max:25734 - global_seqlen/minmax_diff:7682 - global_seqlen/balanced_min:22625 - global_seqlen/balanced_max:22629 - global_seqlen/mean:22627.125 - actor/entropy:0.0921759158372879 - perf/mfu/actor_infer:0 - training/rollout_probs_diff_valid:1 - training/rollout_probs_diff_max:0.2549144923686981 - training/rollout_probs_diff_mean:0.0022082990035414696 - training/rollout_probs_diff_std:0.009376613423228264 - training/rollout_actor_probs_pearson_corr:0.9981832504272461 - self_distillation/success_group_fraction:0.6875 - self_distillation/success_sample_fraction:0.6875 - self_distillation/correct_sample_fraction:0.6171875 - self_distillation/feedback_available_fraction:0.0 - self_distillation/feedback_used_fraction:0.0 - self_distillation/reprompt_sample_fraction:0.6875 - rollout_corr/rollout_is_mean:1.00003182888031 - rollout_corr/rollout_is_ratio_fraction_high:0.0 - rollout_corr/rollout_is_ratio_fraction_low:2.582611341495067e-05 - rollout_corr/rollout_is_max:1.6483628749847412 - rollout_corr/rollout_is_min:0.4751218259334564 - rollout_corr/rollout_is_std:0.02595939300954342 - rollout_corr/rollout_is_eff_sample_size:0.9993266827570269 - rollout_corr/rollout_is_seq_mean:1.0000574588775635 - rollout_corr/rollout_is_seq_std:0.0017196550033986568 - rollout_corr/rollout_is_seq_max:1.006497859954834 - rollout_corr/rollout_is_seq_min:0.9940721988677979 - rollout_corr/rollout_is_seq_max_deviation:0.006497859954833984 - rollout_corr/rollout_is_seq_fraction_high:0.0 - rollout_corr/rollout_is_seq_fraction_low:0.0 - rollout_corr/training_ppl:np.float64(1.0933433049358428) - rollout_corr/training_log_ppl:np.float64(0.08807740310294321) - rollout_corr/kl:np.float64(0.00027317932717929594) - rollout_corr/k3_kl:np.float64(0.0003994494008381366) - rollout_corr/rollout_ppl:np.float64(1.0930359880439937) - rollout_corr/rollout_log_ppl:np.float64(0.08780422273412114) - rollout_corr/log_ppl_diff:np.float64(0.000273180368822068) - rollout_corr/log_ppl_abs_diff:np.float64(0.0014128925540717319) - rollout_corr/log_ppl_diff_max:np.float64(0.00730479508638382) - rollout_corr/log_ppl_diff_min:np.float64(-0.0051046013832092285) - rollout_corr/ppl_ratio:np.float64(1.0002750388812274) - rollout_corr/chi2_token:np.float64(0.0010566439013928175) - rollout_corr/chi2_seq:np.float64(0.3449188000522554) - actor/pg_loss:np.float64(0.0009986156364902854) - actor/kl_loss:np.float64(0.0) - actor/pg_clipfrac:np.float64(0.00014245773490983993) - actor/ppo_kl:np.float64(-1.469018872057859e-05) - actor/pg_clipfrac_lower:np.float64(0.0) - self_distillation/token_reweight_lambda:np.float64(0.5) - self_distillation/token_reweight_eps_w:np.float64(0.20000000000000004) - self_distillation/token_reweight_active_token_fraction:np.float64(0.6875) - self_distillation/token_reweight_context_sample_fraction:np.float64(0.6875) - self_distillation/token_reweight_correct_sample_fraction:np.float64(0.6171875) - self_distillation/token_reweight_w_mean:np.float64(23437.328467473853) - self_distillation/token_reweight_w_std:np.float64(427153.70655143255) - self_distillation/token_reweight_w_clipped_mean:np.float64(0.9993212011177093) - self_distillation/token_reweight_w_clip_frac:np.float64(0.02301981179334689) - self_distillation/empty_target_batch:np.float64(0.3125) - actor/grad_norm:np.float64(0.1457035057246685) - perf/mfu/actor:np.float64(0.05253759434802585) - perf/max_memory_allocated_gb:np.float64(130.98325157165527) - perf/max_memory_reserved_gb:np.float64(140.86328125) - perf/cpu_memory_used_gb:np.float64(198.28145599365234) - actor/lr:np.float64(1e-06) - training/global_step:29 - training/epoch:1 - critic/score/mean:0.6171875 - critic/score/max:1.0 - critic/score/min:0.0 - critic/rewards/mean:0.6171875 - critic/rewards/max:1.0 - critic/rewards/min:0.0 - critic/advantages/mean:-0.0004229026089888066 - critic/advantages/max:0.625 - critic/advantages/min:-0.875 - critic/returns/mean:-0.0004229026089888066 - critic/returns/max:0.625 - critic/returns/min:-0.875 - response_length/mean:302.50390625 - response_length/max:568.0 - res
|
||
test_gen_batch meta info: {'eos_token_id': 151645, 'pad_token_id': 151643, 'recompute_log_prob': False, 'do_sample': True, 'validate': True, 'global_steps': 30}
|
||
validation generation end
|
||
[prompt] system
|
||
|
||
Given a question and four options, please select the right answer. Respond in the following format:
|
||
<reasoning>
|
||
...
|
||
</reasoning>
|
||
<answer>
|
||
...
|
||
</answer>
|
||
|
||
For the answer, only output the letter corresponding to the correct option (A, B, C, or D), and nothing else. Do not restate the answer text. For example, if the answer is "A", just output:
|
||
<answer>
|
||
A
|
||
</answer>
|
||
|
||
user
|
||
Given the following crystal structure parameters for the material RbLa9(IrO6)4 (Material ID: mp-560657), calculate the volume of the unit cell (in ų).
|
||
|
||
Lattice Parameters:
|
||
|
||
a = 7.82 Å
|
||
b = 7.82 Å
|
||
c = 17.88 Å
|
||
α = 90.00º
|
||
β = 90.00º
|
||
γ = 90.00º
|
||
|
||
|
||
|
||
A: 1025.67 ų
|
||
|
||
B: 1094.31 ų
|
||
|
||
C: 1200.45 ų
|
||
|
||
D: 1150.78 ų
|
||
|
||
Please reason step by step.
|
||
assistant
|
||
<think>
|
||
|
||
</think>
|
||
|
||
|
||
[response] <reasoning>
|
||
The given crystal structure has lattice parameters a = 7.82 Å, b = 7.82 Å, c = 17.88 Å, and all angles (α, β, γ) are 90.00º, indicating a orthorhombic unit cell.
|
||
|
||
The volume $ V $ of a unit cell in an orthorhombic system is calculated using the formula:
|
||
|
||
$$
|
||
V = a \times b \times c
|
||
$$
|
||
|
||
Substituting the given values:
|
||
|
||
$$
|
||
V = 7.82 \times 7.82 \times 17.88
|
||
$$
|
||
|
||
First, calculate $ 7.82 \times 7.82 $:
|
||
|
||
$$
|
||
7.82 \times 7.82 = 61.1524
|
||
$$
|
||
|
||
Then multiply by 17.88:
|
||
|
||
$$
|
||
61.1524 \times 17.88 = 1094.31
|
||
$$
|
||
|
||
Thus, the volume of the unit cell is 1094.31 ų.
|
||
</reasoning>
|
||
<answer>
|
||
B
|
||
</answer>
|
||
[ground_truth] B
|
||
[score] 1.0
|
||
[acc] 1.0
|
||
[pred] B
|
||
[incorrect_format] 1
|
||
[feedback]
|
||
local_global_step_folder: /mnt/mole/SDPO/L2T/checkpoints/datasets/sciknoweval/material/qwen3gen-material-RLSD_TR-Qwen-Qwen3-8B-mbs8-decay0-tr0.1-train32-rollout8-lr1e-6-vllm0.8/global_step_30
|
||
step:30 - global_seqlen/min:20198 - global_seqlen/max:26628 - global_seqlen/minmax_diff:6430 - global_seqlen/balanced_min:23735 - global_seqlen/balanced_max:23742 - global_seqlen/mean:23739.625 - actor/entropy:0.10219845175743103 - perf/mfu/actor_infer:0 - training/rollout_probs_diff_valid:1 - training/rollout_probs_diff_max:0.6695281267166138 - training/rollout_probs_diff_mean:0.0023985919542610645 - training/rollout_probs_diff_std:0.010449547320604324 - training/rollout_actor_probs_pearson_corr:0.9980133175849915 - self_distillation/success_group_fraction:0.6875 - self_distillation/success_sample_fraction:0.67578125 - self_distillation/correct_sample_fraction:0.53125 - self_distillation/feedback_available_fraction:0.0 - self_distillation/feedback_used_fraction:0.0 - self_distillation/reprompt_sample_fraction:0.67578125 - rollout_corr/rollout_is_mean:1.0001095533370972 - rollout_corr/rollout_is_ratio_fraction_high:1.1889334018633235e-05 - rollout_corr/rollout_is_ratio_fraction_low:4.755733607453294e-05 - rollout_corr/rollout_is_max:5.522396087646484 - rollout_corr/rollout_is_min:0.21412654221057892 - rollout_corr/rollout_is_std:0.027441667392849922 - rollout_corr/rollout_is_eff_sample_size:0.9992476405980566 - rollout_corr/rollout_is_seq_mean:1.0000834465026855 - rollout_corr/rollout_is_seq_std:0.0018353002378717065 - rollout_corr/rollout_is_seq_max:1.0131018161773682 - rollout_corr/rollout_is_seq_min:0.994994580745697 - rollout_corr/rollout_is_seq_max_deviation:0.013101816177368164 - rollout_corr/rollout_is_seq_fraction_high:0.0 - rollout_corr/rollout_is_seq_fraction_low:0.0 - rollout_corr/training_ppl:np.float64(1.108945808839053) - rollout_corr/training_log_ppl:np.float64(0.10102521417138632) - rollout_corr/kl:np.float64(0.0006417104669356188) - rollout_corr/k3_kl:np.float64(0.0007572973181737552) - rollout_corr/rollout_ppl:np.float64(1.1082097035832703) - rollout_corr/rollout_log_ppl:np.float64(0.10038350410468411) - rollout_corr/log_ppl_diff:np.float64(0.0006417100667022169) - rollout_corr/log_ppl_abs_diff:np.float64(0.0018683964008232579) - rollout_corr/log_ppl_diff_max:np.float64(0.014286383986473083) - rollout_corr/log_ppl_diff_min:np.float64(-0.00767090916633606) - rollout_corr/ppl_ratio:np.float64(1.0006450458895415) - rollout_corr/chi2_token:np.float64(0.0018558201845735312) - rollout_corr/chi2_seq:np.float64(0.8411497273482382) - actor/pg_loss:np.float64(0.0017660357989370823) - actor/kl_loss:np.float64(0.0) - actor/pg_clipfrac:np.float64(0.0006848026164334442) - actor/ppo_kl:np.float64(0.00030027530577569905) - actor/pg_clipfrac_lower:np.float64(0.0) - self_distillation/token_reweight_lambda:np.float64(0.5) - self_distillation/token_reweight_eps_w:np.float64(0.20000000000000004) - self_distillation/token_reweight_active_token_fraction:np.float64(0.67578125) - self_distillation/token_reweight_context_sample_fraction:np.float64(0.67578125) - self_distillation/token_reweight_correct_sample_fraction:np.float64(0.53125) - self_distillation/token_reweight_w_mean:np.float64(22148.840008513536) - self_distillation/token_reweight_w_std:np.float64(361685.4959547251) - self_distillation/token_reweight_w_clipped_mean:np.float64(1.0007154732011259) - self_distillation/token_reweight_w_clip_frac:np.float64(0.029575713473604992) - self_distillation/empty_target_batch:np.float64(0.32421875) - actor/grad_norm:np.float64(0.22181391529738903) - perf/mfu/actor:np.float64(0.05468114347361665) - perf/max_memory_allocated_gb:np.float64(130.98325157165527) - perf/max_memory_reserved_gb:np.float64(140.86328125) - perf/cpu_memory_used_gb:np.float64(198.37141799926758) - actor/lr:np.float64(1e-06) - val-aux/sciknoweval/reward/mean@16:np.float64(0.6974734042553191) - val-aux/sciknoweval/reward/std@16:np.float64(0.10297707501391735) - val-aux/sciknoweval/reward/best@2/mean:np.float64(0.7381914893617021) - val-aux/sciknoweval/reward/best@2/std:np.float64(0.07451746654780546) - val-aux/sciknoweval/reward/worst@2/mean:np.float64(0.6570531914893617) - val-aux/sciknoweval/reward/worst@2/std:np.float64(0.0985696922009786) - val-aux/sciknowev
|
||
step:31 - global_seqlen/min:20844 - global_seqlen/max:25023 - global_seqlen/minmax_diff:4179 - global_seqlen/balanced_min:22623 - global_seqlen/balanced_max:22625 - global_seqlen/mean:22624.125 - actor/entropy:0.10231570899486542 - perf/mfu/actor_infer:0 - training/rollout_probs_diff_valid:1 - training/rollout_probs_diff_max:0.33947548270225525 - training/rollout_probs_diff_mean:0.002441213233396411 - training/rollout_probs_diff_std:0.00997840240597725 - training/rollout_actor_probs_pearson_corr:0.9981839060783386 - self_distillation/success_group_fraction:0.75 - self_distillation/success_sample_fraction:0.75 - self_distillation/correct_sample_fraction:0.65234375 - self_distillation/feedback_available_fraction:0.0 - self_distillation/feedback_used_fraction:0.0 - self_distillation/reprompt_sample_fraction:0.75 - rollout_corr/rollout_is_mean:0.9999410510063171 - rollout_corr/rollout_is_ratio_fraction_high:3.758974708034657e-05 - rollout_corr/rollout_is_ratio_fraction_low:1.2529914783954155e-05 - rollout_corr/rollout_is_max:3.3646881580352783 - rollout_corr/rollout_is_min:0.4018668532371521 - rollout_corr/rollout_is_std:0.028051747009158134 - rollout_corr/rollout_is_eff_sample_size:0.9992135991833438 - rollout_corr/rollout_is_seq_mean:0.9999500513076782 - rollout_corr/rollout_is_seq_std:0.00178877345751971 - rollout_corr/rollout_is_seq_max:1.0059913396835327 - rollout_corr/rollout_is_seq_min:0.9913996458053589 - rollout_corr/rollout_is_seq_max_deviation:0.008600354194641113 - rollout_corr/rollout_is_seq_fraction_high:0.0 - rollout_corr/rollout_is_seq_fraction_low:0.0 - rollout_corr/training_ppl:np.float64(1.1037001409567893) - rollout_corr/training_log_ppl:np.float64(0.09721801686100662) - rollout_corr/kl:np.float64(0.0006632166562940967) - rollout_corr/k3_kl:np.float64(0.0005725183017233348) - rollout_corr/rollout_ppl:np.float64(1.102965306956321) - rollout_corr/rollout_log_ppl:np.float64(0.09655479896900943) - rollout_corr/log_ppl_diff:np.float64(0.0006632178919971921) - rollout_corr/log_ppl_abs_diff:np.float64(0.0015754591222503223) - rollout_corr/log_ppl_diff_max:np.float64(0.010033667087554932) - rollout_corr/log_ppl_diff_min:np.float64(-0.007137298583984375) - rollout_corr/ppl_ratio:np.float64(1.0006655387114733) - rollout_corr/chi2_token:np.float64(0.0010873002465814352) - rollout_corr/chi2_seq:np.float64(0.2302088625729084) - actor/pg_loss:np.float64(0.001406671479344368) - actor/kl_loss:np.float64(0.0) - actor/pg_clipfrac:np.float64(0.0002917049291681906) - actor/ppo_kl:np.float64(0.00019202986122568433) - actor/pg_clipfrac_lower:np.float64(0.0) - self_distillation/token_reweight_lambda:np.float64(0.5) - self_distillation/token_reweight_eps_w:np.float64(0.20000000000000004) - self_distillation/token_reweight_active_token_fraction:np.float64(0.75) - self_distillation/token_reweight_context_sample_fraction:np.float64(0.75) - self_distillation/token_reweight_correct_sample_fraction:np.float64(0.65234375) - self_distillation/token_reweight_w_mean:np.float64(105553.12214704277) - self_distillation/token_reweight_w_std:np.float64(1536630.4769199342) - self_distillation/token_reweight_w_clipped_mean:np.float64(0.9998775997664779) - self_distillation/token_reweight_w_clip_frac:np.float64(0.023525838390924037) - self_distillation/empty_target_batch:np.float64(0.25) - actor/grad_norm:np.float64(0.20417143404483795) - perf/mfu/actor:np.float64(0.05114181998467157) - perf/max_memory_allocated_gb:np.float64(130.98325157165527) - perf/max_memory_reserved_gb:np.float64(140.86328125) - perf/cpu_memory_used_gb:np.float64(196.02905082702637) - actor/lr:np.float64(1e-06) - training/global_step:31 - training/epoch:1 - critic/score/mean:0.65234375 - critic/score/max:1.0 - critic/score/min:0.0 - critic/rewards/mean:0.65234375 - critic/rewards/max:1.0 - critic/rewards/min:0.0 - critic/advantages/mean:0.0034034382551908493 - critic/advantages/max:0.75 - critic/advantages/min:-0.875 - critic/returns/mean:0.0034034382551908493 - critic/returns/max:0.75 - critic/returns/min:-0.875 - response_length/mean:311.75390625 - response_length/max:716
|
||
step:32 - global_seqlen/min:17465 - global_seqlen/max:26557 - global_seqlen/minmax_diff:9092 - global_seqlen/balanced_min:22365 - global_seqlen/balanced_max:22370 - global_seqlen/mean:22367.75 - actor/entropy:0.10624657571315765 - perf/mfu/actor_infer:0 - training/rollout_probs_diff_valid:1 - training/rollout_probs_diff_max:0.24491804838180542 - training/rollout_probs_diff_mean:0.0025010062381625175 - training/rollout_probs_diff_std:0.010108744725584984 - training/rollout_actor_probs_pearson_corr:0.9982028007507324 - self_distillation/success_group_fraction:0.71875 - self_distillation/success_sample_fraction:0.71875 - self_distillation/correct_sample_fraction:0.65625 - self_distillation/feedback_available_fraction:0.0 - self_distillation/feedback_used_fraction:0.0 - self_distillation/reprompt_sample_fraction:0.71875 - rollout_corr/rollout_is_mean:1.0000066757202148 - rollout_corr/rollout_is_ratio_fraction_high:0.0 - rollout_corr/rollout_is_ratio_fraction_low:0.0 - rollout_corr/rollout_is_max:1.8165538311004639 - rollout_corr/rollout_is_min:0.5453736186027527 - rollout_corr/rollout_is_std:0.02843792922794819 - rollout_corr/rollout_is_eff_sample_size:0.9991918186557794 - rollout_corr/rollout_is_seq_mean:0.9999681711196899 - rollout_corr/rollout_is_seq_std:0.0018278236966580153 - rollout_corr/rollout_is_seq_max:1.0054386854171753 - rollout_corr/rollout_is_seq_min:0.9935697317123413 - rollout_corr/rollout_is_seq_max_deviation:0.006430268287658691 - rollout_corr/rollout_is_seq_fraction_high:0.0 - rollout_corr/rollout_is_seq_fraction_low:0.0 - rollout_corr/training_ppl:np.float64(1.1141844862140715) - rollout_corr/training_log_ppl:np.float64(0.10589008548413403) - rollout_corr/kl:np.float64(0.0006955666955104078) - rollout_corr/k3_kl:np.float64(0.0005526502317962922) - rollout_corr/rollout_ppl:np.float64(1.1134190135635436) - rollout_corr/rollout_log_ppl:np.float64(0.1051945186118246) - rollout_corr/log_ppl_diff:np.float64(0.0006955668723094277) - rollout_corr/log_ppl_abs_diff:np.float64(0.0016634032290312462) - rollout_corr/log_ppl_diff_max:np.float64(0.006527021527290344) - rollout_corr/log_ppl_diff_min:np.float64(-0.0044723451137542725) - rollout_corr/ppl_ratio:np.float64(1.0006979582831264) - rollout_corr/chi2_token:np.float64(0.0009256161283701658) - rollout_corr/chi2_seq:np.float64(0.372445534914732) - actor/pg_loss:np.float64(0.0014663469046354294) - actor/kl_loss:np.float64(0.0) - actor/pg_clipfrac:np.float64(0.00011217115661565913) - actor/ppo_kl:np.float64(0.00025597080171557707) - actor/pg_clipfrac_lower:np.float64(0.0) - self_distillation/token_reweight_lambda:np.float64(0.5) - self_distillation/token_reweight_eps_w:np.float64(0.20000000000000004) - self_distillation/token_reweight_active_token_fraction:np.float64(0.71875) - self_distillation/token_reweight_context_sample_fraction:np.float64(0.71875) - self_distillation/token_reweight_correct_sample_fraction:np.float64(0.65625) - self_distillation/token_reweight_w_mean:np.float64(12649.956588850822) - self_distillation/token_reweight_w_std:np.float64(114389.63001844386) - self_distillation/token_reweight_w_clipped_mean:np.float64(1.0000564639922231) - self_distillation/token_reweight_w_clip_frac:np.float64(0.019761866453336552) - self_distillation/empty_target_batch:np.float64(0.28125) - actor/grad_norm:np.float64(0.17176248785108328) - perf/mfu/actor:np.float64(0.050811027766343705) - perf/max_memory_allocated_gb:np.float64(130.98325157165527) - perf/max_memory_reserved_gb:np.float64(140.86328125) - perf/cpu_memory_used_gb:np.float64(195.6191520690918) - actor/lr:np.float64(1e-06) - training/global_step:32 - training/epoch:1 - critic/score/mean:0.65625 - critic/score/max:1.0 - critic/score/min:0.0 - critic/rewards/mean:0.65625 - critic/rewards/max:1.0 - critic/rewards/min:0.0 - critic/advantages/mean:0.0010969723807647824 - critic/advantages/max:0.75 - critic/advantages/min:-0.875 - critic/returns/mean:0.0010969723807647824 - critic/returns/max:0.75 - critic/returns/min:-0.875 - response_length/mean:302.6796875 - response_length/max:620.0 - response_length/min:72.
|
||
step:33 - global_seqlen/min:17970 - global_seqlen/max:23726 - global_seqlen/minmax_diff:5756 - global_seqlen/balanced_min:21881 - global_seqlen/balanced_max:21884 - global_seqlen/mean:21882.25 - actor/entropy:0.11248673498630524 - perf/mfu/actor_infer:0 - training/rollout_probs_diff_valid:1 - training/rollout_probs_diff_max:0.37974897027015686 - training/rollout_probs_diff_mean:0.002393264090642333 - training/rollout_probs_diff_std:0.00961360428482294 - training/rollout_actor_probs_pearson_corr:0.998481810092926 - self_distillation/success_group_fraction:0.84375 - self_distillation/success_sample_fraction:0.83984375 - self_distillation/correct_sample_fraction:0.7109375 - self_distillation/feedback_available_fraction:0.0 - self_distillation/feedback_used_fraction:0.0 - self_distillation/reprompt_sample_fraction:0.83984375 - rollout_corr/rollout_is_mean:0.9996835589408875 - rollout_corr/rollout_is_ratio_fraction_high:1.3081813449389301e-05 - rollout_corr/rollout_is_ratio_fraction_low:3.924544216715731e-05 - rollout_corr/rollout_is_max:2.4000132083892822 - rollout_corr/rollout_is_min:0.3203427791595459 - rollout_corr/rollout_is_std:0.026572102680802345 - rollout_corr/rollout_is_eff_sample_size:0.9992938263651499 - rollout_corr/rollout_is_seq_mean:0.9996833801269531 - rollout_corr/rollout_is_seq_std:0.0017737209564074874 - rollout_corr/rollout_is_seq_max:1.0062614679336548 - rollout_corr/rollout_is_seq_min:0.9934943318367004 - rollout_corr/rollout_is_seq_max_deviation:0.0065056681632995605 - rollout_corr/rollout_is_seq_fraction_high:0.0 - rollout_corr/rollout_is_seq_fraction_low:0.0 - rollout_corr/training_ppl:np.float64(1.1257116086781025) - rollout_corr/training_log_ppl:np.float64(0.1153708925266983) - rollout_corr/kl:np.float64(0.0009000642832432959) - rollout_corr/k3_kl:np.float64(0.0006722159668157701) - rollout_corr/rollout_ppl:np.float64(1.1247029760852456) - rollout_corr/rollout_log_ppl:np.float64(0.1144708267529495) - rollout_corr/log_ppl_diff:np.float64(0.0009000657737487927) - rollout_corr/log_ppl_abs_diff:np.float64(0.0018840501870727167) - rollout_corr/log_ppl_diff_max:np.float64(0.01474740356206894) - rollout_corr/log_ppl_diff_min:np.float64(-0.008892104029655457) - rollout_corr/ppl_ratio:np.float64(1.000903464620933) - rollout_corr/chi2_token:np.float64(0.0012851853389292955) - rollout_corr/chi2_seq:np.float64(0.42543286248110235) - actor/pg_loss:np.float64(0.0017157087568193674) - actor/kl_loss:np.float64(0.0) - actor/pg_clipfrac:np.float64(0.0005343864218048111) - actor/ppo_kl:np.float64(0.00020211363660394) - actor/pg_clipfrac_lower:np.float64(0.0) - self_distillation/token_reweight_lambda:np.float64(0.5) - self_distillation/token_reweight_eps_w:np.float64(0.20000000000000004) - self_distillation/token_reweight_active_token_fraction:np.float64(0.83984375) - self_distillation/token_reweight_context_sample_fraction:np.float64(0.83984375) - self_distillation/token_reweight_correct_sample_fraction:np.float64(0.7109375) - self_distillation/token_reweight_w_mean:np.float64(61808.9106729636) - self_distillation/token_reweight_w_std:np.float64(852764.9511296176) - self_distillation/token_reweight_w_clipped_mean:np.float64(1.0005815126933157) - self_distillation/token_reweight_w_clip_frac:np.float64(0.039750390249537304) - self_distillation/empty_target_batch:np.float64(0.16015625) - actor/grad_norm:np.float64(0.23709694296121597) - perf/mfu/actor:np.float64(0.049707760742796335) - perf/max_memory_allocated_gb:np.float64(130.98325157165527) - perf/max_memory_reserved_gb:np.float64(140.86328125) - perf/cpu_memory_used_gb:np.float64(195.51437759399414) - actor/lr:np.float64(1e-06) - training/global_step:33 - training/epoch:1 - critic/score/mean:0.7109375 - critic/score/max:1.0 - critic/score/min:0.0 - critic/rewards/mean:0.7109375 - critic/rewards/max:1.0 - critic/rewards/min:0.0 - critic/advantages/mean:0.0004366055363789201 - critic/advantages/max:0.875 - critic/advantages/min:-0.875 - critic/returns/mean:0.0004366055363789201 - critic/returns/max:0.875 - critic/returns/min:-0.875 - response_length/mean:298.6015625
|
||
step:34 - global_seqlen/min:21088 - global_seqlen/max:25454 - global_seqlen/minmax_diff:4366 - global_seqlen/balanced_min:23192 - global_seqlen/balanced_max:23193 - global_seqlen/mean:23192.5 - actor/entropy:0.12696923315525055 - perf/mfu/actor_infer:0 - training/rollout_probs_diff_valid:1 - training/rollout_probs_diff_max:0.317455530166626 - training/rollout_probs_diff_mean:0.0027052101213485003 - training/rollout_probs_diff_std:0.009988086298108101 - training/rollout_actor_probs_pearson_corr:0.9985281229019165 - self_distillation/success_group_fraction:0.90625 - self_distillation/success_sample_fraction:0.90625 - self_distillation/correct_sample_fraction:0.80859375 - self_distillation/feedback_available_fraction:0.0 - self_distillation/feedback_used_fraction:0.0 - self_distillation/reprompt_sample_fraction:0.90625 - rollout_corr/rollout_is_mean:1.0001001358032227 - rollout_corr/rollout_is_ratio_fraction_high:2.5154700779239647e-05 - rollout_corr/rollout_is_ratio_fraction_low:0.0 - rollout_corr/rollout_is_max:2.463893175125122 - rollout_corr/rollout_is_min:0.5469041466712952 - rollout_corr/rollout_is_std:0.02853209152817726 - rollout_corr/rollout_is_eff_sample_size:0.9991869389956493 - rollout_corr/rollout_is_seq_mean:1.0001479387283325 - rollout_corr/rollout_is_seq_std:0.0017741305055096745 - rollout_corr/rollout_is_seq_max:1.0061523914337158 - rollout_corr/rollout_is_seq_min:0.993314802646637 - rollout_corr/rollout_is_seq_max_deviation:0.006685197353363037 - rollout_corr/rollout_is_seq_fraction_high:0.0 - rollout_corr/rollout_is_seq_fraction_low:0.0 - rollout_corr/training_ppl:np.float64(1.1375689092092216) - rollout_corr/training_log_ppl:np.float64(0.1276737543521449) - rollout_corr/kl:np.float64(0.0003955075519228579) - rollout_corr/k3_kl:np.float64(0.0004891105344881908) - rollout_corr/rollout_ppl:np.float64(1.137121386360377) - rollout_corr/rollout_log_ppl:np.float64(0.1272782460291637) - rollout_corr/log_ppl_diff:np.float64(0.0003955083229811862) - rollout_corr/log_ppl_abs_diff:np.float64(0.0014175720716593787) - rollout_corr/log_ppl_diff_max:np.float64(0.007905177772045135) - rollout_corr/log_ppl_diff_min:np.float64(-0.006664559245109558) - rollout_corr/ppl_ratio:np.float64(1.000397405354306) - rollout_corr/chi2_token:np.float64(0.0012063093017786741) - rollout_corr/chi2_seq:np.float64(0.39643845544196665) - actor/pg_loss:np.float64(0.0009817357640713453) - actor/kl_loss:np.float64(0.0) - actor/pg_clipfrac:np.float64(0.00016826897945065866) - actor/ppo_kl:np.float64(0.0001194505678521196) - actor/pg_clipfrac_lower:np.float64(0.0) - self_distillation/token_reweight_lambda:np.float64(0.5) - self_distillation/token_reweight_eps_w:np.float64(0.20000000000000004) - self_distillation/token_reweight_active_token_fraction:np.float64(0.90625) - self_distillation/token_reweight_context_sample_fraction:np.float64(0.90625) - self_distillation/token_reweight_correct_sample_fraction:np.float64(0.80859375) - self_distillation/token_reweight_w_mean:np.float64(43711.79978899378) - self_distillation/token_reweight_w_std:np.float64(769875.5125299292) - self_distillation/token_reweight_w_clipped_mean:np.float64(1.000270399497822) - self_distillation/token_reweight_w_clip_frac:np.float64(0.02182861461187713) - self_distillation/empty_target_batch:np.float64(0.09375) - actor/grad_norm:np.float64(0.18393283523619175) - perf/mfu/actor:np.float64(0.05226619083988275) - perf/max_memory_allocated_gb:np.float64(130.98325157165527) - perf/max_memory_reserved_gb:np.float64(140.86328125) - perf/cpu_memory_used_gb:np.float64(195.52132034301758) - actor/lr:np.float64(1e-06) - training/global_step:34 - training/epoch:1 - critic/score/mean:0.80859375 - critic/score/max:1.0 - critic/score/min:0.0 - critic/rewards/mean:0.80859375 - critic/rewards/max:1.0 - critic/rewards/min:0.0 - critic/advantages/mean:0.0007043316145427525 - critic/advantages/max:0.75 - critic/advantages/min:-0.875 - critic/returns/mean:0.0007043316145427525 - critic/returns/max:0.75 - critic/returns/min:-0.875 - response_length/mean:310.578125 - response_length/max:597.0 - respo
|
||
step:35 - global_seqlen/min:19976 - global_seqlen/max:24551 - global_seqlen/minmax_diff:4575 - global_seqlen/balanced_min:21273 - global_seqlen/balanced_max:21278 - global_seqlen/mean:21276.125 - actor/entropy:0.1320156753063202 - perf/mfu/actor_infer:0 - training/rollout_probs_diff_valid:1 - training/rollout_probs_diff_max:0.36655372381210327 - training/rollout_probs_diff_mean:0.002745934296399355 - training/rollout_probs_diff_std:0.010113532654941082 - training/rollout_actor_probs_pearson_corr:0.9984952807426453 - self_distillation/success_group_fraction:0.875 - self_distillation/success_sample_fraction:0.875 - self_distillation/correct_sample_fraction:0.81640625 - self_distillation/feedback_available_fraction:0.0 - self_distillation/feedback_used_fraction:0.0 - self_distillation/reprompt_sample_fraction:0.875 - rollout_corr/rollout_is_mean:0.9999025464057922 - rollout_corr/rollout_is_ratio_fraction_high:0.0 - rollout_corr/rollout_is_ratio_fraction_low:5.574679744313471e-05 - rollout_corr/rollout_is_max:1.6714837551116943 - rollout_corr/rollout_is_min:0.029101360589265823 - rollout_corr/rollout_is_std:0.02873193845152855 - rollout_corr/rollout_is_eff_sample_size:0.999174918572789 - rollout_corr/rollout_is_seq_mean:0.999925971031189 - rollout_corr/rollout_is_seq_std:0.0018350593745708466 - rollout_corr/rollout_is_seq_max:1.0064617395401 - rollout_corr/rollout_is_seq_min:0.9949533343315125 - rollout_corr/rollout_is_seq_max_deviation:0.006461739540100098 - rollout_corr/rollout_is_seq_fraction_high:0.0 - rollout_corr/rollout_is_seq_fraction_low:0.0 - rollout_corr/training_ppl:np.float64(1.140158954076469) - rollout_corr/training_log_ppl:np.float64(0.12918748372612754) - rollout_corr/kl:np.float64(0.0005167052121564097) - rollout_corr/k3_kl:np.float64(0.0005370462090237282) - rollout_corr/rollout_ppl:np.float64(1.1395544721744955) - rollout_corr/rollout_log_ppl:np.float64(0.1286707783001475) - rollout_corr/log_ppl_diff:np.float64(0.0005167054259800352) - rollout_corr/log_ppl_abs_diff:np.float64(0.0016419052044511773) - rollout_corr/log_ppl_diff_max:np.float64(0.011951439082622528) - rollout_corr/log_ppl_diff_min:np.float64(-0.00703805685043335) - rollout_corr/ppl_ratio:np.float64(1.0005191075615585) - rollout_corr/chi2_token:np.float64(0.001026106532663107) - rollout_corr/chi2_seq:np.float64(0.30341047793626785) - actor/pg_loss:np.float64(0.0011919063981622458) - actor/kl_loss:np.float64(0.0) - actor/pg_clipfrac:np.float64(0.0002517128054932982) - actor/ppo_kl:np.float64(-4.123877571160506e-05) - actor/pg_clipfrac_lower:np.float64(0.0) - self_distillation/token_reweight_lambda:np.float64(0.5) - self_distillation/token_reweight_eps_w:np.float64(0.20000000000000004) - self_distillation/token_reweight_active_token_fraction:np.float64(0.875) - self_distillation/token_reweight_context_sample_fraction:np.float64(0.875) - self_distillation/token_reweight_correct_sample_fraction:np.float64(0.81640625) - self_distillation/token_reweight_w_mean:np.float64(22245.829055046663) - self_distillation/token_reweight_w_std:np.float64(408636.0869032652) - self_distillation/token_reweight_w_clipped_mean:np.float64(0.9995780531316996) - self_distillation/token_reweight_w_clip_frac:np.float64(0.02098444675357314) - self_distillation/empty_target_batch:np.float64(0.125) - actor/grad_norm:np.float64(0.1606460027396679) - perf/mfu/actor:np.float64(0.04926837037338072) - perf/max_memory_allocated_gb:np.float64(130.98325157165527) - perf/max_memory_reserved_gb:np.float64(140.86328125) - perf/cpu_memory_used_gb:np.float64(195.49019145965576) - actor/lr:np.float64(1e-06) - training/global_step:35 - training/epoch:1 - critic/score/mean:0.81640625 - critic/score/max:1.0 - critic/score/min:0.0 - critic/rewards/mean:0.81640625 - critic/rewards/max:1.0 - critic/rewards/min:0.0 - critic/advantages/mean:-0.000839686137624085 - critic/advantages/max:0.625 - critic/advantages/min:-0.875 - critic/returns/mean:-0.000839686137624085 - critic/returns/max:0.625 - critic/returns/min:-0.875 - response_length/mean:280.28515625 - response_length/max:596.0 - response_len
|
||
step:36 - global_seqlen/min:20282 - global_seqlen/max:25914 - global_seqlen/minmax_diff:5632 - global_seqlen/balanced_min:22874 - global_seqlen/balanced_max:22876 - global_seqlen/mean:22874.875 - actor/entropy:0.1515883058309555 - perf/mfu/actor_infer:0 - training/rollout_probs_diff_valid:1 - training/rollout_probs_diff_max:0.317455530166626 - training/rollout_probs_diff_mean:0.002942807972431183 - training/rollout_probs_diff_std:0.010348225012421608 - training/rollout_actor_probs_pearson_corr:0.9985588192939758 - self_distillation/success_group_fraction:0.6875 - self_distillation/success_sample_fraction:0.68359375 - self_distillation/correct_sample_fraction:0.61328125 - self_distillation/feedback_available_fraction:0.0 - self_distillation/feedback_used_fraction:0.0 - self_distillation/reprompt_sample_fraction:0.68359375 - rollout_corr/rollout_is_mean:1.0001130104064941 - rollout_corr/rollout_is_ratio_fraction_high:0.0 - rollout_corr/rollout_is_ratio_fraction_low:0.0 - rollout_corr/rollout_is_max:1.585033655166626 - rollout_corr/rollout_is_min:0.6092519164085388 - rollout_corr/rollout_is_std:0.029064016416668892 - rollout_corr/rollout_is_eff_sample_size:0.9991562339352135 - rollout_corr/rollout_is_seq_mean:1.0001581907272339 - rollout_corr/rollout_is_seq_std:0.0017724381759762764 - rollout_corr/rollout_is_seq_max:1.0063711404800415 - rollout_corr/rollout_is_seq_min:0.9951498508453369 - rollout_corr/rollout_is_seq_max_deviation:0.006371140480041504 - rollout_corr/rollout_is_seq_fraction_high:0.0 - rollout_corr/rollout_is_seq_fraction_low:0.0 - rollout_corr/training_ppl:np.float64(1.1587157743051648) - rollout_corr/training_log_ppl:np.float64(0.14470403948507737) - rollout_corr/kl:np.float64(0.0005400172712077822) - rollout_corr/k3_kl:np.float64(0.0005347190636655341) - rollout_corr/rollout_ppl:np.float64(1.1580887311138213) - rollout_corr/rollout_log_ppl:np.float64(0.14416402138158446) - rollout_corr/log_ppl_diff:np.float64(0.0005400181034929119) - rollout_corr/log_ppl_abs_diff:np.float64(0.0015646610627300106) - rollout_corr/log_ppl_diff_max:np.float64(0.00749245285987854) - rollout_corr/log_ppl_diff_min:np.float64(-0.005699381232261658) - rollout_corr/ppl_ratio:np.float64(1.0005421619862318) - rollout_corr/chi2_token:np.float64(0.0010581978131085634) - rollout_corr/chi2_seq:np.float64(0.3867321067955345) - actor/pg_loss:np.float64(0.0007704137824475765) - actor/kl_loss:np.float64(0.0) - actor/pg_clipfrac:np.float64(0.0002766783450169896) - actor/ppo_kl:np.float64(0.00026795737259011077) - actor/pg_clipfrac_lower:np.float64(0.0) - self_distillation/token_reweight_lambda:np.float64(0.5) - self_distillation/token_reweight_eps_w:np.float64(0.20000000000000004) - self_distillation/token_reweight_active_token_fraction:np.float64(0.68359375) - self_distillation/token_reweight_context_sample_fraction:np.float64(0.68359375) - self_distillation/token_reweight_correct_sample_fraction:np.float64(0.61328125) - self_distillation/token_reweight_w_mean:np.float64(13437.756535453489) - self_distillation/token_reweight_w_std:np.float64(253038.20973212097) - self_distillation/token_reweight_w_clipped_mean:np.float64(1.0000933702103794) - self_distillation/token_reweight_w_clip_frac:np.float64(0.022569283057237044) - self_distillation/empty_target_batch:np.float64(0.31640625) - actor/grad_norm:np.float64(0.13522806763648987) - perf/mfu/actor:np.float64(0.05244350556818289) - perf/max_memory_allocated_gb:np.float64(130.98325157165527) - perf/max_memory_reserved_gb:np.float64(140.86328125) - perf/cpu_memory_used_gb:np.float64(195.4292984008789) - actor/lr:np.float64(1e-06) - training/global_step:36 - training/epoch:1 - critic/score/mean:0.61328125 - critic/score/max:1.0 - critic/score/min:0.0 - critic/rewards/mean:0.61328125 - critic/rewards/max:1.0 - critic/rewards/min:0.0 - critic/advantages/mean:0.003078529378399253 - critic/advantages/max:0.875 - critic/advantages/min:-0.875 - critic/returns/mean:0.003078529378399253 - critic/returns/max:0.875 - critic/returns/min:-0.875 - response_length/mean:317.05859375 - response_length/max:613.0 - r
|
||
step:37 - global_seqlen/min:18214 - global_seqlen/max:23268 - global_seqlen/minmax_diff:5054 - global_seqlen/balanced_min:21453 - global_seqlen/balanced_max:21458 - global_seqlen/mean:21456.0 - actor/entropy:0.15314580500125885 - perf/mfu/actor_infer:0 - training/rollout_probs_diff_valid:1 - training/rollout_probs_diff_max:0.41912177205085754 - training/rollout_probs_diff_mean:0.002847726223990321 - training/rollout_probs_diff_std:0.009974293410778046 - training/rollout_actor_probs_pearson_corr:0.9986763000488281 - self_distillation/success_group_fraction:0.9375 - self_distillation/success_sample_fraction:0.9375 - self_distillation/correct_sample_fraction:0.8515625 - self_distillation/feedback_available_fraction:0.0 - self_distillation/feedback_used_fraction:0.0 - self_distillation/reprompt_sample_fraction:0.9375 - rollout_corr/rollout_is_mean:1.0001238584518433 - rollout_corr/rollout_is_ratio_fraction_high:1.2167818567832e-05 - rollout_corr/rollout_is_ratio_fraction_low:1.2167818567832e-05 - rollout_corr/rollout_is_max:2.73596453666687 - rollout_corr/rollout_is_min:0.2720000147819519 - rollout_corr/rollout_is_std:0.029622619971632957 - rollout_corr/rollout_is_eff_sample_size:0.9991236267512279 - rollout_corr/rollout_is_seq_mean:1.0001177787780762 - rollout_corr/rollout_is_seq_std:0.001862207194790244 - rollout_corr/rollout_is_seq_max:1.0055677890777588 - rollout_corr/rollout_is_seq_min:0.9919551014900208 - rollout_corr/rollout_is_seq_max_deviation:0.008044898509979248 - rollout_corr/rollout_is_seq_fraction_high:0.0 - rollout_corr/rollout_is_seq_fraction_low:0.0 - rollout_corr/training_ppl:np.float64(1.164930485188961) - rollout_corr/training_log_ppl:np.float64(0.14994544762885198) - rollout_corr/kl:np.float64(0.00047978433622719585) - rollout_corr/k3_kl:np.float64(0.0005115730669871255) - rollout_corr/rollout_ppl:np.float64(1.1643610489554703) - rollout_corr/rollout_log_ppl:np.float64(0.1494656624199706) - rollout_corr/log_ppl_diff:np.float64(0.000479785208881367) - rollout_corr/log_ppl_abs_diff:np.float64(0.0015769913807162084) - rollout_corr/log_ppl_diff_max:np.float64(0.008887842297554016) - rollout_corr/log_ppl_diff_min:np.float64(-0.005079410970211029) - rollout_corr/ppl_ratio:np.float64(1.0004820607136935) - rollout_corr/chi2_token:np.float64(0.0011041394900530577) - rollout_corr/chi2_seq:np.float64(0.8046807374339551) - actor/pg_loss:np.float64(0.0009413559455424547) - actor/kl_loss:np.float64(0.0) - actor/pg_clipfrac:np.float64(0.00015754969035697286) - actor/ppo_kl:np.float64(0.0001336421836066748) - actor/pg_clipfrac_lower:np.float64(0.0) - self_distillation/token_reweight_lambda:np.float64(0.5) - self_distillation/token_reweight_eps_w:np.float64(0.20000000000000004) - self_distillation/token_reweight_active_token_fraction:np.float64(0.9375) - self_distillation/token_reweight_context_sample_fraction:np.float64(0.9375) - self_distillation/token_reweight_correct_sample_fraction:np.float64(0.8515625) - self_distillation/token_reweight_w_mean:np.float64(30945.948427035706) - self_distillation/token_reweight_w_std:np.float64(398418.81097031885) - self_distillation/token_reweight_w_clipped_mean:np.float64(0.9998940988443792) - self_distillation/token_reweight_w_clip_frac:np.float64(0.031333529215771705) - self_distillation/empty_target_batch:np.float64(0.0625) - actor/grad_norm:np.float64(0.17728198319673538) - perf/mfu/actor:np.float64(0.048920584162657293) - perf/max_memory_allocated_gb:np.float64(130.98325157165527) - perf/max_memory_reserved_gb:np.float64(140.86328125) - perf/cpu_memory_used_gb:np.float64(195.46635818481445) - actor/lr:np.float64(1e-06) - training/global_step:37 - training/epoch:1 - critic/score/mean:0.8515625 - critic/score/max:1.0 - critic/score/min:0.0 - critic/rewards/mean:0.8515625 - critic/rewards/max:1.0 - critic/rewards/min:0.0 - critic/advantages/mean:-0.0021643508225679398 - critic/advantages/max:0.75 - critic/advantages/min:-0.875 - critic/returns/mean:-0.0021643508225679398 - critic/returns/max:0.75 - critic/returns/min:-0.875 - response_length/mean:321.03125 - response_length/max:78
|
||
step:38 - global_seqlen/min:19052 - global_seqlen/max:25764 - global_seqlen/minmax_diff:6712 - global_seqlen/balanced_min:23011 - global_seqlen/balanced_max:23014 - global_seqlen/mean:23012.25 - actor/entropy:0.1633475422859192 - perf/mfu/actor_infer:0 - training/rollout_probs_diff_valid:1 - training/rollout_probs_diff_max:0.508357584476471 - training/rollout_probs_diff_mean:0.0028399399016052485 - training/rollout_probs_diff_std:0.009923688136041164 - training/rollout_actor_probs_pearson_corr:0.9988518953323364 - self_distillation/success_group_fraction:0.8125 - self_distillation/success_sample_fraction:0.80078125 - self_distillation/correct_sample_fraction:0.60546875 - self_distillation/feedback_available_fraction:0.0 - self_distillation/feedback_used_fraction:0.0 - self_distillation/reprompt_sample_fraction:0.80078125 - rollout_corr/rollout_is_mean:1.000065565109253 - rollout_corr/rollout_is_ratio_fraction_high:1.1911284673260525e-05 - rollout_corr/rollout_is_ratio_fraction_low:1.1911284673260525e-05 - rollout_corr/rollout_is_max:2.890216827392578 - rollout_corr/rollout_is_min:0.3545355796813965 - rollout_corr/rollout_is_std:0.02890154905617237 - rollout_corr/rollout_is_eff_sample_size:0.9991653976510357 - rollout_corr/rollout_is_seq_mean:0.9999657273292542 - rollout_corr/rollout_is_seq_std:0.0017425058176741004 - rollout_corr/rollout_is_seq_max:1.009063482284546 - rollout_corr/rollout_is_seq_min:0.9950194358825684 - rollout_corr/rollout_is_seq_max_deviation:0.009063482284545898 - rollout_corr/rollout_is_seq_fraction_high:0.0 - rollout_corr/rollout_is_seq_fraction_low:0.0 - rollout_corr/training_ppl:np.float64(1.1723122037947178) - rollout_corr/training_log_ppl:np.float64(0.15486336171306903) - rollout_corr/kl:np.float64(0.000746124695474748) - rollout_corr/k3_kl:np.float64(0.0006364840925243698) - rollout_corr/rollout_ppl:np.float64(1.1714562741108239) - rollout_corr/rollout_log_ppl:np.float64(0.15411723549914313) - rollout_corr/log_ppl_diff:np.float64(0.0007461262139258906) - rollout_corr/log_ppl_abs_diff:np.float64(0.001779028432792984) - rollout_corr/log_ppl_diff_max:np.float64(0.01745888590812683) - rollout_corr/log_ppl_diff_min:np.float64(-0.005290612578392029) - rollout_corr/ppl_ratio:np.float64(1.000749502563849) - rollout_corr/chi2_token:np.float64(0.0008848432917147875) - rollout_corr/chi2_seq:np.float64(0.41483693197369576) - actor/pg_loss:np.float64(0.0016468875110149384) - actor/kl_loss:np.float64(0.0) - actor/pg_clipfrac:np.float64(0.00032333156786990003) - actor/ppo_kl:np.float64(0.0002796881955546837) - actor/pg_clipfrac_lower:np.float64(0.0) - self_distillation/token_reweight_lambda:np.float64(0.5) - self_distillation/token_reweight_eps_w:np.float64(0.20000000000000004) - self_distillation/token_reweight_active_token_fraction:np.float64(0.80078125) - self_distillation/token_reweight_context_sample_fraction:np.float64(0.80078125) - self_distillation/token_reweight_correct_sample_fraction:np.float64(0.60546875) - self_distillation/token_reweight_w_mean:np.float64(109265.09118858888) - self_distillation/token_reweight_w_std:np.float64(1775391.109999408) - self_distillation/token_reweight_w_clipped_mean:np.float64(1.001167719019577) - self_distillation/token_reweight_w_clip_frac:np.float64(0.0509182026507915) - self_distillation/empty_target_batch:np.float64(0.19921875) - actor/grad_norm:np.float64(0.25122565031051636) - perf/mfu/actor:np.float64(0.05191056955392058) - perf/max_memory_allocated_gb:np.float64(130.98325157165527) - perf/max_memory_reserved_gb:np.float64(140.86328125) - perf/cpu_memory_used_gb:np.float64(195.49505996704102) - actor/lr:np.float64(1e-06) - training/global_step:38 - training/epoch:1 - critic/score/mean:0.60546875 - critic/score/max:1.0 - critic/score/min:0.0 - critic/rewards/mean:0.60546875 - critic/rewards/max:1.0 - critic/rewards/min:0.0 - critic/advantages/mean:0.0012462182203307748 - critic/advantages/max:0.875 - critic/advantages/min:-0.875 - critic/returns/mean:0.0012462182203307748 - critic/returns/max:0.875 - critic/returns/min:-0.875 - response_length/mean:327.9453125
|
||
step:39 - global_seqlen/min:19236 - global_seqlen/max:27409 - global_seqlen/minmax_diff:8173 - global_seqlen/balanced_min:22689 - global_seqlen/balanced_max:22691 - global_seqlen/mean:22690.5 - actor/entropy:0.1420234888792038 - perf/mfu/actor_infer:0 - training/rollout_probs_diff_valid:1 - training/rollout_probs_diff_max:0.6971966028213501 - training/rollout_probs_diff_mean:0.0028620732482522726 - training/rollout_probs_diff_std:0.01034072507172823 - training/rollout_actor_probs_pearson_corr:0.9985620975494385 - self_distillation/success_group_fraction:0.8125 - self_distillation/success_sample_fraction:0.80859375 - self_distillation/correct_sample_fraction:0.7578125 - self_distillation/feedback_available_fraction:0.0 - self_distillation/feedback_used_fraction:0.0 - self_distillation/reprompt_sample_fraction:0.80859375 - rollout_corr/rollout_is_mean:1.0000659227371216 - rollout_corr/rollout_is_ratio_fraction_high:0.0 - rollout_corr/rollout_is_ratio_fraction_low:2.4261245926027186e-05 - rollout_corr/rollout_is_max:1.6453677415847778 - rollout_corr/rollout_is_min:0.20416899025440216 - rollout_corr/rollout_is_std:0.02925819344818592 - rollout_corr/rollout_is_eff_sample_size:0.9991448092781762 - rollout_corr/rollout_is_seq_mean:1.0000933408737183 - rollout_corr/rollout_is_seq_std:0.0020522912964224815 - rollout_corr/rollout_is_seq_max:1.0066521167755127 - rollout_corr/rollout_is_seq_min:0.99286949634552 - rollout_corr/rollout_is_seq_max_deviation:0.0071305036544799805 - rollout_corr/rollout_is_seq_fraction_high:0.0 - rollout_corr/rollout_is_seq_fraction_low:0.0 - rollout_corr/training_ppl:np.float64(1.1582270697690547) - rollout_corr/training_log_ppl:np.float64(0.1451856552012032) - rollout_corr/kl:np.float64(0.0004661589703989222) - rollout_corr/k3_kl:np.float64(0.0005055180065909326) - rollout_corr/rollout_ppl:np.float64(1.1576528307050467) - rollout_corr/rollout_log_ppl:np.float64(0.1447194957145257) - rollout_corr/log_ppl_diff:np.float64(0.0004661594866774976) - rollout_corr/log_ppl_abs_diff:np.float64(0.0015421572024933994) - rollout_corr/log_ppl_diff_max:np.float64(0.010017722845077515) - rollout_corr/log_ppl_diff_min:np.float64(-0.006251499056816101) - rollout_corr/ppl_ratio:np.float64(1.0004685285966843) - rollout_corr/chi2_token:np.float64(0.0010773444082587957) - rollout_corr/chi2_seq:np.float64(0.30354314390569925) - actor/pg_loss:np.float64(0.00034410133957862854) - actor/kl_loss:np.float64(0.0) - actor/pg_clipfrac:np.float64(9.910650533129228e-05) - actor/ppo_kl:np.float64(7.239252952295772e-05) - actor/pg_clipfrac_lower:np.float64(0.0) - self_distillation/token_reweight_lambda:np.float64(0.5) - self_distillation/token_reweight_eps_w:np.float64(0.20000000000000004) - self_distillation/token_reweight_active_token_fraction:np.float64(0.80859375) - self_distillation/token_reweight_context_sample_fraction:np.float64(0.80859375) - self_distillation/token_reweight_correct_sample_fraction:np.float64(0.7578125) - self_distillation/token_reweight_w_mean:np.float64(16097.881471956614) - self_distillation/token_reweight_w_std:np.float64(291332.0561478417) - self_distillation/token_reweight_w_clipped_mean:np.float64(1.000020751496777) - self_distillation/token_reweight_w_clip_frac:np.float64(0.00949771508021513) - self_distillation/empty_target_batch:np.float64(0.19140625) - actor/grad_norm:np.float64(0.07992621650919318) - perf/mfu/actor:np.float64(0.05144791179433724) - perf/max_memory_allocated_gb:np.float64(130.98325157165527) - perf/max_memory_reserved_gb:np.float64(140.86328125) - perf/cpu_memory_used_gb:np.float64(195.4413948059082) - actor/lr:np.float64(1e-06) - training/global_step:39 - training/epoch:1 - critic/score/mean:0.7578125 - critic/score/max:1.0 - critic/score/min:0.0 - critic/rewards/mean:0.7578125 - critic/rewards/max:1.0 - critic/rewards/min:0.0 - critic/advantages/mean:0.002805206459015608 - critic/advantages/max:0.875 - critic/advantages/min:-0.875 - critic/returns/mean:0.002805206459015608 - critic/returns/max:0.875 - critic/returns/min:-0.875 - response_length/mean:322.015625 - response_length/max:
|
||
test_gen_batch meta info: {'eos_token_id': 151645, 'pad_token_id': 151643, 'recompute_log_prob': False, 'do_sample': True, 'validate': True, 'global_steps': 40}
|
||
validation generation end
|
||
[prompt] system
|
||
|
||
Given a question and four options, please select the right answer. Respond in the following format:
|
||
<reasoning>
|
||
...
|
||
</reasoning>
|
||
<answer>
|
||
...
|
||
</answer>
|
||
|
||
For the answer, only output the letter corresponding to the correct option (A, B, C, or D), and nothing else. Do not restate the answer text. For example, if the answer is "A", just output:
|
||
<answer>
|
||
A
|
||
</answer>
|
||
|
||
user
|
||
Given the following crystal structure parameters for the material RbLa9(IrO6)4 (Material ID: mp-560657), calculate the volume of the unit cell (in ų).
|
||
|
||
Lattice Parameters:
|
||
|
||
a = 7.82 Å
|
||
b = 7.82 Å
|
||
c = 17.88 Å
|
||
α = 90.00º
|
||
β = 90.00º
|
||
γ = 90.00º
|
||
|
||
|
||
|
||
A: 1025.67 ų
|
||
|
||
B: 1094.31 ų
|
||
|
||
C: 1200.45 ų
|
||
|
||
D: 1150.78 ų
|
||
|
||
Please reason step by step.
|
||
assistant
|
||
<think>
|
||
|
||
</think>
|
||
|
||
|
||
[response] <reasoning>
|
||
The given crystal structure has lattice parameters a = 7.82 Å, b = 7.82 Å, c = 17.88 Å, and all angles are 90.00º, which indicates a orthorhombic crystal system.
|
||
|
||
For an orthorhombic system, the volume $ V $ of the unit cell is calculated using the formula:
|
||
|
||
$$
|
||
V = a \times b \times c
|
||
$$
|
||
|
||
Substituting the given values:
|
||
|
||
$$
|
||
V = 7.82 \times 7.82 \times 17.88
|
||
$$
|
||
|
||
First, calculate $ 7.82 \times 7.82 $:
|
||
|
||
$$
|
||
7.82 \times 7.82 = 61.1524
|
||
$$
|
||
|
||
Then multiply by 17.88:
|
||
|
||
$$
|
||
61.1524 \times 17.88 = 1094.31
|
||
$$
|
||
|
||
Thus, the volume of the unit cell is approximately 1094.31 ų.
|
||
</reasoning>
|
||
<answer>
|
||
B
|
||
</answer>
|
||
[ground_truth] B
|
||
[score] 1.0
|
||
[acc] 1.0
|
||
[pred] B
|
||
[incorrect_format] 1
|
||
[feedback]
|
||
local_global_step_folder: /mnt/mole/SDPO/L2T/checkpoints/datasets/sciknoweval/material/qwen3gen-material-RLSD_TR-Qwen-Qwen3-8B-mbs8-decay0-tr0.1-train32-rollout8-lr1e-6-vllm0.8/global_step_40
|
||
step:40 - global_seqlen/min:17835 - global_seqlen/max:24761 - global_seqlen/minmax_diff:6926 - global_seqlen/balanced_min:21932 - global_seqlen/balanced_max:22013 - global_seqlen/mean:21943.625 - actor/entropy:0.14098937809467316 - perf/mfu/actor_infer:0 - training/rollout_probs_diff_valid:1 - training/rollout_probs_diff_max:0.2553865313529968 - training/rollout_probs_diff_mean:0.0029350039549171925 - training/rollout_probs_diff_std:0.01030927523970604 - training/rollout_actor_probs_pearson_corr:0.9984802007675171 - self_distillation/success_group_fraction:0.90625 - self_distillation/success_sample_fraction:0.8984375 - self_distillation/correct_sample_fraction:0.8515625 - self_distillation/feedback_available_fraction:0.0 - self_distillation/feedback_used_fraction:0.0 - self_distillation/reprompt_sample_fraction:0.8984375 - rollout_corr/rollout_is_mean:1.0000852346420288 - rollout_corr/rollout_is_ratio_fraction_high:0.0 - rollout_corr/rollout_is_ratio_fraction_low:1.2781676559825428e-05 - rollout_corr/rollout_is_max:1.9516299962997437 - rollout_corr/rollout_is_min:0.41667383909225464 - rollout_corr/rollout_is_std:0.02979115955531597 - rollout_corr/rollout_is_eff_sample_size:0.9991133928184341 - rollout_corr/rollout_is_seq_mean:1.0000969171524048 - rollout_corr/rollout_is_seq_std:0.0018581942422315478 - rollout_corr/rollout_is_seq_max:1.0077598094940186 - rollout_corr/rollout_is_seq_min:0.9940147995948792 - rollout_corr/rollout_is_seq_max_deviation:0.007759809494018555 - rollout_corr/rollout_is_seq_fraction_high:0.0 - rollout_corr/rollout_is_seq_fraction_low:0.0 - rollout_corr/training_ppl:np.float64(1.1458545695059001) - rollout_corr/training_log_ppl:np.float64(0.1349656182937906) - rollout_corr/kl:np.float64(0.0004070129167867975) - rollout_corr/k3_kl:np.float64(0.0004980892093442435) - rollout_corr/rollout_ppl:np.float64(1.1453717714175582) - rollout_corr/rollout_log_ppl:np.float64(0.13455860382237006) - rollout_corr/log_ppl_diff:np.float64(0.0004070144714205526) - rollout_corr/log_ppl_abs_diff:np.float64(0.0015183983850874938) - rollout_corr/log_ppl_diff_max:np.float64(0.006893813610076904) - rollout_corr/log_ppl_diff_min:np.float64(-0.00674157589673996) - rollout_corr/ppl_ratio:np.float64(1.0004091300070286) - rollout_corr/chi2_token:np.float64(0.0011687756050378084) - rollout_corr/chi2_seq:np.float64(0.38695761072449386) - actor/pg_loss:np.float64(8.210085798054934e-05) - actor/kl_loss:np.float64(0.0) - actor/pg_clipfrac:np.float64(5.997138350721798e-05) - actor/ppo_kl:np.float64(4.544163184760919e-05) - actor/pg_clipfrac_lower:np.float64(0.0) - self_distillation/token_reweight_lambda:np.float64(0.5) - self_distillation/token_reweight_eps_w:np.float64(0.20000000000000004) - self_distillation/token_reweight_active_token_fraction:np.float64(0.8984375) - self_distillation/token_reweight_context_sample_fraction:np.float64(0.8984375) - self_distillation/token_reweight_correct_sample_fraction:np.float64(0.8515625) - self_distillation/token_reweight_w_mean:np.float64(13679.636977609713) - self_distillation/token_reweight_w_std:np.float64(279038.5410264209) - self_distillation/token_reweight_w_clipped_mean:np.float64(1.0003306025173515) - self_distillation/token_reweight_w_clip_frac:np.float64(0.005064020268036984) - self_distillation/empty_target_batch:np.float64(0.1015625) - actor/grad_norm:np.float64(0.03898233640938997) - perf/mfu/actor:np.float64(0.04943869679337855) - perf/max_memory_allocated_gb:np.float64(130.98326301574707) - perf/max_memory_reserved_gb:np.float64(140.86328125) - perf/cpu_memory_used_gb:np.float64(195.3865146636963) - actor/lr:np.float64(1e-06) - val-aux/sciknoweval/reward/mean@16:np.float64(0.7340425531914894) - val-aux/sciknoweval/reward/std@16:np.float64(0.07158614242680836) - val-aux/sciknoweval/reward/best@2/mean:np.float64(0.7607872340425531) - val-aux/sciknoweval/reward/best@2/std:np.float64(0.06279756648335182) - val-aux/sciknoweval/reward/worst@2/mean:np.float64(0.7073085106382978) - val-aux/sciknoweval/reward/worst@2/std:np.float64(0.05685132383706091) - val-aux/sciknoweval/reward/maj@2/me
|
||
step:41 - global_seqlen/min:20580 - global_seqlen/max:26975 - global_seqlen/minmax_diff:6395 - global_seqlen/balanced_min:24255 - global_seqlen/balanced_max:24260 - global_seqlen/mean:24257.625 - actor/entropy:0.13618877530097961 - perf/mfu/actor_infer:0 - training/rollout_probs_diff_valid:1 - training/rollout_probs_diff_max:0.4967532455921173 - training/rollout_probs_diff_mean:0.0028353757224977016 - training/rollout_probs_diff_std:0.010483196936547756 - training/rollout_actor_probs_pearson_corr:0.9983766078948975 - self_distillation/success_group_fraction:0.71875 - self_distillation/success_sample_fraction:0.71484375 - self_distillation/correct_sample_fraction:0.59765625 - self_distillation/feedback_available_fraction:0.0 - self_distillation/feedback_used_fraction:0.0 - self_distillation/reprompt_sample_fraction:0.71484375 - rollout_corr/rollout_is_mean:1.0000534057617188 - rollout_corr/rollout_is_ratio_fraction_high:1.0927887160505634e-05 - rollout_corr/rollout_is_ratio_fraction_low:4.3711548642022535e-05 - rollout_corr/rollout_is_max:2.463893175125122 - rollout_corr/rollout_is_min:0.3244990110397339 - rollout_corr/rollout_is_std:0.029735084623098373 - rollout_corr/rollout_is_eff_sample_size:0.9991167247735356 - rollout_corr/rollout_is_seq_mean:1.000001072883606 - rollout_corr/rollout_is_seq_std:0.0018027505138888955 - rollout_corr/rollout_is_seq_max:1.005057454109192 - rollout_corr/rollout_is_seq_min:0.9950371384620667 - rollout_corr/rollout_is_seq_max_deviation:0.0050574541091918945 - rollout_corr/rollout_is_seq_fraction_high:0.0 - rollout_corr/rollout_is_seq_fraction_low:0.0 - rollout_corr/training_ppl:np.float64(1.1445708661340177) - rollout_corr/training_log_ppl:np.float64(0.1329260604143201) - rollout_corr/kl:np.float64(0.0004669340913103781) - rollout_corr/k3_kl:np.float64(0.0004507001112799003) - rollout_corr/rollout_ppl:np.float64(1.1440314431674778) - rollout_corr/rollout_log_ppl:np.float64(0.1324591252778191) - rollout_corr/log_ppl_diff:np.float64(0.00046693513650097884) - rollout_corr/log_ppl_abs_diff:np.float64(0.001347396111668786) - rollout_corr/log_ppl_diff_max:np.float64(0.005796477198600769) - rollout_corr/log_ppl_diff_min:np.float64(-0.0040665604174137115) - rollout_corr/ppl_ratio:np.float64(1.0004686149768531) - rollout_corr/chi2_token:np.float64(0.0008493810892105103) - rollout_corr/chi2_seq:np.float64(0.331406096694991) - actor/pg_loss:np.float64(0.0016671433113515377) - actor/kl_loss:np.float64(0.0) - actor/pg_clipfrac:np.float64(0.000255978321092698) - actor/ppo_kl:np.float64(1.2966889821441896e-05) - actor/pg_clipfrac_lower:np.float64(0.0) - self_distillation/token_reweight_lambda:np.float64(0.5) - self_distillation/token_reweight_eps_w:np.float64(0.20000000000000004) - self_distillation/token_reweight_active_token_fraction:np.float64(0.71484375) - self_distillation/token_reweight_context_sample_fraction:np.float64(0.71484375) - self_distillation/token_reweight_correct_sample_fraction:np.float64(0.59765625) - self_distillation/token_reweight_w_mean:np.float64(22113.47238512803) - self_distillation/token_reweight_w_std:np.float64(394551.27365861356) - self_distillation/token_reweight_w_clipped_mean:np.float64(1.0004445768427104) - self_distillation/token_reweight_w_clip_frac:np.float64(0.02935507533402415) - self_distillation/empty_target_batch:np.float64(0.28515625) - actor/grad_norm:np.float64(0.1600772701203823) - perf/mfu/actor:np.float64(0.05537688033283876) - perf/max_memory_allocated_gb:np.float64(130.98326301574707) - perf/max_memory_reserved_gb:np.float64(140.86328125) - perf/cpu_memory_used_gb:np.float64(196.28454208374023) - actor/lr:np.float64(1e-06) - training/global_step:41 - training/epoch:1 - critic/score/mean:0.59765625 - critic/score/max:1.0 - critic/score/min:0.0 - critic/rewards/mean:0.59765625 - critic/rewards/max:1.0 - critic/rewards/min:0.0 - critic/advantages/mean:-0.002195139182731509 - critic/advantages/max:0.875 - critic/advantages/min:-0.875 - critic/returns/mean:-0.002195139182731509 - critic/returns/max:0.875 - critic/returns/min:-0.875 - response_length/mean:357.4
|
||
step:42 - global_seqlen/min:22411 - global_seqlen/max:27248 - global_seqlen/minmax_diff:4837 - global_seqlen/balanced_min:24246 - global_seqlen/balanced_max:24253 - global_seqlen/mean:24250.5 - actor/entropy:0.14001302421092987 - perf/mfu/actor_infer:0 - training/rollout_probs_diff_valid:1 - training/rollout_probs_diff_max:0.5599758625030518 - training/rollout_probs_diff_mean:0.002688286593183875 - training/rollout_probs_diff_std:0.009930191561579704 - training/rollout_actor_probs_pearson_corr:0.9986210465431213 - self_distillation/success_group_fraction:0.75 - self_distillation/success_sample_fraction:0.734375 - self_distillation/correct_sample_fraction:0.59375 - self_distillation/feedback_available_fraction:0.0 - self_distillation/feedback_used_fraction:0.0 - self_distillation/reprompt_sample_fraction:0.734375 - rollout_corr/rollout_is_mean:0.9999199509620667 - rollout_corr/rollout_is_ratio_fraction_high:1.1234440535190515e-05 - rollout_corr/rollout_is_ratio_fraction_low:3.370331978658214e-05 - rollout_corr/rollout_is_max:2.4000132083892822 - rollout_corr/rollout_is_min:0.3642396628856659 - rollout_corr/rollout_is_std:0.028286617249250412 - rollout_corr/rollout_is_eff_sample_size:0.9992003879303721 - rollout_corr/rollout_is_seq_mean:0.9999122023582458 - rollout_corr/rollout_is_seq_std:0.0016149813309311867 - rollout_corr/rollout_is_seq_max:1.0048774480819702 - rollout_corr/rollout_is_seq_min:0.995093584060669 - rollout_corr/rollout_is_seq_max_deviation:0.004906415939331055 - rollout_corr/rollout_is_seq_fraction_high:0.0 - rollout_corr/rollout_is_seq_fraction_low:0.0 - rollout_corr/training_ppl:np.float64(1.1585845770314336) - rollout_corr/training_log_ppl:np.float64(0.14303961391851772) - rollout_corr/kl:np.float64(0.0006641505782014612) - rollout_corr/k3_kl:np.float64(0.0005312824764303059) - rollout_corr/rollout_ppl:np.float64(1.1577940359711647) - rollout_corr/rollout_log_ppl:np.float64(0.14237546218646457) - rollout_corr/log_ppl_diff:np.float64(0.000664151732053142) - rollout_corr/log_ppl_abs_diff:np.float64(0.0014161761864670552) - rollout_corr/log_ppl_diff_max:np.float64(0.009874150156974792) - rollout_corr/log_ppl_diff_min:np.float64(-0.004382699728012085) - rollout_corr/ppl_ratio:np.float64(1.000665974104777) - rollout_corr/chi2_token:np.float64(0.0007279803976416588) - rollout_corr/chi2_seq:np.float64(0.08038255502469838) - actor/pg_loss:np.float64(0.0011525413719937205) - actor/kl_loss:np.float64(0.0) - actor/pg_clipfrac:np.float64(0.00020648631925723748) - actor/ppo_kl:np.float64(0.00015610256760312247) - actor/pg_clipfrac_lower:np.float64(0.0) - self_distillation/token_reweight_lambda:np.float64(0.5) - self_distillation/token_reweight_eps_w:np.float64(0.20000000000000004) - self_distillation/token_reweight_active_token_fraction:np.float64(0.734375) - self_distillation/token_reweight_context_sample_fraction:np.float64(0.734375) - self_distillation/token_reweight_correct_sample_fraction:np.float64(0.59375) - self_distillation/token_reweight_w_mean:np.float64(26345.013721656986) - self_distillation/token_reweight_w_std:np.float64(467469.8246827084) - self_distillation/token_reweight_w_clipped_mean:np.float64(1.0015952868852764) - self_distillation/token_reweight_w_clip_frac:np.float64(0.03433779346232768) - self_distillation/empty_target_batch:np.float64(0.265625) - actor/grad_norm:np.float64(0.193668182939291) - perf/mfu/actor:np.float64(0.054983356597584125) - perf/max_memory_allocated_gb:np.float64(130.98326301574707) - perf/max_memory_reserved_gb:np.float64(140.86328125) - perf/cpu_memory_used_gb:np.float64(195.84840393066406) - actor/lr:np.float64(1e-06) - training/global_step:42 - training/epoch:1 - critic/score/mean:0.59375 - critic/score/max:1.0 - critic/score/min:0.0 - critic/rewards/mean:0.59375 - critic/rewards/max:1.0 - critic/rewards/min:0.0 - critic/advantages/mean:0.0037874106783419847 - critic/advantages/max:0.875 - critic/advantages/min:-0.875 - critic/returns/mean:0.0037874106783419847 - critic/returns/max:0.875 - critic/returns/min:-0.875 - response_length/mean:347.703125 - response_length/m
|
||
step:43 - global_seqlen/min:19395 - global_seqlen/max:28000 - global_seqlen/minmax_diff:8605 - global_seqlen/balanced_min:24044 - global_seqlen/balanced_max:24049 - global_seqlen/mean:24046.875 - actor/entropy:0.1420712172985077 - perf/mfu/actor_infer:0 - training/rollout_probs_diff_valid:1 - training/rollout_probs_diff_max:0.6292536854743958 - training/rollout_probs_diff_mean:0.002893645316362381 - training/rollout_probs_diff_std:0.01048985868692398 - training/rollout_actor_probs_pearson_corr:0.9984310269355774 - self_distillation/success_group_fraction:0.59375 - self_distillation/success_sample_fraction:0.58984375 - self_distillation/correct_sample_fraction:0.515625 - self_distillation/feedback_available_fraction:0.0 - self_distillation/feedback_used_fraction:0.0 - self_distillation/reprompt_sample_fraction:0.58984375 - rollout_corr/rollout_is_mean:0.9999460577964783 - rollout_corr/rollout_is_ratio_fraction_high:0.0 - rollout_corr/rollout_is_ratio_fraction_low:4.5401404349831864e-05 - rollout_corr/rollout_is_max:1.6487189531326294 - rollout_corr/rollout_is_min:0.2613983452320099 - rollout_corr/rollout_is_std:0.028771327808499336 - rollout_corr/rollout_is_eff_sample_size:0.9991727763495746 - rollout_corr/rollout_is_seq_mean:1.0000054836273193 - rollout_corr/rollout_is_seq_std:0.0017316376324743032 - rollout_corr/rollout_is_seq_max:1.0067939758300781 - rollout_corr/rollout_is_seq_min:0.9957586526870728 - rollout_corr/rollout_is_seq_max_deviation:0.006793975830078125 - rollout_corr/rollout_is_seq_fraction_high:0.0 - rollout_corr/rollout_is_seq_fraction_low:0.0 - rollout_corr/training_ppl:np.float64(1.157421734649688) - rollout_corr/training_log_ppl:np.float64(0.14304501365404576) - rollout_corr/kl:np.float64(0.0005448067362863185) - rollout_corr/k3_kl:np.float64(0.0005209827327945504) - rollout_corr/rollout_ppl:np.float64(1.1567890807054937) - rollout_corr/rollout_log_ppl:np.float64(0.14250020634790417) - rollout_corr/log_ppl_diff:np.float64(0.0005448073061415926) - rollout_corr/log_ppl_abs_diff:np.float64(0.0015114768029889092) - rollout_corr/log_ppl_diff_max:np.float64(0.006321355700492859) - rollout_corr/log_ppl_diff_min:np.float64(-0.00696495920419693) - rollout_corr/ppl_ratio:np.float64(1.0005467752926052) - rollout_corr/chi2_token:np.float64(0.0010000739712268114) - rollout_corr/chi2_seq:np.float64(0.3491819892078638) - actor/pg_loss:np.float64(0.0009307680884376168) - actor/kl_loss:np.float64(0.0) - actor/pg_clipfrac:np.float64(4.09595422752318e-05) - actor/ppo_kl:np.float64(0.00011152088650501213) - actor/pg_clipfrac_lower:np.float64(0.0) - self_distillation/token_reweight_lambda:np.float64(0.5) - self_distillation/token_reweight_eps_w:np.float64(0.20000000000000004) - self_distillation/token_reweight_active_token_fraction:np.float64(0.58984375) - self_distillation/token_reweight_context_sample_fraction:np.float64(0.58984375) - self_distillation/token_reweight_correct_sample_fraction:np.float64(0.515625) - self_distillation/token_reweight_w_mean:np.float64(14319.957484231563) - self_distillation/token_reweight_w_std:np.float64(238024.43798052194) - self_distillation/token_reweight_w_clipped_mean:np.float64(1.0005522321444005) - self_distillation/token_reweight_w_clip_frac:np.float64(0.0240835617587436) - self_distillation/empty_target_batch:np.float64(0.41015625) - actor/grad_norm:np.float64(0.12713662534952164) - perf/mfu/actor:np.float64(0.0551132476644326) - perf/max_memory_allocated_gb:np.float64(130.98326301574707) - perf/max_memory_reserved_gb:np.float64(140.86328125) - perf/cpu_memory_used_gb:np.float64(195.6787872314453) - actor/lr:np.float64(1e-06) - training/global_step:43 - training/epoch:1 - critic/score/mean:0.515625 - critic/score/max:1.0 - critic/score/min:0.0 - critic/rewards/mean:0.515625 - critic/rewards/max:1.0 - critic/rewards/min:0.0 - critic/advantages/mean:0.0012457011034712195 - critic/advantages/max:0.875 - critic/advantages/min:-0.875 - critic/returns/mean:0.0012457011034712195 - critic/returns/max:0.875 - critic/returns/min:-0.875 - response_length/mean:344.15234375 - response_length/max
|
||
step:44 - global_seqlen/min:21992 - global_seqlen/max:25456 - global_seqlen/minmax_diff:3464 - global_seqlen/balanced_min:23423 - global_seqlen/balanced_max:23432 - global_seqlen/mean:23429.125 - actor/entropy:0.13345500826835632 - perf/mfu/actor_infer:0 - training/rollout_probs_diff_valid:1 - training/rollout_probs_diff_max:0.3015652000904083 - training/rollout_probs_diff_mean:0.0027469515334814787 - training/rollout_probs_diff_std:0.009971688501536846 - training/rollout_actor_probs_pearson_corr:0.9986183643341064 - self_distillation/success_group_fraction:0.78125 - self_distillation/success_sample_fraction:0.78125 - self_distillation/correct_sample_fraction:0.72265625 - self_distillation/feedback_available_fraction:0.0 - self_distillation/feedback_used_fraction:0.0 - self_distillation/reprompt_sample_fraction:0.78125 - rollout_corr/rollout_is_mean:1.0000137090682983 - rollout_corr/rollout_is_ratio_fraction_high:0.0 - rollout_corr/rollout_is_ratio_fraction_low:1.1071621884184424e-05 - rollout_corr/rollout_is_max:1.8683162927627563 - rollout_corr/rollout_is_min:0.49738770723342896 - rollout_corr/rollout_is_std:0.02936190739274025 - rollout_corr/rollout_is_eff_sample_size:0.9991387400353914 - rollout_corr/rollout_is_seq_mean:0.9999842047691345 - rollout_corr/rollout_is_seq_std:0.0018778472440317273 - rollout_corr/rollout_is_seq_max:1.0074628591537476 - rollout_corr/rollout_is_seq_min:0.992678165435791 - rollout_corr/rollout_is_seq_max_deviation:0.007462859153747559 - rollout_corr/rollout_is_seq_fraction_high:0.0 - rollout_corr/rollout_is_seq_fraction_low:0.0 - rollout_corr/training_ppl:np.float64(1.1445557121187449) - rollout_corr/training_log_ppl:np.float64(0.13367409240163397) - rollout_corr/kl:np.float64(0.0005900809133603246) - rollout_corr/k3_kl:np.float64(0.00046168545707558906) - rollout_corr/rollout_ppl:np.float64(1.1438749125227332) - rollout_corr/rollout_log_ppl:np.float64(0.13308401135873282) - rollout_corr/log_ppl_diff:np.float64(0.0005900810429011472) - rollout_corr/log_ppl_abs_diff:np.float64(0.0015492462043766864) - rollout_corr/log_ppl_diff_max:np.float64(0.0074571967124938965) - rollout_corr/log_ppl_diff_min:np.float64(-0.008081652224063873) - rollout_corr/ppl_ratio:np.float64(1.0005922082345933) - rollout_corr/chi2_token:np.float64(0.0006860834546387196) - rollout_corr/chi2_seq:np.float64(0.6962482798844576) - actor/pg_loss:np.float64(0.0007348335348069668) - actor/kl_loss:np.float64(0.0) - actor/pg_clipfrac:np.float64(8.418954439548543e-05) - actor/ppo_kl:np.float64(0.00012582078771217198) - actor/pg_clipfrac_lower:np.float64(0.0) - self_distillation/token_reweight_lambda:np.float64(0.5) - self_distillation/token_reweight_eps_w:np.float64(0.20000000000000004) - self_distillation/token_reweight_active_token_fraction:np.float64(0.78125) - self_distillation/token_reweight_context_sample_fraction:np.float64(0.78125) - self_distillation/token_reweight_correct_sample_fraction:np.float64(0.72265625) - self_distillation/token_reweight_w_mean:np.float64(20517.69756563753) - self_distillation/token_reweight_w_std:np.float64(358047.4882045661) - self_distillation/token_reweight_w_clipped_mean:np.float64(1.000059798359871) - self_distillation/token_reweight_w_clip_frac:np.float64(0.0146756786343758) - self_distillation/empty_target_batch:np.float64(0.21875) - actor/grad_norm:np.float64(0.18048310466110706) - perf/mfu/actor:np.float64(0.05380951610340004) - perf/max_memory_allocated_gb:np.float64(130.98326301574707) - perf/max_memory_reserved_gb:np.float64(140.86328125) - perf/cpu_memory_used_gb:np.float64(195.57110977172852) - actor/lr:np.float64(1e-06) - training/global_step:44 - training/epoch:1 - critic/score/mean:0.72265625 - critic/score/max:1.0 - critic/score/min:0.0 - critic/rewards/mean:0.72265625 - critic/rewards/max:1.0 - critic/rewards/min:0.0 - critic/advantages/mean:0.006885165348649025 - critic/advantages/max:0.75 - critic/advantages/min:-0.875 - critic/returns/mean:0.006885165348649025 - critic/returns/max:0.75 - critic/returns/min:-0.875 - response_length/mean:352.81640625 - response_length/max:799.
|
||
step:45 - global_seqlen/min:18176 - global_seqlen/max:26087 - global_seqlen/minmax_diff:7911 - global_seqlen/balanced_min:22796 - global_seqlen/balanced_max:22799 - global_seqlen/mean:22797.625 - actor/entropy:0.1399414837360382 - perf/mfu/actor_infer:0 - training/rollout_probs_diff_valid:1 - training/rollout_probs_diff_max:0.4102373421192169 - training/rollout_probs_diff_mean:0.002902822569012642 - training/rollout_probs_diff_std:0.010340701788663864 - training/rollout_actor_probs_pearson_corr:0.9984586834907532 - self_distillation/success_group_fraction:0.84375 - self_distillation/success_sample_fraction:0.84375 - self_distillation/correct_sample_fraction:0.7734375 - self_distillation/feedback_available_fraction:0.0 - self_distillation/feedback_used_fraction:0.0 - self_distillation/reprompt_sample_fraction:0.84375 - rollout_corr/rollout_is_mean:1.0001752376556396 - rollout_corr/rollout_is_ratio_fraction_high:1.1424262993386947e-05 - rollout_corr/rollout_is_ratio_fraction_low:2.2848525986773893e-05 - rollout_corr/rollout_is_max:2.3329057693481445 - rollout_corr/rollout_is_min:0.3676825761795044 - rollout_corr/rollout_is_std:0.030085813254117966 - rollout_corr/rollout_is_eff_sample_size:0.9990960194125961 - rollout_corr/rollout_is_seq_mean:1.0002131462097168 - rollout_corr/rollout_is_seq_std:0.0018944222247228026 - rollout_corr/rollout_is_seq_max:1.0114330053329468 - rollout_corr/rollout_is_seq_min:0.9952909350395203 - rollout_corr/rollout_is_seq_max_deviation:0.011433005332946777 - rollout_corr/rollout_is_seq_fraction_high:0.0 - rollout_corr/rollout_is_seq_fraction_low:0.0 - rollout_corr/training_ppl:np.float64(1.1445479299873114) - rollout_corr/training_log_ppl:np.float64(0.13379971245012712) - rollout_corr/kl:np.float64(0.0005008000899939802) - rollout_corr/k3_kl:np.float64(0.0004867743777481337) - rollout_corr/rollout_ppl:np.float64(1.1439769458957016) - rollout_corr/rollout_log_ppl:np.float64(0.13329891020839568) - rollout_corr/log_ppl_diff:np.float64(0.0005008022417314351) - rollout_corr/log_ppl_abs_diff:np.float64(0.0014860984665574506) - rollout_corr/log_ppl_diff_max:np.float64(0.00550118088722229) - rollout_corr/log_ppl_diff_min:np.float64(-0.009168602526187897) - rollout_corr/ppl_ratio:np.float64(1.00050284457393) - rollout_corr/chi2_token:np.float64(0.0009574964642524719) - rollout_corr/chi2_seq:np.float64(0.4922240450978279) - actor/pg_loss:np.float64(0.0010258276015520096) - actor/kl_loss:np.float64(0.0) - actor/pg_clipfrac:np.float64(0.00016894487498575472) - actor/ppo_kl:np.float64(0.0002360405962065215) - actor/pg_clipfrac_lower:np.float64(0.0) - self_distillation/token_reweight_lambda:np.float64(0.5) - self_distillation/token_reweight_eps_w:np.float64(0.20000000000000004) - self_distillation/token_reweight_active_token_fraction:np.float64(0.84375) - self_distillation/token_reweight_context_sample_fraction:np.float64(0.84375) - self_distillation/token_reweight_correct_sample_fraction:np.float64(0.7734375) - self_distillation/token_reweight_w_mean:np.float64(9821.606562155532) - self_distillation/token_reweight_w_std:np.float64(178767.35334247636) - self_distillation/token_reweight_w_clipped_mean:np.float64(1.0001066522672772) - self_distillation/token_reweight_w_clip_frac:np.float64(0.01871949274209328) - self_distillation/empty_target_batch:np.float64(0.15625) - actor/grad_norm:np.float64(0.1153863884974271) - perf/mfu/actor:np.float64(0.05208565184680937) - perf/max_memory_allocated_gb:np.float64(130.98326301574707) - perf/max_memory_reserved_gb:np.float64(140.86328125) - perf/cpu_memory_used_gb:np.float64(195.7615509033203) - actor/lr:np.float64(1e-06) - training/global_step:45 - training/epoch:1 - critic/score/mean:0.7734375 - critic/score/max:1.0 - critic/score/min:0.0 - critic/rewards/mean:0.7734375 - critic/rewards/max:1.0 - critic/rewards/min:0.0 - critic/advantages/mean:-0.0014894382329657674 - critic/advantages/max:0.75 - critic/advantages/min:-0.875 - critic/returns/mean:-0.0014894382329657674 - critic/returns/max:0.75 - critic/returns/min:-0.875 - response_length/mean:341.92578125 - response_l
|
||
step:46 - global_seqlen/min:21540 - global_seqlen/max:26143 - global_seqlen/minmax_diff:4603 - global_seqlen/balanced_min:23587 - global_seqlen/balanced_max:23591 - global_seqlen/mean:23589.25 - actor/entropy:0.1451496034860611 - perf/mfu/actor_infer:0 - training/rollout_probs_diff_valid:1 - training/rollout_probs_diff_max:0.21502655744552612 - training/rollout_probs_diff_mean:0.002851153025403619 - training/rollout_probs_diff_std:0.009968206286430359 - training/rollout_actor_probs_pearson_corr:0.9986371994018555 - self_distillation/success_group_fraction:0.75 - self_distillation/success_sample_fraction:0.74609375 - self_distillation/correct_sample_fraction:0.703125 - self_distillation/feedback_available_fraction:0.0 - self_distillation/feedback_used_fraction:0.0 - self_distillation/reprompt_sample_fraction:0.74609375 - rollout_corr/rollout_is_mean:1.0002301931381226 - rollout_corr/rollout_is_ratio_fraction_high:0.0 - rollout_corr/rollout_is_ratio_fraction_low:0.0 - rollout_corr/rollout_is_max:1.655013918876648 - rollout_corr/rollout_is_min:0.5358285903930664 - rollout_corr/rollout_is_std:0.028942765668034554 - rollout_corr/rollout_is_eff_sample_size:0.9991633744785383 - rollout_corr/rollout_is_seq_mean:1.0002855062484741 - rollout_corr/rollout_is_seq_std:0.0016697313403710723 - rollout_corr/rollout_is_seq_max:1.005419373512268 - rollout_corr/rollout_is_seq_min:0.9953736066818237 - rollout_corr/rollout_is_seq_max_deviation:0.005419373512268066 - rollout_corr/rollout_is_seq_fraction_high:0.0 - rollout_corr/rollout_is_seq_fraction_low:0.0 - rollout_corr/training_ppl:np.float64(1.1572659551165998) - rollout_corr/training_log_ppl:np.float64(0.14247176656499505) - rollout_corr/kl:np.float64(0.0002932232152055292) - rollout_corr/k3_kl:np.float64(0.00048011249347723606) - rollout_corr/rollout_ppl:np.float64(1.156941320747137) - rollout_corr/rollout_log_ppl:np.float64(0.14217854263552) - rollout_corr/log_ppl_diff:np.float64(0.0002932239294750616) - rollout_corr/log_ppl_abs_diff:np.float64(0.0013674008805537596) - rollout_corr/log_ppl_diff_max:np.float64(0.006510406732559204) - rollout_corr/log_ppl_diff_min:np.float64(-0.004583634436130524) - rollout_corr/ppl_ratio:np.float64(1.0002947926986963) - rollout_corr/chi2_token:np.float64(0.0013452090788632631) - rollout_corr/chi2_seq:np.float64(0.4254507820587605) - actor/pg_loss:np.float64(0.0008674579439684749) - actor/kl_loss:np.float64(0.0) - actor/pg_clipfrac:np.float64(0.00011161280372107285) - actor/ppo_kl:np.float64(0.000148489029569987) - actor/pg_clipfrac_lower:np.float64(0.0) - self_distillation/token_reweight_lambda:np.float64(0.5) - self_distillation/token_reweight_eps_w:np.float64(0.20000000000000004) - self_distillation/token_reweight_active_token_fraction:np.float64(0.74609375) - self_distillation/token_reweight_context_sample_fraction:np.float64(0.74609375) - self_distillation/token_reweight_correct_sample_fraction:np.float64(0.703125) - self_distillation/token_reweight_w_mean:np.float64(14773.669593083207) - self_distillation/token_reweight_w_std:np.float64(243772.67391494405) - self_distillation/token_reweight_w_clipped_mean:np.float64(1.000106336781755) - self_distillation/token_reweight_w_clip_frac:np.float64(0.021612054741126485) - self_distillation/empty_target_batch:np.float64(0.25390625) - actor/grad_norm:np.float64(0.1143045425415039) - perf/mfu/actor:np.float64(0.05422613845648462) - perf/max_memory_allocated_gb:np.float64(130.98326301574707) - perf/max_memory_reserved_gb:np.float64(140.86328125) - perf/cpu_memory_used_gb:np.float64(195.55764389038086) - actor/lr:np.float64(1e-06) - training/global_step:46 - training/epoch:1 - critic/score/mean:0.703125 - critic/score/max:1.0 - critic/score/min:0.0 - critic/rewards/mean:0.703125 - critic/rewards/max:1.0 - critic/rewards/min:0.0 - critic/advantages/mean:-0.0023552831262350082 - critic/advantages/max:0.875 - critic/advantages/min:-0.875 - critic/returns/mean:-0.0023552831262350082 - critic/returns/max:0.875 - critic/returns/min:-0.875 - response_length/mean:356.1640625 - response_length/max:654.0 - response_len
|
||
step:47 - global_seqlen/min:20476 - global_seqlen/max:26782 - global_seqlen/minmax_diff:6306 - global_seqlen/balanced_min:24446 - global_seqlen/balanced_max:24449 - global_seqlen/mean:24447.5 - actor/entropy:0.1583218276500702 - perf/mfu/actor_infer:0 - training/rollout_probs_diff_valid:1 - training/rollout_probs_diff_max:0.30163803696632385 - training/rollout_probs_diff_mean:0.003097226144745946 - training/rollout_probs_diff_std:0.010430659167468548 - training/rollout_actor_probs_pearson_corr:0.998622477054596 - self_distillation/success_group_fraction:0.75 - self_distillation/success_sample_fraction:0.74609375 - self_distillation/correct_sample_fraction:0.6640625 - self_distillation/feedback_available_fraction:0.0 - self_distillation/feedback_used_fraction:0.0 - self_distillation/reprompt_sample_fraction:0.74609375 - rollout_corr/rollout_is_mean:1.0001773834228516 - rollout_corr/rollout_is_ratio_fraction_high:0.0 - rollout_corr/rollout_is_ratio_fraction_low:0.0 - rollout_corr/rollout_is_max:1.939037799835205 - rollout_corr/rollout_is_min:0.5226768255233765 - rollout_corr/rollout_is_std:0.03018668293952942 - rollout_corr/rollout_is_eff_sample_size:0.9990899507625369 - rollout_corr/rollout_is_seq_mean:1.0002132654190063 - rollout_corr/rollout_is_seq_std:0.0017920753452926874 - rollout_corr/rollout_is_seq_max:1.00586998462677 - rollout_corr/rollout_is_seq_min:0.9940651059150696 - rollout_corr/rollout_is_seq_max_deviation:0.00593489408493042 - rollout_corr/rollout_is_seq_fraction_high:0.0 - rollout_corr/rollout_is_seq_fraction_low:0.0 - rollout_corr/training_ppl:np.float64(1.177405840717256) - rollout_corr/training_log_ppl:np.float64(0.1596459786160267) - rollout_corr/kl:np.float64(0.000444874939256934) - rollout_corr/k3_kl:np.float64(0.0005352445203357092) - rollout_corr/rollout_ppl:np.float64(1.1769076697528362) - rollout_corr/rollout_log_ppl:np.float64(0.15920110348815797) - rollout_corr/log_ppl_diff:np.float64(0.00044487512786872685) - rollout_corr/log_ppl_abs_diff:np.float64(0.0015562142943963408) - rollout_corr/log_ppl_diff_max:np.float64(0.008173204958438873) - rollout_corr/log_ppl_diff_min:np.float64(-0.00523337721824646) - rollout_corr/ppl_ratio:np.float64(1.0004469151608646) - rollout_corr/chi2_token:np.float64(0.0012778392992913723) - rollout_corr/chi2_seq:np.float64(0.631908955052495) - actor/pg_loss:np.float64(0.001201812643557787) - actor/kl_loss:np.float64(0.0) - actor/pg_clipfrac:np.float64(0.00034777516520989593) - actor/ppo_kl:np.float64(0.0001929089082366442) - actor/pg_clipfrac_lower:np.float64(0.0) - self_distillation/token_reweight_lambda:np.float64(0.5) - self_distillation/token_reweight_eps_w:np.float64(0.20000000000000004) - self_distillation/token_reweight_active_token_fraction:np.float64(0.74609375) - self_distillation/token_reweight_context_sample_fraction:np.float64(0.74609375) - self_distillation/token_reweight_correct_sample_fraction:np.float64(0.6640625) - self_distillation/token_reweight_w_mean:np.float64(33618.56932144542) - self_distillation/token_reweight_w_std:np.float64(630104.3924066984) - self_distillation/token_reweight_w_clipped_mean:np.float64(0.9997440660372376) - self_distillation/token_reweight_w_clip_frac:np.float64(0.027509495565027464) - self_distillation/empty_target_batch:np.float64(0.25390625) - actor/grad_norm:np.float64(0.1583699844777584) - perf/mfu/actor:np.float64(0.05528347187737279) - perf/max_memory_allocated_gb:np.float64(130.98326301574707) - perf/max_memory_reserved_gb:np.float64(140.86328125) - perf/cpu_memory_used_gb:np.float64(195.58264541625977) - actor/lr:np.float64(1e-06) - training/global_step:47 - training/epoch:1 - critic/score/mean:0.6640625 - critic/score/max:1.0 - critic/score/min:0.0 - critic/rewards/mean:0.6640625 - critic/rewards/max:1.0 - critic/rewards/min:0.0 - critic/advantages/mean:-0.002365098800510168 - critic/advantages/max:0.875 - critic/advantages/min:-0.875 - critic/returns/mean:-0.002365098800510168 - critic/returns/max:0.875 - critic/returns/min:-0.875 - response_length/mean:372.234375 - response_length/max:590.0 - response_length/mi
|
||
step:48 - global_seqlen/min:21150 - global_seqlen/max:26798 - global_seqlen/minmax_diff:5648 - global_seqlen/balanced_min:24129 - global_seqlen/balanced_max:24132 - global_seqlen/mean:24130.75 - actor/entropy:0.14493174850940704 - perf/mfu/actor_infer:0 - training/rollout_probs_diff_valid:1 - training/rollout_probs_diff_max:0.5314710736274719 - training/rollout_probs_diff_mean:0.002818688051775098 - training/rollout_probs_diff_std:0.010344676673412323 - training/rollout_actor_probs_pearson_corr:0.9985598921775818 - self_distillation/success_group_fraction:0.90625 - self_distillation/success_sample_fraction:0.90234375 - self_distillation/correct_sample_fraction:0.84375 - self_distillation/feedback_available_fraction:0.0 - self_distillation/feedback_used_fraction:0.0 - self_distillation/reprompt_sample_fraction:0.90234375 - rollout_corr/rollout_is_mean:1.0001366138458252 - rollout_corr/rollout_is_ratio_fraction_high:2.1507688870769925e-05 - rollout_corr/rollout_is_ratio_fraction_low:2.1507688870769925e-05 - rollout_corr/rollout_is_max:3.622957468032837 - rollout_corr/rollout_is_min:0.05226994305849075 - rollout_corr/rollout_is_std:0.02917659282684326 - rollout_corr/rollout_is_eff_sample_size:0.999149688526822 - rollout_corr/rollout_is_seq_mean:1.000110387802124 - rollout_corr/rollout_is_seq_std:0.0019526410615071654 - rollout_corr/rollout_is_seq_max:1.0087071657180786 - rollout_corr/rollout_is_seq_min:0.9937211275100708 - rollout_corr/rollout_is_seq_max_deviation:0.008707165718078613 - rollout_corr/rollout_is_seq_fraction_high:0.0 - rollout_corr/rollout_is_seq_fraction_low:0.0 - rollout_corr/training_ppl:np.float64(1.1530222110450268) - rollout_corr/training_log_ppl:np.float64(0.14057794290420134) - rollout_corr/kl:np.float64(0.00048637970533782493) - rollout_corr/k3_kl:np.float64(0.0006143509694993554) - rollout_corr/rollout_ppl:np.float64(1.1524676536209881) - rollout_corr/rollout_log_ppl:np.float64(0.14009156363317743) - rollout_corr/log_ppl_diff:np.float64(0.00048637927102390677) - rollout_corr/log_ppl_abs_diff:np.float64(0.0016521671932423487) - rollout_corr/log_ppl_diff_max:np.float64(0.007564276456832886) - rollout_corr/log_ppl_diff_min:np.float64(-0.007610246539115906) - rollout_corr/ppl_ratio:np.float64(1.000488800695166) - rollout_corr/chi2_token:np.float64(0.001616457011550665) - rollout_corr/chi2_seq:np.float64(0.7807452781125903) - actor/pg_loss:np.float64(0.000730266678147018) - actor/kl_loss:np.float64(0.0) - actor/pg_clipfrac:np.float64(0.0002534931982154376) - actor/ppo_kl:np.float64(0.00012199131381329664) - actor/pg_clipfrac_lower:np.float64(0.0) - self_distillation/token_reweight_lambda:np.float64(0.5) - self_distillation/token_reweight_eps_w:np.float64(0.20000000000000004) - self_distillation/token_reweight_active_token_fraction:np.float64(0.90234375) - self_distillation/token_reweight_context_sample_fraction:np.float64(0.90234375) - self_distillation/token_reweight_correct_sample_fraction:np.float64(0.84375) - self_distillation/token_reweight_w_mean:np.float64(31238.350169518264) - self_distillation/token_reweight_w_std:np.float64(474846.4062488966) - self_distillation/token_reweight_w_clipped_mean:np.float64(0.9999360444489866) - self_distillation/token_reweight_w_clip_frac:np.float64(0.017788645636755973) - self_distillation/empty_target_batch:np.float64(0.09765625) - actor/grad_norm:np.float64(0.18741019908338785) - perf/mfu/actor:np.float64(0.054881577129153286) - perf/max_memory_allocated_gb:np.float64(130.98326301574707) - perf/max_memory_reserved_gb:np.float64(140.86328125) - perf/cpu_memory_used_gb:np.float64(195.60995864868164) - actor/lr:np.float64(1e-06) - training/global_step:48 - training/epoch:1 - critic/score/mean:0.84375 - critic/score/max:1.0 - critic/score/min:0.0 - critic/rewards/mean:0.84375 - critic/rewards/max:1.0 - critic/rewards/min:0.0 - critic/advantages/mean:0.0010646305745467544 - critic/advantages/max:0.875 - critic/advantages/min:-0.875 - critic/returns/mean:0.0010646305745467544 - critic/returns/max:0.875 - critic/returns/min:-0.875 - response_length/mean:363.2421875 - r
|
||
step:49 - global_seqlen/min:21430 - global_seqlen/max:27560 - global_seqlen/minmax_diff:6130 - global_seqlen/balanced_min:24100 - global_seqlen/balanced_max:24104 - global_seqlen/mean:24102.75 - actor/entropy:0.15728282928466797 - perf/mfu/actor_infer:0 - training/rollout_probs_diff_valid:1 - training/rollout_probs_diff_max:0.277299165725708 - training/rollout_probs_diff_mean:0.002870401367545128 - training/rollout_probs_diff_std:0.009825697168707848 - training/rollout_actor_probs_pearson_corr:0.9987659454345703 - self_distillation/success_group_fraction:0.8125 - self_distillation/success_sample_fraction:0.8046875 - self_distillation/correct_sample_fraction:0.66796875 - self_distillation/feedback_available_fraction:0.0 - self_distillation/feedback_used_fraction:0.0 - self_distillation/reprompt_sample_fraction:0.8046875 - rollout_corr/rollout_is_mean:1.0000709295272827 - rollout_corr/rollout_is_ratio_fraction_high:2.1716941773775034e-05 - rollout_corr/rollout_is_ratio_fraction_low:0.0 - rollout_corr/rollout_is_max:2.1670196056365967 - rollout_corr/rollout_is_min:0.593317985534668 - rollout_corr/rollout_is_std:0.028590524569153786 - rollout_corr/rollout_is_eff_sample_size:0.9991836065722173 - rollout_corr/rollout_is_seq_mean:1.0001214742660522 - rollout_corr/rollout_is_seq_std:0.0017916224896907806 - rollout_corr/rollout_is_seq_max:1.0051995515823364 - rollout_corr/rollout_is_seq_min:0.9944879412651062 - rollout_corr/rollout_is_seq_max_deviation:0.005512058734893799 - rollout_corr/rollout_is_seq_fraction_high:0.0 - rollout_corr/rollout_is_seq_fraction_low:0.0 - rollout_corr/training_ppl:np.float64(1.1734219631180167) - rollout_corr/training_log_ppl:np.float64(0.15717786941240774) - rollout_corr/kl:np.float64(0.0005783919656501268) - rollout_corr/k3_kl:np.float64(0.0004959159838193727) - rollout_corr/rollout_ppl:np.float64(1.1727482806891203) - rollout_corr/rollout_log_ppl:np.float64(0.15659947633685078) - rollout_corr/log_ppl_diff:np.float64(0.00057839307555696) - rollout_corr/log_ppl_abs_diff:np.float64(0.0014631202866439708) - rollout_corr/log_ppl_diff_max:np.float64(0.008148357272148132) - rollout_corr/log_ppl_diff_min:np.float64(-0.004555851221084595) - rollout_corr/ppl_ratio:np.float64(1.0005802374798805) - rollout_corr/chi2_token:np.float64(0.0008415062911808491) - rollout_corr/chi2_seq:np.float64(0.23110721190460026) - actor/pg_loss:np.float64(0.00141685560811311) - actor/kl_loss:np.float64(0.0) - actor/pg_clipfrac:np.float64(0.00017184585794893792) - actor/ppo_kl:np.float64(0.00026630591867160547) - actor/pg_clipfrac_lower:np.float64(0.0) - self_distillation/token_reweight_lambda:np.float64(0.5) - self_distillation/token_reweight_eps_w:np.float64(0.20000000000000004) - self_distillation/token_reweight_active_token_fraction:np.float64(0.8046875) - self_distillation/token_reweight_context_sample_fraction:np.float64(0.8046875) - self_distillation/token_reweight_correct_sample_fraction:np.float64(0.66796875) - self_distillation/token_reweight_w_mean:np.float64(18197.53233134863) - self_distillation/token_reweight_w_std:np.float64(351916.0242279185) - self_distillation/token_reweight_w_clipped_mean:np.float64(1.0004864516668022) - self_distillation/token_reweight_w_clip_frac:np.float64(0.03428729544975795) - self_distillation/empty_target_batch:np.float64(0.1953125) - actor/grad_norm:np.float64(0.20876117795705795) - perf/mfu/actor:np.float64(0.05458018329697781) - perf/max_memory_allocated_gb:np.float64(130.98326301574707) - perf/max_memory_reserved_gb:np.float64(140.86328125) - perf/cpu_memory_used_gb:np.float64(195.66826248168945) - actor/lr:np.float64(1e-06) - training/global_step:49 - training/epoch:1 - critic/score/mean:0.66796875 - critic/score/max:1.0 - critic/score/min:0.0 - critic/rewards/mean:0.66796875 - critic/rewards/max:1.0 - critic/rewards/min:0.0 - critic/advantages/mean:-0.00484559265896678 - critic/advantages/max:0.875 - critic/advantages/min:-0.875 - critic/returns/mean:-0.00484559265896678 - critic/returns/max:0.875 - critic/returns/min:-0.875 - response_length/mean:359.7421875 - response_length/max
|
||
test_gen_batch meta info: {'eos_token_id': 151645, 'pad_token_id': 151643, 'recompute_log_prob': False, 'do_sample': True, 'validate': True, 'global_steps': 50}
|
||
validation generation end
|
||
[prompt] system
|
||
|
||
Given a question and four options, please select the right answer. Respond in the following format:
|
||
<reasoning>
|
||
...
|
||
</reasoning>
|
||
<answer>
|
||
...
|
||
</answer>
|
||
|
||
For the answer, only output the letter corresponding to the correct option (A, B, C, or D), and nothing else. Do not restate the answer text. For example, if the answer is "A", just output:
|
||
<answer>
|
||
A
|
||
</answer>
|
||
|
||
user
|
||
Given the following crystal structure parameters for the material RbLa9(IrO6)4 (Material ID: mp-560657), calculate the volume of the unit cell (in ų).
|
||
|
||
Lattice Parameters:
|
||
|
||
a = 7.82 Å
|
||
b = 7.82 Å
|
||
c = 17.88 Å
|
||
α = 90.00º
|
||
β = 90.00º
|
||
γ = 90.00º
|
||
|
||
|
||
|
||
A: 1025.67 ų
|
||
|
||
B: 1094.31 ų
|
||
|
||
C: 1200.45 ų
|
||
|
||
D: 1150.78 ų
|
||
|
||
Please reason step by step.
|
||
assistant
|
||
<think>
|
||
|
||
</think>
|
||
|
||
|
||
[response] <reasoning>
|
||
The given crystal structure has lattice parameters a = 7.82 Å, b = 7.82 Å, c = 17.88 Å, and all angles α, β, γ are 90.00°, which indicates that the unit cell is orthorhombic.
|
||
|
||
For an orthorhombic unit cell, the volume $ V $ of the unit cell is calculated using the formula:
|
||
|
||
$$ V = a \times b \times c $$
|
||
|
||
Substituting the given values:
|
||
|
||
$$ V = 7.82 \times 7.82 \times 17.88 $$
|
||
|
||
First, calculate $ 7.82 \times 7.82 $:
|
||
|
||
$$ 7.82 \times 7.82 = 61.1524 $$
|
||
|
||
Then multiply by 17.88:
|
||
|
||
$$ 61.1524 \times 17.88 = 1094.31 $$
|
||
|
||
Thus, the volume of the unit cell is approximately 1094.31 ų.
|
||
</reasoning>
|
||
<answer>
|
||
B
|
||
</answer>
|
||
[ground_truth] B
|
||
[score] 1.0
|
||
[acc] 1.0
|
||
[pred] B
|
||
[incorrect_format] 1
|
||
[feedback]
|
||
local_global_step_folder: /mnt/mole/SDPO/L2T/checkpoints/datasets/sciknoweval/material/qwen3gen-material-RLSD_TR-Qwen-Qwen3-8B-mbs8-decay0-tr0.1-train32-rollout8-lr1e-6-vllm0.8/global_step_50
|
||
step:50 - global_seqlen/min:18129 - global_seqlen/max:28037 - global_seqlen/minmax_diff:9908 - global_seqlen/balanced_min:24010 - global_seqlen/balanced_max:24016 - global_seqlen/mean:24014.5 - actor/entropy:0.16035732626914978 - perf/mfu/actor_infer:0 - training/rollout_probs_diff_valid:1 - training/rollout_probs_diff_max:0.4744114577770233 - training/rollout_probs_diff_mean:0.0029135870281606913 - training/rollout_probs_diff_std:0.009974236600100994 - training/rollout_actor_probs_pearson_corr:0.998715877532959 - self_distillation/success_group_fraction:0.8125 - self_distillation/success_sample_fraction:0.80859375 - self_distillation/correct_sample_fraction:0.7265625 - self_distillation/feedback_available_fraction:0.0 - self_distillation/feedback_used_fraction:0.0 - self_distillation/reprompt_sample_fraction:0.80859375 - rollout_corr/rollout_is_mean:0.9999343156814575 - rollout_corr/rollout_is_ratio_fraction_high:1.1190188160981052e-05 - rollout_corr/rollout_is_ratio_fraction_low:1.1190188160981052e-05 - rollout_corr/rollout_is_max:2.2565836906433105 - rollout_corr/rollout_is_min:0.2993377447128296 - rollout_corr/rollout_is_std:0.02937408536672592 - rollout_corr/rollout_is_eff_sample_size:0.9991379070078227 - rollout_corr/rollout_is_seq_mean:0.9999151229858398 - rollout_corr/rollout_is_seq_std:0.0018135372083634138 - rollout_corr/rollout_is_seq_max:1.0075640678405762 - rollout_corr/rollout_is_seq_min:0.9934759140014648 - rollout_corr/rollout_is_seq_max_deviation:0.007564067840576172 - rollout_corr/rollout_is_seq_fraction_high:0.0 - rollout_corr/rollout_is_seq_fraction_low:0.0 - rollout_corr/training_ppl:np.float64(1.175829870160669) - rollout_corr/training_log_ppl:np.float64(0.1581116948655108) - rollout_corr/kl:np.float64(0.00059627767032433) - rollout_corr/k3_kl:np.float64(0.0005526660350341217) - rollout_corr/rollout_ppl:np.float64(1.1751363426446915) - rollout_corr/rollout_log_ppl:np.float64(0.15751541723147966) - rollout_corr/log_ppl_diff:np.float64(0.0005962776340311393) - rollout_corr/log_ppl_abs_diff:np.float64(0.0015397017996292561) - rollout_corr/log_ppl_diff_max:np.float64(0.007978282868862152) - rollout_corr/log_ppl_diff_min:np.float64(-0.0065612345933914185) - rollout_corr/ppl_ratio:np.float64(1.0005985118914396) - rollout_corr/chi2_token:np.float64(0.001080595189705491) - rollout_corr/chi2_seq:np.float64(0.42997400439344347) - actor/pg_loss:np.float64(0.0008469953900203109) - actor/kl_loss:np.float64(0.0) - actor/pg_clipfrac:np.float64(0.0002414540476820548) - actor/ppo_kl:np.float64(6.015790615965244e-05) - actor/pg_clipfrac_lower:np.float64(0.0) - self_distillation/token_reweight_lambda:np.float64(0.5) - self_distillation/token_reweight_eps_w:np.float64(0.20000000000000004) - self_distillation/token_reweight_active_token_fraction:np.float64(0.80859375) - self_distillation/token_reweight_context_sample_fraction:np.float64(0.80859375) - self_distillation/token_reweight_correct_sample_fraction:np.float64(0.7265625) - self_distillation/token_reweight_w_mean:np.float64(43890.94956536731) - self_distillation/token_reweight_w_std:np.float64(817201.1323379974) - self_distillation/token_reweight_w_clipped_mean:np.float64(0.9998418700415641) - self_distillation/token_reweight_w_clip_frac:np.float64(0.027138499077409506) - self_distillation/empty_target_batch:np.float64(0.19140625) - actor/grad_norm:np.float64(0.1494966372847557) - perf/mfu/actor:np.float64(0.05476425496633601) - perf/max_memory_allocated_gb:np.float64(130.98326301574707) - perf/max_memory_reserved_gb:np.float64(140.86328125) - perf/cpu_memory_used_gb:np.float64(195.49509620666504) - actor/lr:np.float64(1e-06) - val-aux/sciknoweval/reward/mean@16:np.float64(0.7466755319148937) - val-aux/sciknoweval/reward/std@16:np.float64(0.053407092795511396) - val-aux/sciknoweval/reward/best@2/mean:np.float64(0.7685744680851063) - val-aux/sciknoweval/reward/best@2/std:np.float64(0.04014169769484946) - val-aux/sciknoweval/reward/worst@2/mean:np.float64(0.7242765957446807) - val-aux/sciknoweval/reward/worst@2/std:np.float64(0.05065782360309899) - val-aux/sciknowev
|
||
step:51 - global_seqlen/min:20681 - global_seqlen/max:27967 - global_seqlen/minmax_diff:7286 - global_seqlen/balanced_min:23835 - global_seqlen/balanced_max:23840 - global_seqlen/mean:23838.25 - actor/entropy:0.1650850474834442 - perf/mfu/actor_infer:0 - training/rollout_probs_diff_valid:1 - training/rollout_probs_diff_max:0.2733956575393677 - training/rollout_probs_diff_mean:0.0027729461435228586 - training/rollout_probs_diff_std:0.009361513890326023 - training/rollout_actor_probs_pearson_corr:0.9989805221557617 - self_distillation/success_group_fraction:0.75 - self_distillation/success_sample_fraction:0.74609375 - self_distillation/correct_sample_fraction:0.66796875 - self_distillation/feedback_available_fraction:0.0 - self_distillation/feedback_used_fraction:0.0 - self_distillation/reprompt_sample_fraction:0.74609375 - rollout_corr/rollout_is_mean:0.9999463558197021 - rollout_corr/rollout_is_ratio_fraction_high:0.0 - rollout_corr/rollout_is_ratio_fraction_low:1.1035334864573088e-05 - rollout_corr/rollout_is_max:1.6391828060150146 - rollout_corr/rollout_is_min:0.492061048746109 - rollout_corr/rollout_is_std:0.02771829441189766 - rollout_corr/rollout_is_eff_sample_size:0.999232285966509 - rollout_corr/rollout_is_seq_mean:0.9999741315841675 - rollout_corr/rollout_is_seq_std:0.0017780638299882412 - rollout_corr/rollout_is_seq_max:1.0069881677627563 - rollout_corr/rollout_is_seq_min:0.992406964302063 - rollout_corr/rollout_is_seq_max_deviation:0.007593035697937012 - rollout_corr/rollout_is_seq_fraction_high:0.0 - rollout_corr/rollout_is_seq_fraction_low:0.0 - rollout_corr/training_ppl:np.float64(1.1927359052933753) - rollout_corr/training_log_ppl:np.float64(0.17125490526814247) - rollout_corr/kl:np.float64(0.00048572005391089856) - rollout_corr/k3_kl:np.float64(0.0004628497603818005) - rollout_corr/rollout_ppl:np.float64(1.1921314564533532) - rollout_corr/rollout_log_ppl:np.float64(0.17076918519160245) - rollout_corr/log_ppl_diff:np.float64(0.00048572007654001936) - rollout_corr/log_ppl_abs_diff:np.float64(0.0014220301018212922) - rollout_corr/log_ppl_diff_max:np.float64(0.009596168994903564) - rollout_corr/log_ppl_diff_min:np.float64(-0.006043359637260437) - rollout_corr/ppl_ratio:np.float64(1.0004876188468188) - rollout_corr/chi2_token:np.float64(0.0008928193710744381) - rollout_corr/chi2_seq:np.float64(0.36601235694251955) - actor/pg_loss:np.float64(0.0007545447442680597) - actor/kl_loss:np.float64(0.0) - actor/pg_clipfrac:np.float64(0.00012426727698766626) - actor/ppo_kl:np.float64(4.306889947258696e-05) - actor/pg_clipfrac_lower:np.float64(0.0) - self_distillation/token_reweight_lambda:np.float64(0.5) - self_distillation/token_reweight_eps_w:np.float64(0.20000000000000004) - self_distillation/token_reweight_active_token_fraction:np.float64(0.74609375) - self_distillation/token_reweight_context_sample_fraction:np.float64(0.74609375) - self_distillation/token_reweight_correct_sample_fraction:np.float64(0.66796875) - self_distillation/token_reweight_w_mean:np.float64(8465.622773182346) - self_distillation/token_reweight_w_std:np.float64(146480.68177438865) - self_distillation/token_reweight_w_clipped_mean:np.float64(1.0010926395189017) - self_distillation/token_reweight_w_clip_frac:np.float64(0.02168405632255599) - self_distillation/empty_target_batch:np.float64(0.25390625) - actor/grad_norm:np.float64(0.14392557181417942) - perf/mfu/actor:np.float64(0.053210625937775015) - perf/max_memory_allocated_gb:np.float64(130.98326301574707) - perf/max_memory_reserved_gb:np.float64(140.86328125) - perf/cpu_memory_used_gb:np.float64(196.1166114807129) - actor/lr:np.float64(1e-06) - training/global_step:51 - training/epoch:1 - critic/score/mean:0.66796875 - critic/score/max:1.0 - critic/score/min:0.0 - critic/rewards/mean:0.66796875 - critic/rewards/max:1.0 - critic/rewards/min:0.0 - critic/advantages/mean:-0.0029919552616775036 - critic/advantages/max:0.875 - critic/advantages/min:-0.75 - critic/returns/mean:-0.0029919552616775036 - critic/returns/max:0.875 - critic/returns/min:-0.75 - response_length/mean:353.9765625 - response_l
|
||
step:52 - global_seqlen/min:20929 - global_seqlen/max:28953 - global_seqlen/minmax_diff:8024 - global_seqlen/balanced_min:25185 - global_seqlen/balanced_max:25193 - global_seqlen/mean:25191.625 - actor/entropy:0.16280372440814972 - perf/mfu/actor_infer:0 - training/rollout_probs_diff_valid:1 - training/rollout_probs_diff_max:0.2151232361793518 - training/rollout_probs_diff_mean:0.0027323930989950895 - training/rollout_probs_diff_std:0.00925284344702959 - training/rollout_actor_probs_pearson_corr:0.9989708662033081 - self_distillation/success_group_fraction:0.875 - self_distillation/success_sample_fraction:0.8671875 - self_distillation/correct_sample_fraction:0.78515625 - self_distillation/feedback_available_fraction:0.0 - self_distillation/feedback_used_fraction:0.0 - self_distillation/reprompt_sample_fraction:0.8671875 - rollout_corr/rollout_is_mean:0.9999663829803467 - rollout_corr/rollout_is_ratio_fraction_high:0.0 - rollout_corr/rollout_is_ratio_fraction_low:1.0168698281631805e-05 - rollout_corr/rollout_is_max:1.868242859840393 - rollout_corr/rollout_is_min:0.3074229955673218 - rollout_corr/rollout_is_std:0.027021415531635284 - rollout_corr/rollout_is_eff_sample_size:0.9992702568062312 - rollout_corr/rollout_is_seq_mean:0.9999210238456726 - rollout_corr/rollout_is_seq_std:0.0015943794278427958 - rollout_corr/rollout_is_seq_max:1.005804181098938 - rollout_corr/rollout_is_seq_min:0.9952024221420288 - rollout_corr/rollout_is_seq_max_deviation:0.005804181098937988 - rollout_corr/rollout_is_seq_fraction_high:0.0 - rollout_corr/rollout_is_seq_fraction_low:0.0 - rollout_corr/training_ppl:np.float64(1.1793708237819374) - rollout_corr/training_log_ppl:np.float64(0.16213442391745048) - rollout_corr/kl:np.float64(0.0005378835334433063) - rollout_corr/k3_kl:np.float64(0.0005385009871474722) - rollout_corr/rollout_ppl:np.float64(1.1787225706502795) - rollout_corr/rollout_log_ppl:np.float64(0.1615965386881726) - rollout_corr/log_ppl_diff:np.float64(0.0005378852292778902) - rollout_corr/log_ppl_abs_diff:np.float64(0.0014847702332190238) - rollout_corr/log_ppl_diff_max:np.float64(0.00623573362827301) - rollout_corr/log_ppl_diff_min:np.float64(-0.004963576793670654) - rollout_corr/ppl_ratio:np.float64(1.000539786182344) - rollout_corr/chi2_token:np.float64(0.0010742407757788897) - rollout_corr/chi2_seq:np.float64(0.3871990798506886) - actor/pg_loss:np.float64(0.0007529193535447121) - actor/kl_loss:np.float64(0.0) - actor/pg_clipfrac:np.float64(0.0003004613563462044) - actor/ppo_kl:np.float64(6.476392650078822e-05) - actor/pg_clipfrac_lower:np.float64(0.0) - self_distillation/token_reweight_lambda:np.float64(0.5) - self_distillation/token_reweight_eps_w:np.float64(0.20000000000000004) - self_distillation/token_reweight_active_token_fraction:np.float64(0.8671875) - self_distillation/token_reweight_context_sample_fraction:np.float64(0.8671875) - self_distillation/token_reweight_correct_sample_fraction:np.float64(0.78515625) - self_distillation/token_reweight_w_mean:np.float64(14713.069492256269) - self_distillation/token_reweight_w_std:np.float64(269632.1662454984) - self_distillation/token_reweight_w_clipped_mean:np.float64(0.9998877602629364) - self_distillation/token_reweight_w_clip_frac:np.float64(0.030055424678721465) - self_distillation/empty_target_batch:np.float64(0.1328125) - actor/grad_norm:np.float64(0.14222942106425762) - perf/mfu/actor:np.float64(0.056224778008784046) - perf/max_memory_allocated_gb:np.float64(130.98326301574707) - perf/max_memory_reserved_gb:np.float64(140.86328125) - perf/cpu_memory_used_gb:np.float64(195.7989044189453) - actor/lr:np.float64(1e-06) - training/global_step:52 - training/epoch:1 - critic/score/mean:0.78515625 - critic/score/max:1.0 - critic/score/min:0.0 - critic/rewards/mean:0.78515625 - critic/rewards/max:1.0 - critic/rewards/min:0.0 - critic/advantages/mean:-0.003757334081456065 - critic/advantages/max:0.875 - critic/advantages/min:-0.875 - critic/returns/mean:-0.003757334081456065 - critic/returns/max:0.875 - critic/returns/min:-0.875 - response_length/mean:384.14453125 - response_length/
|
||
step:53 - global_seqlen/min:20480 - global_seqlen/max:24979 - global_seqlen/minmax_diff:4499 - global_seqlen/balanced_min:23009 - global_seqlen/balanced_max:23014 - global_seqlen/mean:23012.625 - actor/entropy:0.15226857364177704 - perf/mfu/actor_infer:0 - training/rollout_probs_diff_valid:1 - training/rollout_probs_diff_max:0.4149150848388672 - training/rollout_probs_diff_mean:0.0029795358423143625 - training/rollout_probs_diff_std:0.010255304165184498 - training/rollout_actor_probs_pearson_corr:0.9986225366592407 - self_distillation/success_group_fraction:0.71875 - self_distillation/success_sample_fraction:0.71484375 - self_distillation/correct_sample_fraction:0.64453125 - self_distillation/feedback_available_fraction:0.0 - self_distillation/feedback_used_fraction:0.0 - self_distillation/reprompt_sample_fraction:0.71484375 - rollout_corr/rollout_is_mean:1.0000274181365967 - rollout_corr/rollout_is_ratio_fraction_high:2.4199891413445584e-05 - rollout_corr/rollout_is_ratio_fraction_low:0.0 - rollout_corr/rollout_is_max:5.8867363929748535 - rollout_corr/rollout_is_min:0.612064003944397 - rollout_corr/rollout_is_std:0.028760967776179314 - rollout_corr/rollout_is_eff_sample_size:0.9991734904229588 - rollout_corr/rollout_is_seq_mean:0.9999820590019226 - rollout_corr/rollout_is_seq_std:0.0018618700560182333 - rollout_corr/rollout_is_seq_max:1.0082342624664307 - rollout_corr/rollout_is_seq_min:0.9929176568984985 - rollout_corr/rollout_is_seq_max_deviation:0.008234262466430664 - rollout_corr/rollout_is_seq_fraction_high:0.0 - rollout_corr/rollout_is_seq_fraction_low:0.0 - rollout_corr/training_ppl:np.float64(1.1658900491893291) - rollout_corr/training_log_ppl:np.float64(0.1514768806300708) - rollout_corr/kl:np.float64(0.0005496431590059636) - rollout_corr/k3_kl:np.float64(0.0004694314374376063) - rollout_corr/rollout_ppl:np.float64(1.1652461695484817) - rollout_corr/rollout_log_ppl:np.float64(0.15092723770067096) - rollout_corr/log_ppl_diff:np.float64(0.0005496429293998517) - rollout_corr/log_ppl_abs_diff:np.float64(0.001543470520118717) - rollout_corr/log_ppl_diff_max:np.float64(0.0066513121128082275) - rollout_corr/log_ppl_diff_min:np.float64(-0.004514351487159729) - rollout_corr/ppl_ratio:np.float64(1.0005516416858882) - rollout_corr/chi2_token:np.float64(0.0007865177467465401) - rollout_corr/chi2_seq:np.float64(0.43467385205440223) - actor/pg_loss:np.float64(0.0009989301906898618) - actor/kl_loss:np.float64(0.0) - actor/pg_clipfrac:np.float64(0.00015393912690342404) - actor/ppo_kl:np.float64(8.647114800908184e-05) - actor/pg_clipfrac_lower:np.float64(0.0) - self_distillation/token_reweight_lambda:np.float64(0.5) - self_distillation/token_reweight_eps_w:np.float64(0.20000000000000004) - self_distillation/token_reweight_active_token_fraction:np.float64(0.71484375) - self_distillation/token_reweight_context_sample_fraction:np.float64(0.71484375) - self_distillation/token_reweight_correct_sample_fraction:np.float64(0.64453125) - self_distillation/token_reweight_w_mean:np.float64(24308.17669938528) - self_distillation/token_reweight_w_std:np.float64(413428.8588547886) - self_distillation/token_reweight_w_clipped_mean:np.float64(0.9998141932301223) - self_distillation/token_reweight_w_clip_frac:np.float64(0.02757751938770525) - self_distillation/empty_target_batch:np.float64(0.28515625) - actor/grad_norm:np.float64(0.17244473472237587) - perf/mfu/actor:np.float64(0.05212055127204503) - perf/max_memory_allocated_gb:np.float64(130.98326301574707) - perf/max_memory_reserved_gb:np.float64(140.86328125) - perf/cpu_memory_used_gb:np.float64(195.67004013061523) - actor/lr:np.float64(1e-06) - training/global_step:53 - training/epoch:2 - critic/score/mean:0.64453125 - critic/score/max:1.0 - critic/score/min:0.0 - critic/rewards/mean:0.64453125 - critic/rewards/max:1.0 - critic/rewards/min:0.0 - critic/advantages/mean:0.0003796357777900994 - critic/advantages/max:0.875 - critic/advantages/min:-0.875 - critic/returns/mean:0.0003796357777900994 - critic/returns/max:0.875 - critic/returns/min:-0.875 - response_length/mean:322.83203125 - respo
|
||
step:54 - global_seqlen/min:20279 - global_seqlen/max:27363 - global_seqlen/minmax_diff:7084 - global_seqlen/balanced_min:25193 - global_seqlen/balanced_max:25196 - global_seqlen/mean:25194.875 - actor/entropy:0.15839405357837677 - perf/mfu/actor_infer:0 - training/rollout_probs_diff_valid:1 - training/rollout_probs_diff_max:0.30164676904678345 - training/rollout_probs_diff_mean:0.0026747898664325476 - training/rollout_probs_diff_std:0.009149343706667423 - training/rollout_actor_probs_pearson_corr:0.9989473223686218 - self_distillation/success_group_fraction:0.625 - self_distillation/success_sample_fraction:0.62109375 - self_distillation/correct_sample_fraction:0.57421875 - self_distillation/feedback_available_fraction:0.0 - self_distillation/feedback_used_fraction:0.0 - self_distillation/reprompt_sample_fraction:0.62109375 - rollout_corr/rollout_is_mean:0.9999547600746155 - rollout_corr/rollout_is_ratio_fraction_high:0.0 - rollout_corr/rollout_is_ratio_fraction_low:0.0 - rollout_corr/rollout_is_max:1.6425002813339233 - rollout_corr/rollout_is_min:0.5558498501777649 - rollout_corr/rollout_is_std:0.026692956686019897 - rollout_corr/rollout_is_eff_sample_size:0.9992879933909776 - rollout_corr/rollout_is_seq_mean:0.9999342560768127 - rollout_corr/rollout_is_seq_std:0.0015845621237531304 - rollout_corr/rollout_is_seq_max:1.0048959255218506 - rollout_corr/rollout_is_seq_min:0.9955595135688782 - rollout_corr/rollout_is_seq_max_deviation:0.004895925521850586 - rollout_corr/rollout_is_seq_fraction_high:0.0 - rollout_corr/rollout_is_seq_fraction_low:0.0 - rollout_corr/training_ppl:np.float64(1.1783004244789481) - rollout_corr/training_log_ppl:np.float64(0.1593520959140733) - rollout_corr/kl:np.float64(0.0006535952649509014) - rollout_corr/k3_kl:np.float64(0.0004820261234499412) - rollout_corr/rollout_ppl:np.float64(1.177526906132698) - rollout_corr/rollout_log_ppl:np.float64(0.15869850096351001) - rollout_corr/log_ppl_diff:np.float64(0.0006535949505632743) - rollout_corr/log_ppl_abs_diff:np.float64(0.0013726948818657547) - rollout_corr/log_ppl_diff_max:np.float64(0.0071489810943603516) - rollout_corr/log_ppl_diff_min:np.float64(-0.005629494786262512) - rollout_corr/ppl_ratio:np.float64(1.0006552611012012) - rollout_corr/chi2_token:np.float64(0.0006367254536598921) - rollout_corr/chi2_seq:np.float64(0.6323353212792426) - actor/pg_loss:np.float64(0.0005831223679706454) - actor/kl_loss:np.float64(0.0) - actor/pg_clipfrac:np.float64(0.00011379809757272596) - actor/ppo_kl:np.float64(0.00020858292008796298) - actor/pg_clipfrac_lower:np.float64(0.0) - self_distillation/token_reweight_lambda:np.float64(0.5) - self_distillation/token_reweight_eps_w:np.float64(0.20000000000000004) - self_distillation/token_reweight_active_token_fraction:np.float64(0.62109375) - self_distillation/token_reweight_context_sample_fraction:np.float64(0.62109375) - self_distillation/token_reweight_correct_sample_fraction:np.float64(0.57421875) - self_distillation/token_reweight_w_mean:np.float64(9313.207618546905) - self_distillation/token_reweight_w_std:np.float64(180056.0983797087) - self_distillation/token_reweight_w_clipped_mean:np.float64(0.999447213485837) - self_distillation/token_reweight_w_clip_frac:np.float64(0.019204000534955412) - self_distillation/empty_target_batch:np.float64(0.37890625) - actor/grad_norm:np.float64(0.10884904395788908) - perf/mfu/actor:np.float64(0.05665816464996518) - perf/max_memory_allocated_gb:np.float64(130.98326301574707) - perf/max_memory_reserved_gb:np.float64(140.86328125) - perf/cpu_memory_used_gb:np.float64(195.66030502319336) - actor/lr:np.float64(1e-06) - training/global_step:54 - training/epoch:2 - critic/score/mean:0.57421875 - critic/score/max:1.0 - critic/score/min:0.0 - critic/rewards/mean:0.57421875 - critic/rewards/max:1.0 - critic/rewards/min:0.0 - critic/advantages/mean:-0.0007078859489411116 - critic/advantages/max:0.875 - critic/advantages/min:-0.875 - critic/returns/mean:-0.0007078859489411116 - critic/returns/max:0.875 - critic/returns/min:-0.875 - response_length/mean:377.99609375 - response_length/max:91
|
||
step:55 - global_seqlen/min:22028 - global_seqlen/max:26786 - global_seqlen/minmax_diff:4758 - global_seqlen/balanced_min:24965 - global_seqlen/balanced_max:25170 - global_seqlen/mean:24992.625 - actor/entropy:0.15181651711463928 - perf/mfu/actor_infer:0 - training/rollout_probs_diff_valid:1 - training/rollout_probs_diff_max:0.17278850078582764 - training/rollout_probs_diff_mean:0.0025789199862629175 - training/rollout_probs_diff_std:0.008787320926785469 - training/rollout_actor_probs_pearson_corr:0.9990404844284058 - self_distillation/success_group_fraction:0.78125 - self_distillation/success_sample_fraction:0.78125 - self_distillation/correct_sample_fraction:0.7421875 - self_distillation/feedback_available_fraction:0.0 - self_distillation/feedback_used_fraction:0.0 - self_distillation/reprompt_sample_fraction:0.78125 - rollout_corr/rollout_is_mean:0.9999983310699463 - rollout_corr/rollout_is_ratio_fraction_high:0.0 - rollout_corr/rollout_is_ratio_fraction_low:1.0607941476337146e-05 - rollout_corr/rollout_is_max:1.6370457410812378 - rollout_corr/rollout_is_min:0.49754756689071655 - rollout_corr/rollout_is_std:0.02620846964418888 - rollout_corr/rollout_is_eff_sample_size:0.9993134685136027 - rollout_corr/rollout_is_seq_mean:0.9999988079071045 - rollout_corr/rollout_is_seq_std:0.0015600925544276834 - rollout_corr/rollout_is_seq_max:1.0062010288238525 - rollout_corr/rollout_is_seq_min:0.9956908226013184 - rollout_corr/rollout_is_seq_max_deviation:0.006201028823852539 - rollout_corr/rollout_is_seq_fraction_high:0.0 - rollout_corr/rollout_is_seq_fraction_low:0.0 - rollout_corr/training_ppl:np.float64(1.1677626050077379) - rollout_corr/training_log_ppl:np.float64(0.15257516079873312) - rollout_corr/kl:np.float64(0.00042375086632562287) - rollout_corr/k3_kl:np.float64(0.00040278795792403344) - rollout_corr/rollout_ppl:np.float64(1.1672604512423277) - rollout_corr/rollout_log_ppl:np.float64(0.15215140884538414) - rollout_corr/log_ppl_diff:np.float64(0.0004237519533489831) - rollout_corr/log_ppl_abs_diff:np.float64(0.0013195802093832754) - rollout_corr/log_ppl_diff_max:np.float64(0.005208849906921387) - rollout_corr/log_ppl_diff_min:np.float64(-0.00468212366104126) - rollout_corr/ppl_ratio:np.float64(1.000425181351602) - rollout_corr/chi2_token:np.float64(0.0007754352409392595) - rollout_corr/chi2_seq:np.float64(0.32551098824478686) - actor/pg_loss:np.float64(0.0008454262278974056) - actor/kl_loss:np.float64(0.0) - actor/pg_clipfrac:np.float64(7.39435245122877e-05) - actor/ppo_kl:np.float64(6.83178090241654e-05) - actor/pg_clipfrac_lower:np.float64(0.0) - self_distillation/token_reweight_lambda:np.float64(0.5) - self_distillation/token_reweight_eps_w:np.float64(0.20000000000000004) - self_distillation/token_reweight_active_token_fraction:np.float64(0.78125) - self_distillation/token_reweight_context_sample_fraction:np.float64(0.78125) - self_distillation/token_reweight_correct_sample_fraction:np.float64(0.7421875) - self_distillation/token_reweight_w_mean:np.float64(6225.795452147257) - self_distillation/token_reweight_w_std:np.float64(108468.37706527536) - self_distillation/token_reweight_w_clipped_mean:np.float64(0.9998256473336369) - self_distillation/token_reweight_w_clip_frac:np.float64(0.01854172901221318) - self_distillation/empty_target_batch:np.float64(0.21875) - actor/grad_norm:np.float64(0.12622096296399832) - perf/mfu/actor:np.float64(0.0559628797720936) - perf/max_memory_allocated_gb:np.float64(130.98332166671753) - perf/max_memory_reserved_gb:np.float64(140.86328125) - perf/cpu_memory_used_gb:np.float64(195.84178924560547) - actor/lr:np.float64(1e-06) - training/global_step:55 - training/epoch:2 - critic/score/mean:0.7421875 - critic/score/max:1.0 - critic/score/min:0.0 - critic/rewards/mean:0.7421875 - critic/rewards/max:1.0 - critic/rewards/min:0.0 - critic/advantages/mean:-0.0006364764412865043 - critic/advantages/max:0.5 - critic/advantages/min:-0.875 - critic/returns/mean:-0.0006364764412865043 - critic/returns/max:0.5 - critic/returns/min:-0.875 - response_length/mean:368.23828125 - response_length/max:1313.0
|
||
step:56 - global_seqlen/min:19002 - global_seqlen/max:28125 - global_seqlen/minmax_diff:9123 - global_seqlen/balanced_min:24289 - global_seqlen/balanced_max:24294 - global_seqlen/mean:24291.875 - actor/entropy:0.17489738762378693 - perf/mfu/actor_infer:0 - training/rollout_probs_diff_valid:1 - training/rollout_probs_diff_max:0.28977641463279724 - training/rollout_probs_diff_mean:0.0027698748745024204 - training/rollout_probs_diff_std:0.009292028844356537 - training/rollout_actor_probs_pearson_corr:0.9989330768585205 - self_distillation/success_group_fraction:0.84375 - self_distillation/success_sample_fraction:0.84375 - self_distillation/correct_sample_fraction:0.765625 - self_distillation/feedback_available_fraction:0.0 - self_distillation/feedback_used_fraction:0.0 - self_distillation/reprompt_sample_fraction:0.84375 - rollout_corr/rollout_is_mean:0.9998672604560852 - rollout_corr/rollout_is_ratio_fraction_high:0.0 - rollout_corr/rollout_is_ratio_fraction_low:2.068701542157214e-05 - rollout_corr/rollout_is_max:1.6370550394058228 - rollout_corr/rollout_is_min:0.3381432890892029 - rollout_corr/rollout_is_std:0.02769678272306919 - rollout_corr/rollout_is_eff_sample_size:0.9992333572045602 - rollout_corr/rollout_is_seq_mean:0.9998617768287659 - rollout_corr/rollout_is_seq_std:0.0015884299064055085 - rollout_corr/rollout_is_seq_max:1.0066787004470825 - rollout_corr/rollout_is_seq_min:0.994590699672699 - rollout_corr/rollout_is_seq_max_deviation:0.0066787004470825195 - rollout_corr/rollout_is_seq_fraction_high:0.0 - rollout_corr/rollout_is_seq_fraction_low:0.0 - rollout_corr/training_ppl:np.float64(1.1882359734736383) - rollout_corr/training_log_ppl:np.float64(0.16863422583264764) - rollout_corr/kl:np.float64(0.0004804436645375887) - rollout_corr/k3_kl:np.float64(0.00045078001338083595) - rollout_corr/rollout_ppl:np.float64(1.1876508300192654) - rollout_corr/rollout_log_ppl:np.float64(0.16815378133469494) - rollout_corr/log_ppl_diff:np.float64(0.0004804444979527034) - rollout_corr/log_ppl_abs_diff:np.float64(0.0013180674068280496) - rollout_corr/log_ppl_diff_max:np.float64(0.005176126956939697) - rollout_corr/log_ppl_diff_min:np.float64(-0.006251998245716095) - rollout_corr/ppl_ratio:np.float64(1.0004819862078875) - rollout_corr/chi2_token:np.float64(0.0008565529715269804) - rollout_corr/chi2_seq:np.float64(0.26315169618465006) - actor/pg_loss:np.float64(0.0010086237452924252) - actor/kl_loss:np.float64(0.0) - actor/pg_clipfrac:np.float64(0.00010706515604397282) - actor/ppo_kl:np.float64(-6.85965127367183e-05) - actor/pg_clipfrac_lower:np.float64(0.0) - self_distillation/token_reweight_lambda:np.float64(0.5) - self_distillation/token_reweight_eps_w:np.float64(0.20000000000000004) - self_distillation/token_reweight_active_token_fraction:np.float64(0.84375) - self_distillation/token_reweight_context_sample_fraction:np.float64(0.84375) - self_distillation/token_reweight_correct_sample_fraction:np.float64(0.765625) - self_distillation/token_reweight_w_mean:np.float64(18737.445998198586) - self_distillation/token_reweight_w_std:np.float64(351841.7787493873) - self_distillation/token_reweight_w_clipped_mean:np.float64(1.0000774189829826) - self_distillation/token_reweight_w_clip_frac:np.float64(0.019534573119017296) - self_distillation/empty_target_batch:np.float64(0.15625) - actor/grad_norm:np.float64(0.17343188263475895) - perf/mfu/actor:np.float64(0.05484556116620652) - perf/max_memory_allocated_gb:np.float64(130.98332166671753) - perf/max_memory_reserved_gb:np.float64(140.86328125) - perf/cpu_memory_used_gb:np.float64(195.69879531860352) - actor/lr:np.float64(1e-06) - training/global_step:56 - training/epoch:2 - critic/score/mean:0.765625 - critic/score/max:1.0 - critic/score/min:0.0 - critic/rewards/mean:0.765625 - critic/rewards/max:1.0 - critic/rewards/min:0.0 - critic/advantages/mean:0.003864593105390668 - critic/advantages/max:0.75 - critic/advantages/min:-0.625 - critic/returns/mean:0.003864593105390668 - critic/returns/max:0.75 - critic/returns/min:-0.625 - response_length/mean:377.65234375 - response_length/max:863.0 -
|
||
step:57 - global_seqlen/min:22180 - global_seqlen/max:26730 - global_seqlen/minmax_diff:4550 - global_seqlen/balanced_min:24761 - global_seqlen/balanced_max:24766 - global_seqlen/mean:24764.375 - actor/entropy:0.18208277225494385 - perf/mfu/actor_infer:0 - training/rollout_probs_diff_valid:1 - training/rollout_probs_diff_max:0.380797803401947 - training/rollout_probs_diff_mean:0.0030766604468226433 - training/rollout_probs_diff_std:0.009836580604314804 - training/rollout_actor_probs_pearson_corr:0.9989367127418518 - self_distillation/success_group_fraction:0.84375 - self_distillation/success_sample_fraction:0.84375 - self_distillation/correct_sample_fraction:0.765625 - self_distillation/feedback_available_fraction:0.0 - self_distillation/feedback_used_fraction:0.0 - self_distillation/reprompt_sample_fraction:0.84375 - rollout_corr/rollout_is_mean:1.000044822692871 - rollout_corr/rollout_is_ratio_fraction_high:1.0690728231566027e-05 - rollout_corr/rollout_is_ratio_fraction_low:2.1381456463132054e-05 - rollout_corr/rollout_is_max:4.1945343017578125 - rollout_corr/rollout_is_min:0.13594895601272583 - rollout_corr/rollout_is_std:0.028684185817837715 - rollout_corr/rollout_is_eff_sample_size:0.9991777748846973 - rollout_corr/rollout_is_seq_mean:1.000114917755127 - rollout_corr/rollout_is_seq_std:0.0016589241567999125 - rollout_corr/rollout_is_seq_max:1.0065020322799683 - rollout_corr/rollout_is_seq_min:0.9957323670387268 - rollout_corr/rollout_is_seq_max_deviation:0.006502032279968262 - rollout_corr/rollout_is_seq_fraction_high:0.0 - rollout_corr/rollout_is_seq_fraction_low:0.0 - rollout_corr/training_ppl:np.float64(1.1985374288633466) - rollout_corr/training_log_ppl:np.float64(0.17924344843777362) - rollout_corr/kl:np.float64(0.0005058041924144163) - rollout_corr/k3_kl:np.float64(0.000529215224418067) - rollout_corr/rollout_ppl:np.float64(1.197921092621982) - rollout_corr/rollout_log_ppl:np.float64(0.17873764344403753) - rollout_corr/log_ppl_diff:np.float64(0.000505804993736092) - rollout_corr/log_ppl_abs_diff:np.float64(0.0013525268659577705) - rollout_corr/log_ppl_diff_max:np.float64(0.005043789744377136) - rollout_corr/log_ppl_diff_min:np.float64(-0.0049214959144592285) - rollout_corr/ppl_ratio:np.float64(1.0005072951316833) - rollout_corr/chi2_token:np.float64(0.0011630894150584936) - rollout_corr/chi2_seq:np.float64(0.3350441094953567) - actor/pg_loss:np.float64(0.0012816599337384105) - actor/kl_loss:np.float64(0.0) - actor/pg_clipfrac:np.float64(0.00018717671855483786) - actor/ppo_kl:np.float64(0.00018078757759099062) - actor/pg_clipfrac_lower:np.float64(0.0) - self_distillation/token_reweight_lambda:np.float64(0.5) - self_distillation/token_reweight_eps_w:np.float64(0.20000000000000004) - self_distillation/token_reweight_active_token_fraction:np.float64(0.84375) - self_distillation/token_reweight_context_sample_fraction:np.float64(0.84375) - self_distillation/token_reweight_correct_sample_fraction:np.float64(0.765625) - self_distillation/token_reweight_w_mean:np.float64(18.977482604794204) - self_distillation/token_reweight_w_std:np.float64(325.0625329564209) - self_distillation/token_reweight_w_clipped_mean:np.float64(0.9999024558346719) - self_distillation/token_reweight_w_clip_frac:np.float64(0.027874489809619263) - self_distillation/empty_target_batch:np.float64(0.15625) - actor/grad_norm:np.float64(0.15066221542656422) - perf/mfu/actor:np.float64(0.054658498083677315) - perf/max_memory_allocated_gb:np.float64(130.98332166671753) - perf/max_memory_reserved_gb:np.float64(140.86328125) - perf/cpu_memory_used_gb:np.float64(195.6921157836914) - actor/lr:np.float64(1e-06) - training/global_step:57 - training/epoch:2 - critic/score/mean:0.765625 - critic/score/max:1.0 - critic/score/min:0.0 - critic/rewards/mean:0.765625 - critic/rewards/max:1.0 - critic/rewards/min:0.0 - critic/advantages/mean:-0.004766728263348341 - critic/advantages/max:0.625 - critic/advantages/min:-0.875 - critic/returns/mean:-0.004766728263348341 - critic/returns/max:0.625 - critic/returns/min:-0.875 - response_length/mean:365.38671875 - response_
|
||
step:58 - global_seqlen/min:22489 - global_seqlen/max:27290 - global_seqlen/minmax_diff:4801 - global_seqlen/balanced_min:24464 - global_seqlen/balanced_max:24473 - global_seqlen/mean:24470.125 - actor/entropy:0.16469363868236542 - perf/mfu/actor_infer:0 - training/rollout_probs_diff_valid:1 - training/rollout_probs_diff_max:0.244918555021286 - training/rollout_probs_diff_mean:0.0026008158456534147 - training/rollout_probs_diff_std:0.008907139301300049 - training/rollout_actor_probs_pearson_corr:0.9990557432174683 - self_distillation/success_group_fraction:0.875 - self_distillation/success_sample_fraction:0.87109375 - self_distillation/correct_sample_fraction:0.8046875 - self_distillation/feedback_available_fraction:0.0 - self_distillation/feedback_used_fraction:0.0 - self_distillation/reprompt_sample_fraction:0.87109375 - rollout_corr/rollout_is_mean:0.9998775720596313 - rollout_corr/rollout_is_ratio_fraction_high:2.0434645193745382e-05 - rollout_corr/rollout_is_ratio_fraction_low:2.0434645193745382e-05 - rollout_corr/rollout_is_max:3.076564073562622 - rollout_corr/rollout_is_min:5.316741589922458e-05 - rollout_corr/rollout_is_std:0.026616927236318588 - rollout_corr/rollout_is_eff_sample_size:0.9992918026725187 - rollout_corr/rollout_is_seq_mean:0.9999071955680847 - rollout_corr/rollout_is_seq_std:0.001745687099173665 - rollout_corr/rollout_is_seq_max:1.0125819444656372 - rollout_corr/rollout_is_seq_min:0.9947338700294495 - rollout_corr/rollout_is_seq_max_deviation:0.012581944465637207 - rollout_corr/rollout_is_seq_fraction_high:0.0 - rollout_corr/rollout_is_seq_fraction_low:0.0 - rollout_corr/training_ppl:np.float64(1.1783719179220498) - rollout_corr/training_log_ppl:np.float64(0.16035878341062926) - rollout_corr/kl:np.float64(0.0006153967960749185) - rollout_corr/k3_kl:np.float64(0.00047639863944937133) - rollout_corr/rollout_ppl:np.float64(1.177631719969213) - rollout_corr/rollout_log_ppl:np.float64(0.1597433851493406) - rollout_corr/log_ppl_diff:np.float64(0.0006153982612886466) - rollout_corr/log_ppl_abs_diff:np.float64(0.0013544214962166734) - rollout_corr/log_ppl_diff_max:np.float64(0.017868779599666595) - rollout_corr/log_ppl_diff_min:np.float64(-0.0059568434953689575) - rollout_corr/ppl_ratio:np.float64(1.0006175043527037) - rollout_corr/chi2_token:np.float64(0.0005664050113409758) - rollout_corr/chi2_seq:np.float64(0.14223428606055677) - actor/pg_loss:np.float64(0.00035006296820938587) - actor/kl_loss:np.float64(0.0) - actor/pg_clipfrac:np.float64(9.173735679723904e-05) - actor/ppo_kl:np.float64(6.755156556303987e-05) - actor/pg_clipfrac_lower:np.float64(0.0) - self_distillation/token_reweight_lambda:np.float64(0.5) - self_distillation/token_reweight_eps_w:np.float64(0.20000000000000004) - self_distillation/token_reweight_active_token_fraction:np.float64(0.87109375) - self_distillation/token_reweight_context_sample_fraction:np.float64(0.87109375) - self_distillation/token_reweight_correct_sample_fraction:np.float64(0.8046875) - self_distillation/token_reweight_w_mean:np.float64(2222.0714181936346) - self_distillation/token_reweight_w_std:np.float64(39239.9520744278) - self_distillation/token_reweight_w_clipped_mean:np.float64(1.0001498765777797) - self_distillation/token_reweight_w_clip_frac:np.float64(0.01538468204671517) - self_distillation/empty_target_batch:np.float64(0.12890625) - actor/grad_norm:np.float64(0.09138654079288244) - perf/mfu/actor:np.float64(0.055187680810946424) - perf/max_memory_allocated_gb:np.float64(130.98332166671753) - perf/max_memory_reserved_gb:np.float64(140.86328125) - perf/cpu_memory_used_gb:np.float64(195.6977996826172) - actor/lr:np.float64(1e-06) - training/global_step:58 - training/epoch:2 - critic/score/mean:0.8046875 - critic/score/max:1.0 - critic/score/min:0.0 - critic/rewards/mean:0.8046875 - critic/rewards/max:1.0 - critic/rewards/min:0.0 - critic/advantages/mean:0.0035351936239749193 - critic/advantages/max:0.875 - critic/advantages/min:-0.875 - critic/returns/mean:0.0035351936239749193 - critic/returns/max:0.875 - critic/returns/min:-0.875 - response_length/mean:382.3
|
||
step:59 - global_seqlen/min:20196 - global_seqlen/max:26982 - global_seqlen/minmax_diff:6786 - global_seqlen/balanced_min:23296 - global_seqlen/balanced_max:23301 - global_seqlen/mean:23299.125 - actor/entropy:0.16596408188343048 - perf/mfu/actor_infer:0 - training/rollout_probs_diff_valid:1 - training/rollout_probs_diff_max:0.30163905024528503 - training/rollout_probs_diff_mean:0.0027721028309315443 - training/rollout_probs_diff_std:0.009304988197982311 - training/rollout_actor_probs_pearson_corr:0.9989809989929199 - self_distillation/success_group_fraction:0.75 - self_distillation/success_sample_fraction:0.75 - self_distillation/correct_sample_fraction:0.7265625 - self_distillation/feedback_available_fraction:0.0 - self_distillation/feedback_used_fraction:0.0 - self_distillation/reprompt_sample_fraction:0.75 - rollout_corr/rollout_is_mean:1.0000334978103638 - rollout_corr/rollout_is_ratio_fraction_high:0.0 - rollout_corr/rollout_is_ratio_fraction_low:1.1246948815823998e-05 - rollout_corr/rollout_is_max:1.9688125848770142 - rollout_corr/rollout_is_min:0.4614606499671936 - rollout_corr/rollout_is_std:0.027634304016828537 - rollout_corr/rollout_is_eff_sample_size:0.9992369280146525 - rollout_corr/rollout_is_seq_mean:1.0000215768814087 - rollout_corr/rollout_is_seq_std:0.0015595081495121121 - rollout_corr/rollout_is_seq_max:1.004768967628479 - rollout_corr/rollout_is_seq_min:0.9953278303146362 - rollout_corr/rollout_is_seq_max_deviation:0.004768967628479004 - rollout_corr/rollout_is_seq_fraction_high:0.0 - rollout_corr/rollout_is_seq_fraction_low:0.0 - rollout_corr/training_ppl:np.float64(1.1875210818834603) - rollout_corr/training_log_ppl:np.float64(0.1683449064294109) - rollout_corr/kl:np.float64(0.0006163650541246568) - rollout_corr/k3_kl:np.float64(0.0004912084286274876) - rollout_corr/rollout_ppl:np.float64(1.1867791321128607) - rollout_corr/rollout_log_ppl:np.float64(0.16772853956354083) - rollout_corr/log_ppl_diff:np.float64(0.0006163668658700772) - rollout_corr/log_ppl_abs_diff:np.float64(0.0014313266074168496) - rollout_corr/log_ppl_diff_max:np.float64(0.0071097612380981445) - rollout_corr/log_ppl_diff_min:np.float64(-0.004202127456665039) - rollout_corr/ppl_ratio:np.float64(1.0006180678028613) - rollout_corr/chi2_token:np.float64(0.000745030352845788) - rollout_corr/chi2_seq:np.float64(0.12269318685866892) - actor/pg_loss:np.float64(0.0005168186035007238) - actor/kl_loss:np.float64(0.0) - actor/pg_clipfrac:np.float64(0.00015859473069212981) - actor/ppo_kl:np.float64(0.00022523770326543513) - actor/pg_clipfrac_lower:np.float64(0.0) - self_distillation/token_reweight_lambda:np.float64(0.5) - self_distillation/token_reweight_eps_w:np.float64(0.20000000000000004) - self_distillation/token_reweight_active_token_fraction:np.float64(0.75) - self_distillation/token_reweight_context_sample_fraction:np.float64(0.75) - self_distillation/token_reweight_correct_sample_fraction:np.float64(0.7265625) - self_distillation/token_reweight_w_mean:np.float64(8113.411336590536) - self_distillation/token_reweight_w_std:np.float64(123141.12176522095) - self_distillation/token_reweight_w_clipped_mean:np.float64(0.9997582165524364) - self_distillation/token_reweight_w_clip_frac:np.float64(0.015467898192582652) - self_distillation/empty_target_batch:np.float64(0.25) - actor/grad_norm:np.float64(0.1160267498344183) - perf/mfu/actor:np.float64(0.05294808405288777) - perf/max_memory_allocated_gb:np.float64(130.98332166671753) - perf/max_memory_reserved_gb:np.float64(140.86328125) - perf/cpu_memory_used_gb:np.float64(195.59362411499023) - actor/lr:np.float64(1e-06) - training/global_step:59 - training/epoch:2 - critic/score/mean:0.7265625 - critic/score/max:1.0 - critic/score/min:0.0 - critic/rewards/mean:0.7265625 - critic/rewards/max:1.0 - critic/rewards/min:0.0 - critic/advantages/mean:0.0011978000402450562 - critic/advantages/max:0.25 - critic/advantages/min:-0.875 - critic/returns/mean:0.0011978000402450562 - critic/returns/max:0.25 - critic/returns/min:-0.875 - response_length/mean:347.31640625 - response_length/max:824.0 - response_leng
|
||
test_gen_batch meta info: {'eos_token_id': 151645, 'pad_token_id': 151643, 'recompute_log_prob': False, 'do_sample': True, 'validate': True, 'global_steps': 60}
|
||
validation generation end
|
||
[prompt] system
|
||
|
||
Given a question and four options, please select the right answer. Respond in the following format:
|
||
<reasoning>
|
||
...
|
||
</reasoning>
|
||
<answer>
|
||
...
|
||
</answer>
|
||
|
||
For the answer, only output the letter corresponding to the correct option (A, B, C, or D), and nothing else. Do not restate the answer text. For example, if the answer is "A", just output:
|
||
<answer>
|
||
A
|
||
</answer>
|
||
|
||
user
|
||
Given the following crystal structure parameters for the material RbLa9(IrO6)4 (Material ID: mp-560657), calculate the volume of the unit cell (in ų).
|
||
|
||
Lattice Parameters:
|
||
|
||
a = 7.82 Å
|
||
b = 7.82 Å
|
||
c = 17.88 Å
|
||
α = 90.00º
|
||
β = 90.00º
|
||
γ = 90.00º
|
||
|
||
|
||
|
||
A: 1025.67 ų
|
||
|
||
B: 1094.31 ų
|
||
|
||
C: 1200.45 ų
|
||
|
||
D: 1150.78 ų
|
||
|
||
Please reason step by step.
|
||
assistant
|
||
<think>
|
||
|
||
</think>
|
||
|
||
|
||
[response] <reasoning>
|
||
The given crystal structure has orthorhombic symmetry, as all angles are 90 degrees. The volume of a unit cell for an orthorhombic system is calculated using the formula:
|
||
|
||
$$ V = a \times b \times c $$
|
||
|
||
Substituting the given lattice parameters:
|
||
|
||
$$ V = 7.82 \times 7.82 \times 17.88 $$
|
||
|
||
First, calculate $ 7.82 \times 7.82 $:
|
||
|
||
$$ 7.82 \times 7.82 = 61.1524 $$
|
||
|
||
Now multiply by 17.88:
|
||
|
||
$$ 61.1524 \times 17.88 = 1094.31 $$
|
||
|
||
Thus, the volume of the unit cell is 1094.31 ų.
|
||
</reasoning>
|
||
<answer>
|
||
B
|
||
</answer>
|
||
[ground_truth] B
|
||
[score] 1.0
|
||
[acc] 1.0
|
||
[pred] B
|
||
[incorrect_format] 1
|
||
[feedback]
|
||
local_global_step_folder: /mnt/mole/SDPO/L2T/checkpoints/datasets/sciknoweval/material/qwen3gen-material-RLSD_TR-Qwen-Qwen3-8B-mbs8-decay0-tr0.1-train32-rollout8-lr1e-6-vllm0.8/global_step_60
|
||
step:60 - global_seqlen/min:20976 - global_seqlen/max:28367 - global_seqlen/minmax_diff:7391 - global_seqlen/balanced_min:24354 - global_seqlen/balanced_max:24362 - global_seqlen/mean:24358.5 - actor/entropy:0.1625373214483261 - perf/mfu/actor_infer:0 - training/rollout_probs_diff_valid:1 - training/rollout_probs_diff_max:0.22152966260910034 - training/rollout_probs_diff_mean:0.002746312413364649 - training/rollout_probs_diff_std:0.009204926900565624 - training/rollout_actor_probs_pearson_corr:0.9989629983901978 - self_distillation/success_group_fraction:0.65625 - self_distillation/success_sample_fraction:0.65625 - self_distillation/correct_sample_fraction:0.6171875 - self_distillation/feedback_available_fraction:0.0 - self_distillation/feedback_used_fraction:0.0 - self_distillation/reprompt_sample_fraction:0.65625 - rollout_corr/rollout_is_mean:0.999926745891571 - rollout_corr/rollout_is_ratio_fraction_high:1.0308002856618259e-05 - rollout_corr/rollout_is_ratio_fraction_low:0.0 - rollout_corr/rollout_is_max:2.767468214035034 - rollout_corr/rollout_is_min:0.6319364309310913 - rollout_corr/rollout_is_std:0.02688652276992798 - rollout_corr/rollout_is_eff_sample_size:0.9992776370535978 - rollout_corr/rollout_is_seq_mean:0.9998831152915955 - rollout_corr/rollout_is_seq_std:0.0015677715418860316 - rollout_corr/rollout_is_seq_max:1.0052909851074219 - rollout_corr/rollout_is_seq_min:0.9949941039085388 - rollout_corr/rollout_is_seq_max_deviation:0.005290985107421875 - rollout_corr/rollout_is_seq_fraction_high:0.0 - rollout_corr/rollout_is_seq_fraction_low:0.0 - rollout_corr/training_ppl:np.float64(1.1889118873514235) - rollout_corr/training_log_ppl:np.float64(0.16804096282430692) - rollout_corr/kl:np.float64(0.0003125149191056664) - rollout_corr/k3_kl:np.float64(0.0004305376780280312) - rollout_corr/rollout_ppl:np.float64(1.1885348898358643) - rollout_corr/rollout_log_ppl:np.float64(0.16772844693332445) - rollout_corr/log_ppl_diff:np.float64(0.0003125158909824677) - rollout_corr/log_ppl_abs_diff:np.float64(0.0013071187931927852) - rollout_corr/log_ppl_diff_max:np.float64(0.008523553609848022) - rollout_corr/log_ppl_diff_min:np.float64(-0.006315112113952637) - rollout_corr/ppl_ratio:np.float64(1.0003141060005873) - rollout_corr/chi2_token:np.float64(0.0011115602683275938) - rollout_corr/chi2_seq:np.float64(0.3905038656666875) - actor/pg_loss:np.float64(0.0006367913447320461) - actor/kl_loss:np.float64(0.0) - actor/pg_clipfrac:np.float64(6.850907811895013e-05) - actor/ppo_kl:np.float64(-0.00019378854725826677) - actor/pg_clipfrac_lower:np.float64(0.0) - self_distillation/token_reweight_lambda:np.float64(0.5) - self_distillation/token_reweight_eps_w:np.float64(0.20000000000000004) - self_distillation/token_reweight_active_token_fraction:np.float64(0.65625) - self_distillation/token_reweight_context_sample_fraction:np.float64(0.65625) - self_distillation/token_reweight_correct_sample_fraction:np.float64(0.6171875) - self_distillation/token_reweight_w_mean:np.float64(10045.848395209527) - self_distillation/token_reweight_w_std:np.float64(136172.85056805442) - self_distillation/token_reweight_w_clipped_mean:np.float64(0.9999325245153159) - self_distillation/token_reweight_w_clip_frac:np.float64(0.016230103617999703) - self_distillation/empty_target_batch:np.float64(0.34375) - actor/grad_norm:np.float64(0.11269044876098633) - perf/mfu/actor:np.float64(0.055351362219646164) - perf/max_memory_allocated_gb:np.float64(130.98332166671753) - perf/max_memory_reserved_gb:np.float64(140.86328125) - perf/cpu_memory_used_gb:np.float64(195.59489059448242) - actor/lr:np.float64(1e-06) - val-aux/sciknoweval/reward/mean@16:np.float64(0.7513297872340425) - val-aux/sciknoweval/reward/std@16:np.float64(0.058702087346603746) - val-aux/sciknoweval/reward/best@2/mean:np.float64(0.7717765957446808) - val-aux/sciknoweval/reward/best@2/std:np.float64(0.03301362979339628) - val-aux/sciknoweval/reward/worst@2/mean:np.float64(0.7308617021276597) - val-aux/sciknoweval/reward/worst@2/std:np.float64(0.0631627294474146) - val-aux/sciknoweval/reward/maj@2/mean:np.flo
|
||
step:61 - global_seqlen/min:20942 - global_seqlen/max:26823 - global_seqlen/minmax_diff:5881 - global_seqlen/balanced_min:25153 - global_seqlen/balanced_max:25161 - global_seqlen/mean:25157.875 - actor/entropy:0.16813786327838898 - perf/mfu/actor_infer:0 - training/rollout_probs_diff_valid:1 - training/rollout_probs_diff_max:0.24496299028396606 - training/rollout_probs_diff_mean:0.002785650547593832 - training/rollout_probs_diff_std:0.009039455093443394 - training/rollout_actor_probs_pearson_corr:0.9989648461341858 - self_distillation/success_group_fraction:0.84375 - self_distillation/success_sample_fraction:0.84375 - self_distillation/correct_sample_fraction:0.80859375 - self_distillation/feedback_available_fraction:0.0 - self_distillation/feedback_used_fraction:0.0 - self_distillation/reprompt_sample_fraction:0.84375 - rollout_corr/rollout_is_mean:1.0000181198120117 - rollout_corr/rollout_is_ratio_fraction_high:1.046649140334921e-05 - rollout_corr/rollout_is_ratio_fraction_low:2.093298280669842e-05 - rollout_corr/rollout_is_max:2.5879404544830322 - rollout_corr/rollout_is_min:0.23645037412643433 - rollout_corr/rollout_is_std:0.026822155341506004 - rollout_corr/rollout_is_eff_sample_size:0.9992810891422065 - rollout_corr/rollout_is_seq_mean:0.9999784231185913 - rollout_corr/rollout_is_seq_std:0.0016213229391723871 - rollout_corr/rollout_is_seq_max:1.007311224937439 - rollout_corr/rollout_is_seq_min:0.9955859780311584 - rollout_corr/rollout_is_seq_max_deviation:0.007311224937438965 - rollout_corr/rollout_is_seq_fraction_high:0.0 - rollout_corr/rollout_is_seq_fraction_low:0.0 - rollout_corr/training_ppl:np.float64(1.1883001108653843) - rollout_corr/training_log_ppl:np.float64(0.17076641073072096) - rollout_corr/kl:np.float64(0.00046844173982929505) - rollout_corr/k3_kl:np.float64(0.00041541116644339127) - rollout_corr/rollout_ppl:np.float64(1.1877418737858534) - rollout_corr/rollout_log_ppl:np.float64(0.17029796764836647) - rollout_corr/log_ppl_diff:np.float64(0.00046844308235449716) - rollout_corr/log_ppl_abs_diff:np.float64(0.001244591076101642) - rollout_corr/log_ppl_diff_max:np.float64(0.005107656121253967) - rollout_corr/log_ppl_diff_min:np.float64(-0.006342723965644836) - rollout_corr/ppl_ratio:np.float64(1.000469823135063) - rollout_corr/chi2_token:np.float64(0.0007278069388121367) - rollout_corr/chi2_seq:np.float64(0.3579556457698345) - actor/pg_loss:np.float64(0.0007300071883946657) - actor/kl_loss:np.float64(0.0) - actor/pg_clipfrac:np.float64(6.12041867498192e-05) - actor/ppo_kl:np.float64(5.782134989851251e-05) - actor/pg_clipfrac_lower:np.float64(0.0) - self_distillation/token_reweight_lambda:np.float64(0.5) - self_distillation/token_reweight_eps_w:np.float64(0.20000000000000004) - self_distillation/token_reweight_active_token_fraction:np.float64(0.84375) - self_distillation/token_reweight_context_sample_fraction:np.float64(0.84375) - self_distillation/token_reweight_correct_sample_fraction:np.float64(0.80859375) - self_distillation/token_reweight_w_mean:np.float64(3637.0788466231897) - self_distillation/token_reweight_w_std:np.float64(82982.7309890319) - self_distillation/token_reweight_w_clipped_mean:np.float64(0.9997178621124476) - self_distillation/token_reweight_w_clip_frac:np.float64(0.020233669521985576) - self_distillation/empty_target_batch:np.float64(0.15625) - actor/grad_norm:np.float64(0.11315641738474369) - perf/mfu/actor:np.float64(0.056785998846264056) - perf/max_memory_allocated_gb:np.float64(130.98332166671753) - perf/max_memory_reserved_gb:np.float64(140.86328125) - perf/cpu_memory_used_gb:np.float64(195.96517181396484) - actor/lr:np.float64(1e-06) - training/global_step:61 - training/epoch:2 - critic/score/mean:0.80859375 - critic/score/max:1.0 - critic/score/min:0.0 - critic/rewards/mean:0.80859375 - critic/rewards/max:1.0 - critic/rewards/min:0.0 - critic/advantages/mean:-0.0031621886882930994 - critic/advantages/max:0.625 - critic/advantages/min:-0.875 - critic/returns/mean:-0.0031621886882930994 - critic/returns/max:0.625 - critic/returns/min:-0.875 - response_length/mean:373.21484375 -
|
||
step:62 - global_seqlen/min:19030 - global_seqlen/max:27927 - global_seqlen/minmax_diff:8897 - global_seqlen/balanced_min:24956 - global_seqlen/balanced_max:24968 - global_seqlen/mean:24964.875 - actor/entropy:0.15622001886367798 - perf/mfu/actor_infer:0 - training/rollout_probs_diff_valid:1 - training/rollout_probs_diff_max:0.4004977345466614 - training/rollout_probs_diff_mean:0.0024245709646493196 - training/rollout_probs_diff_std:0.008639764040708542 - training/rollout_actor_probs_pearson_corr:0.9990466833114624 - self_distillation/success_group_fraction:0.90625 - self_distillation/success_sample_fraction:0.90234375 - self_distillation/correct_sample_fraction:0.8515625 - self_distillation/feedback_available_fraction:0.0 - self_distillation/feedback_used_fraction:0.0 - self_distillation/reprompt_sample_fraction:0.90234375 - rollout_corr/rollout_is_mean:0.9999725222587585 - rollout_corr/rollout_is_ratio_fraction_high:0.0 - rollout_corr/rollout_is_ratio_fraction_low:9.962640433514025e-06 - rollout_corr/rollout_is_max:1.712405800819397 - rollout_corr/rollout_is_min:0.44539976119995117 - rollout_corr/rollout_is_std:0.025484351441264153 - rollout_corr/rollout_is_eff_sample_size:0.9993509693049463 - rollout_corr/rollout_is_seq_mean:0.9999911785125732 - rollout_corr/rollout_is_seq_std:0.0014212263049557805 - rollout_corr/rollout_is_seq_max:1.004982829093933 - rollout_corr/rollout_is_seq_min:0.9956592321395874 - rollout_corr/rollout_is_seq_max_deviation:0.0049828290939331055 - rollout_corr/rollout_is_seq_fraction_high:0.0 - rollout_corr/rollout_is_seq_fraction_low:0.0 - rollout_corr/training_ppl:np.float64(1.1682638903148472) - rollout_corr/training_log_ppl:np.float64(0.15048128882335732) - rollout_corr/kl:np.float64(0.000456686800131223) - rollout_corr/k3_kl:np.float64(0.0003724734151830944) - rollout_corr/rollout_ppl:np.float64(1.1677364106290042) - rollout_corr/rollout_log_ppl:np.float64(0.15002460227697156) - rollout_corr/log_ppl_diff:np.float64(0.0004566865463857539) - rollout_corr/log_ppl_abs_diff:np.float64(0.0011863850013469346) - rollout_corr/log_ppl_diff_max:np.float64(0.007962048053741455) - rollout_corr/log_ppl_diff_min:np.float64(-0.004445493221282959) - rollout_corr/ppl_ratio:np.float64(1.0004580072127283) - rollout_corr/chi2_token:np.float64(0.0005830880254507065) - rollout_corr/chi2_seq:np.float64(0.4403882585465908) - actor/pg_loss:np.float64(0.0009687773417681456) - actor/kl_loss:np.float64(0.0) - actor/pg_clipfrac:np.float64(9.318608908870374e-05) - actor/ppo_kl:np.float64(0.00010262198350829976) - actor/pg_clipfrac_lower:np.float64(0.0) - self_distillation/token_reweight_lambda:np.float64(0.5) - self_distillation/token_reweight_eps_w:np.float64(0.20000000000000004) - self_distillation/token_reweight_active_token_fraction:np.float64(0.90234375) - self_distillation/token_reweight_context_sample_fraction:np.float64(0.90234375) - self_distillation/token_reweight_correct_sample_fraction:np.float64(0.8515625) - self_distillation/token_reweight_w_mean:np.float64(44.64393255230971) - self_distillation/token_reweight_w_std:np.float64(657.1197780608782) - self_distillation/token_reweight_w_clipped_mean:np.float64(1.0001448588445783) - self_distillation/token_reweight_w_clip_frac:np.float64(0.021430096472613513) - self_distillation/empty_target_batch:np.float64(0.09765625) - actor/grad_norm:np.float64(0.09477861924096942) - perf/mfu/actor:np.float64(0.05646728138609545) - perf/max_memory_allocated_gb:np.float64(130.98332166671753) - perf/max_memory_reserved_gb:np.float64(140.86328125) - perf/cpu_memory_used_gb:np.float64(195.61653518676758) - actor/lr:np.float64(1e-06) - training/global_step:62 - training/epoch:2 - critic/score/mean:0.8515625 - critic/score/max:1.0 - critic/score/min:0.0 - critic/rewards/mean:0.8515625 - critic/rewards/max:1.0 - critic/rewards/min:0.0 - critic/advantages/mean:-0.0025554171297699213 - critic/advantages/max:0.875 - critic/advantages/min:-0.875 - critic/returns/mean:-0.0025554171297699213 - critic/returns/max:0.875 - critic/returns/min:-0.875 - response_length/mean:392.08984375 - respons
|
||
step:63 - global_seqlen/min:21676 - global_seqlen/max:29352 - global_seqlen/minmax_diff:7676 - global_seqlen/balanced_min:26177 - global_seqlen/balanced_max:26180 - global_seqlen/mean:26178.375 - actor/entropy:0.18619155883789062 - perf/mfu/actor_infer:0 - training/rollout_probs_diff_valid:1 - training/rollout_probs_diff_max:0.3797507882118225 - training/rollout_probs_diff_mean:0.002799584763124585 - training/rollout_probs_diff_std:0.008926978334784508 - training/rollout_actor_probs_pearson_corr:0.9991098642349243 - self_distillation/success_group_fraction:0.78125 - self_distillation/success_sample_fraction:0.77734375 - self_distillation/correct_sample_fraction:0.6953125 - self_distillation/feedback_available_fraction:0.0 - self_distillation/feedback_used_fraction:0.0 - self_distillation/reprompt_sample_fraction:0.77734375 - rollout_corr/rollout_is_mean:0.9999904632568359 - rollout_corr/rollout_is_ratio_fraction_high:0.0 - rollout_corr/rollout_is_ratio_fraction_low:9.512666110822465e-06 - rollout_corr/rollout_is_max:1.6014662981033325 - rollout_corr/rollout_is_min:0.3244990110397339 - rollout_corr/rollout_is_std:0.026437172666192055 - rollout_corr/rollout_is_eff_sample_size:0.9993015640890296 - rollout_corr/rollout_is_seq_mean:0.9999660849571228 - rollout_corr/rollout_is_seq_std:0.0015565672656521201 - rollout_corr/rollout_is_seq_max:1.0054783821105957 - rollout_corr/rollout_is_seq_min:0.9923509359359741 - rollout_corr/rollout_is_seq_max_deviation:0.007649064064025879 - rollout_corr/rollout_is_seq_fraction_high:0.0 - rollout_corr/rollout_is_seq_fraction_low:0.0 - rollout_corr/training_ppl:np.float64(1.2034375509247184) - rollout_corr/training_log_ppl:np.float64(0.1827049713829183) - rollout_corr/kl:np.float64(0.0005114174473841437) - rollout_corr/k3_kl:np.float64(0.00042598485559608434) - rollout_corr/rollout_ppl:np.float64(1.2028177841566503) - rollout_corr/rollout_log_ppl:np.float64(0.18219355211476795) - rollout_corr/log_ppl_diff:np.float64(0.0005114192681503482) - rollout_corr/log_ppl_abs_diff:np.float64(0.001406845221936237) - rollout_corr/log_ppl_diff_max:np.float64(0.008712932467460632) - rollout_corr/log_ppl_diff_min:np.float64(-0.005025111138820648) - rollout_corr/ppl_ratio:np.float64(1.0005130991339684) - rollout_corr/chi2_token:np.float64(0.0006853425875306129) - rollout_corr/chi2_seq:np.float64(0.5162155868019909) - actor/pg_loss:np.float64(0.0013690560590475798) - actor/kl_loss:np.float64(0.0) - actor/pg_clipfrac:np.float64(0.0002309702149432269) - actor/ppo_kl:np.float64(0.00011990283803520185) - actor/pg_clipfrac_lower:np.float64(0.0) - self_distillation/token_reweight_lambda:np.float64(0.5) - self_distillation/token_reweight_eps_w:np.float64(0.20000000000000004) - self_distillation/token_reweight_active_token_fraction:np.float64(0.77734375) - self_distillation/token_reweight_context_sample_fraction:np.float64(0.77734375) - self_distillation/token_reweight_correct_sample_fraction:np.float64(0.6953125) - self_distillation/token_reweight_w_mean:np.float64(3051.147950245766) - self_distillation/token_reweight_w_std:np.float64(45859.22248870792) - self_distillation/token_reweight_w_clipped_mean:np.float64(0.9996626949869096) - self_distillation/token_reweight_w_clip_frac:np.float64(0.040024787274887785) - self_distillation/empty_target_batch:np.float64(0.22265625) - actor/grad_norm:np.float64(0.14891982078552246) - perf/mfu/actor:np.float64(0.059230344477378155) - perf/max_memory_allocated_gb:np.float64(130.98332166671753) - perf/max_memory_reserved_gb:np.float64(140.86328125) - perf/cpu_memory_used_gb:np.float64(195.58012104034424) - actor/lr:np.float64(1e-06) - training/global_step:63 - training/epoch:2 - critic/score/mean:0.6953125 - critic/score/max:1.0 - critic/score/min:0.0 - critic/rewards/mean:0.6953125 - critic/rewards/max:1.0 - critic/rewards/min:0.0 - critic/advantages/mean:-0.001120116445235908 - critic/advantages/max:0.875 - critic/advantages/min:-0.875 - critic/returns/mean:-0.001120116445235908 - critic/returns/max:0.875 - critic/returns/min:-0.875 - response_length/mean:410.63671875 - response_
|
||
step:64 - global_seqlen/min:21649 - global_seqlen/max:29898 - global_seqlen/minmax_diff:8249 - global_seqlen/balanced_min:25438 - global_seqlen/balanced_max:25443 - global_seqlen/mean:25441.875 - actor/entropy:0.17299145460128784 - perf/mfu/actor_infer:0 - training/rollout_probs_diff_valid:1 - training/rollout_probs_diff_max:0.19256961345672607 - training/rollout_probs_diff_mean:0.0027170064859092236 - training/rollout_probs_diff_std:0.008848386816680431 - training/rollout_actor_probs_pearson_corr:0.9989994168281555 - self_distillation/success_group_fraction:0.84375 - self_distillation/success_sample_fraction:0.83984375 - self_distillation/correct_sample_fraction:0.80078125 - self_distillation/feedback_available_fraction:0.0 - self_distillation/feedback_used_fraction:0.0 - self_distillation/reprompt_sample_fraction:0.83984375 - rollout_corr/rollout_is_mean:0.9999280571937561 - rollout_corr/rollout_is_ratio_fraction_high:0.0 - rollout_corr/rollout_is_ratio_fraction_low:0.0 - rollout_corr/rollout_is_max:1.6264592409133911 - rollout_corr/rollout_is_min:0.6363797187805176 - rollout_corr/rollout_is_std:0.02642815187573433 - rollout_corr/rollout_is_eff_sample_size:0.9993019212176399 - rollout_corr/rollout_is_seq_mean:0.999889612197876 - rollout_corr/rollout_is_seq_std:0.001410823781043291 - rollout_corr/rollout_is_seq_max:1.0057334899902344 - rollout_corr/rollout_is_seq_min:0.9941492676734924 - rollout_corr/rollout_is_seq_max_deviation:0.005850732326507568 - rollout_corr/rollout_is_seq_fraction_high:0.0 - rollout_corr/rollout_is_seq_fraction_low:0.0 - rollout_corr/training_ppl:np.float64(1.1918034963309765) - rollout_corr/training_log_ppl:np.float64(0.1726833581342362) - rollout_corr/kl:np.float64(0.000396452359726851) - rollout_corr/k3_kl:np.float64(0.0003939119860660867) - rollout_corr/rollout_ppl:np.float64(1.1913265604525805) - rollout_corr/rollout_log_ppl:np.float64(0.17228690521005774) - rollout_corr/log_ppl_diff:np.float64(0.0003964529241784476) - rollout_corr/log_ppl_abs_diff:np.float64(0.001206114662636537) - rollout_corr/log_ppl_diff_max:np.float64(0.006472155451774597) - rollout_corr/log_ppl_diff_min:np.float64(-0.003871813416481018) - rollout_corr/ppl_ratio:np.float64(1.0003976812586188) - rollout_corr/chi2_token:np.float64(0.0007929231505841017) - rollout_corr/chi2_seq:np.float64(0.31985570839606225) - actor/pg_loss:np.float64(0.0005421672249212861) - actor/kl_loss:np.float64(0.0) - actor/pg_clipfrac:np.float64(8.674097307448392e-05) - actor/ppo_kl:np.float64(-8.614134704387766e-05) - actor/pg_clipfrac_lower:np.float64(0.0) - self_distillation/token_reweight_lambda:np.float64(0.5) - self_distillation/token_reweight_eps_w:np.float64(0.20000000000000004) - self_distillation/token_reweight_active_token_fraction:np.float64(0.83984375) - self_distillation/token_reweight_context_sample_fraction:np.float64(0.83984375) - self_distillation/token_reweight_correct_sample_fraction:np.float64(0.80078125) - self_distillation/token_reweight_w_mean:np.float64(17.55605116346851) - self_distillation/token_reweight_w_std:np.float64(375.6963393909973) - self_distillation/token_reweight_w_clipped_mean:np.float64(1.0002221886534244) - self_distillation/token_reweight_w_clip_frac:np.float64(0.016844158904859796) - self_distillation/empty_target_batch:np.float64(0.16015625) - actor/grad_norm:np.float64(0.07351323496550322) - perf/mfu/actor:np.float64(0.05771520629925124) - perf/max_memory_allocated_gb:np.float64(130.98332166671753) - perf/max_memory_reserved_gb:np.float64(140.86328125) - perf/cpu_memory_used_gb:np.float64(195.62054443359375) - actor/lr:np.float64(1e-06) - training/global_step:64 - training/epoch:2 - critic/score/mean:0.80078125 - critic/score/max:1.0 - critic/score/min:0.0 - critic/rewards/mean:0.80078125 - critic/rewards/max:1.0 - critic/rewards/min:0.0 - critic/advantages/mean:-0.0016226072330027819 - critic/advantages/max:0.875 - critic/advantages/min:-0.875 - critic/returns/mean:-0.0016226072330027819 - critic/returns/max:0.875 - critic/returns/min:-0.875 - response_length/mean:403.83984375 - response_length/max:802.
|
||
step:65 - global_seqlen/min:22005 - global_seqlen/max:30883 - global_seqlen/minmax_diff:8878 - global_seqlen/balanced_min:26219 - global_seqlen/balanced_max:26223 - global_seqlen/mean:26221.625 - actor/entropy:0.1663258969783783 - perf/mfu/actor_infer:0 - training/rollout_probs_diff_valid:1 - training/rollout_probs_diff_max:0.3175744414329529 - training/rollout_probs_diff_mean:0.0026298724114894867 - training/rollout_probs_diff_std:0.008909679017961025 - training/rollout_actor_probs_pearson_corr:0.9990459680557251 - self_distillation/success_group_fraction:0.6875 - self_distillation/success_sample_fraction:0.6875 - self_distillation/correct_sample_fraction:0.6171875 - self_distillation/feedback_available_fraction:0.0 - self_distillation/feedback_used_fraction:0.0 - self_distillation/reprompt_sample_fraction:0.6875 - rollout_corr/rollout_is_mean:0.9999517202377319 - rollout_corr/rollout_is_ratio_fraction_high:0.0 - rollout_corr/rollout_is_ratio_fraction_low:0.0 - rollout_corr/rollout_is_max:1.9659714698791504 - rollout_corr/rollout_is_min:0.5378821492195129 - rollout_corr/rollout_is_std:0.0260442104190588 - rollout_corr/rollout_is_eff_sample_size:0.9993220398749199 - rollout_corr/rollout_is_seq_mean:0.9998959302902222 - rollout_corr/rollout_is_seq_std:0.0014111259952187538 - rollout_corr/rollout_is_seq_max:1.0054465532302856 - rollout_corr/rollout_is_seq_min:0.9965800642967224 - rollout_corr/rollout_is_seq_max_deviation:0.0054465532302856445 - rollout_corr/rollout_is_seq_fraction_high:0.0 - rollout_corr/rollout_is_seq_fraction_low:0.0 - rollout_corr/training_ppl:np.float64(1.1869128532707691) - rollout_corr/training_log_ppl:np.float64(0.16875576815073146) - rollout_corr/kl:np.float64(0.0004878026736196972) - rollout_corr/k3_kl:np.float64(0.00036951068329926784) - rollout_corr/rollout_ppl:np.float64(1.1863231775350869) - rollout_corr/rollout_log_ppl:np.float64(0.16826796504756203) - rollout_corr/log_ppl_diff:np.float64(0.0004878031031694263) - rollout_corr/log_ppl_abs_diff:np.float64(0.001221742873894982) - rollout_corr/log_ppl_diff_max:np.float64(0.008714795112609863) - rollout_corr/log_ppl_diff_min:np.float64(-0.003134891390800476) - rollout_corr/ppl_ratio:np.float64(1.000489059370011) - rollout_corr/chi2_token:np.float64(0.0004872032441198826) - rollout_corr/chi2_seq:np.float64(0.24569943128153682) - actor/pg_loss:np.float64(0.0006585207302123308) - actor/kl_loss:np.float64(0.0) - actor/pg_clipfrac:np.float64(0.00013484467808666523) - actor/ppo_kl:np.float64(2.8466167123042396e-05) - actor/pg_clipfrac_lower:np.float64(0.0) - self_distillation/token_reweight_lambda:np.float64(0.5) - self_distillation/token_reweight_eps_w:np.float64(0.20000000000000004) - self_distillation/token_reweight_active_token_fraction:np.float64(0.6875) - self_distillation/token_reweight_context_sample_fraction:np.float64(0.6875) - self_distillation/token_reweight_correct_sample_fraction:np.float64(0.6171875) - self_distillation/token_reweight_w_mean:np.float64(62.919017291860655) - self_distillation/token_reweight_w_std:np.float64(1208.9836768980604) - self_distillation/token_reweight_w_clipped_mean:np.float64(1.000118033727631) - self_distillation/token_reweight_w_clip_frac:np.float64(0.02347485371865332) - self_distillation/empty_target_batch:np.float64(0.3125) - actor/grad_norm:np.float64(0.20872666779905558) - perf/mfu/actor:np.float64(0.058806865577430006) - perf/max_memory_allocated_gb:np.float64(130.98332166671753) - perf/max_memory_reserved_gb:np.float64(140.86328125) - perf/cpu_memory_used_gb:np.float64(195.7265510559082) - actor/lr:np.float64(1e-06) - training/global_step:65 - training/epoch:2 - critic/score/mean:0.6171875 - critic/score/max:1.0 - critic/score/min:0.0 - critic/rewards/mean:0.6171875 - critic/rewards/max:1.0 - critic/rewards/min:0.0 - critic/advantages/mean:-0.001335279201157391 - critic/advantages/max:0.75 - critic/advantages/min:-0.875 - critic/returns/mean:-0.001335279201157391 - critic/returns/max:0.75 - critic/returns/min:-0.875 - response_length/mean:420.89453125 - response_length/max:836.0 - response_length/min:16
|
||
step:66 - global_seqlen/min:16342 - global_seqlen/max:28272 - global_seqlen/minmax_diff:11930 - global_seqlen/balanced_min:24902 - global_seqlen/balanced_max:24906 - global_seqlen/mean:24905.0 - actor/entropy:0.1453680694103241 - perf/mfu/actor_infer:0 - training/rollout_probs_diff_valid:1 - training/rollout_probs_diff_max:0.4400331377983093 - training/rollout_probs_diff_mean:0.002347266301512718 - training/rollout_probs_diff_std:0.008429394103586674 - training/rollout_actor_probs_pearson_corr:0.9990342259407043 - self_distillation/success_group_fraction:0.78125 - self_distillation/success_sample_fraction:0.77734375 - self_distillation/correct_sample_fraction:0.69921875 - self_distillation/feedback_available_fraction:0.0 - self_distillation/feedback_used_fraction:0.0 - self_distillation/reprompt_sample_fraction:0.77734375 - rollout_corr/rollout_is_mean:0.9998712539672852 - rollout_corr/rollout_is_ratio_fraction_high:0.0 - rollout_corr/rollout_is_ratio_fraction_low:1.00136185210431e-05 - rollout_corr/rollout_is_max:1.6471433639526367 - rollout_corr/rollout_is_min:0.46178218722343445 - rollout_corr/rollout_is_std:0.024240102618932724 - rollout_corr/rollout_is_eff_sample_size:0.9994124052545171 - rollout_corr/rollout_is_seq_mean:0.9998443722724915 - rollout_corr/rollout_is_seq_std:0.001488116686232388 - rollout_corr/rollout_is_seq_max:1.0049697160720825 - rollout_corr/rollout_is_seq_min:0.9944682717323303 - rollout_corr/rollout_is_seq_max_deviation:0.005531728267669678 - rollout_corr/rollout_is_seq_fraction_high:0.0 - rollout_corr/rollout_is_seq_fraction_low:0.0 - rollout_corr/training_ppl:np.float64(1.1638081297278404) - rollout_corr/training_log_ppl:np.float64(0.14921010341640795) - rollout_corr/kl:np.float64(0.0003408923481164061) - rollout_corr/k3_kl:np.float64(0.00038058763544057683) - rollout_corr/rollout_ppl:np.float64(1.1633959328755736) - rollout_corr/rollout_log_ppl:np.float64(0.1488692093444115) - rollout_corr/log_ppl_diff:np.float64(0.00034089407199644484) - rollout_corr/log_ppl_abs_diff:np.float64(0.0012122128500777762) - rollout_corr/log_ppl_diff_max:np.float64(0.006232038140296936) - rollout_corr/log_ppl_diff_min:np.float64(-0.009414836764335632) - rollout_corr/ppl_ratio:np.float64(1.0003424615133554) - rollout_corr/chi2_token:np.float64(0.0008226935751736164) - rollout_corr/chi2_seq:np.float64(0.423892357153818) - actor/pg_loss:np.float64(0.0008258491288870573) - actor/kl_loss:np.float64(0.0) - actor/pg_clipfrac:np.float64(0.00014899678853907972) - actor/ppo_kl:np.float64(-0.00015697772187905912) - actor/pg_clipfrac_lower:np.float64(0.0) - self_distillation/token_reweight_lambda:np.float64(0.5) - self_distillation/token_reweight_eps_w:np.float64(0.20000000000000004) - self_distillation/token_reweight_active_token_fraction:np.float64(0.77734375) - self_distillation/token_reweight_context_sample_fraction:np.float64(0.77734375) - self_distillation/token_reweight_correct_sample_fraction:np.float64(0.69921875) - self_distillation/token_reweight_w_mean:np.float64(5992.965957323788) - self_distillation/token_reweight_w_std:np.float64(127280.61096522707) - self_distillation/token_reweight_w_clipped_mean:np.float64(1.000441890442744) - self_distillation/token_reweight_w_clip_frac:np.float64(0.019730598985916004) - self_distillation/empty_target_batch:np.float64(0.22265625) - actor/grad_norm:np.float64(0.12438435479998589) - perf/mfu/actor:np.float64(0.05638331722933918) - perf/max_memory_allocated_gb:np.float64(130.98332166671753) - perf/max_memory_reserved_gb:np.float64(140.86328125) - perf/cpu_memory_used_gb:np.float64(195.728759765625) - actor/lr:np.float64(1e-06) - training/global_step:66 - training/epoch:2 - critic/score/mean:0.69921875 - critic/score/max:1.0 - critic/score/min:0.0 - critic/rewards/mean:0.69921875 - critic/rewards/max:1.0 - critic/rewards/min:0.0 - critic/advantages/mean:-0.004558699671179056 - critic/advantages/max:0.875 - critic/advantages/min:-0.625 - critic/returns/mean:-0.004558699671179056 - critic/returns/max:0.875 - critic/returns/min:-0.625 - response_length/mean:390.09375 - response_le
|
||
step:67 - global_seqlen/min:22521 - global_seqlen/max:26374 - global_seqlen/minmax_diff:3853 - global_seqlen/balanced_min:24313 - global_seqlen/balanced_max:24317 - global_seqlen/mean:24315.25 - actor/entropy:0.1448637843132019 - perf/mfu/actor_infer:0 - training/rollout_probs_diff_valid:1 - training/rollout_probs_diff_max:0.42377305030822754 - training/rollout_probs_diff_mean:0.0024852403439581394 - training/rollout_probs_diff_std:0.008906054310500622 - training/rollout_actor_probs_pearson_corr:0.99898362159729 - self_distillation/success_group_fraction:0.8125 - self_distillation/success_sample_fraction:0.8125 - self_distillation/correct_sample_fraction:0.80078125 - self_distillation/feedback_available_fraction:0.0 - self_distillation/feedback_used_fraction:0.0 - self_distillation/reprompt_sample_fraction:0.8125 - rollout_corr/rollout_is_mean:1.0000438690185547 - rollout_corr/rollout_is_ratio_fraction_high:1.043035626935307e-05 - rollout_corr/rollout_is_ratio_fraction_low:2.086071253870614e-05 - rollout_corr/rollout_is_max:2.217512369155884 - rollout_corr/rollout_is_min:0.15181678533554077 - rollout_corr/rollout_is_std:0.025744972750544548 - rollout_corr/rollout_is_eff_sample_size:0.9993376353677934 - rollout_corr/rollout_is_seq_mean:1.0000419616699219 - rollout_corr/rollout_is_seq_std:0.001741582527756691 - rollout_corr/rollout_is_seq_max:1.0054556131362915 - rollout_corr/rollout_is_seq_min:0.9952270984649658 - rollout_corr/rollout_is_seq_max_deviation:0.005455613136291504 - rollout_corr/rollout_is_seq_fraction_high:0.0 - rollout_corr/rollout_is_seq_fraction_low:0.0 - rollout_corr/training_ppl:np.float64(1.1658252156339586) - rollout_corr/training_log_ppl:np.float64(0.15123848517396254) - rollout_corr/kl:np.float64(0.0003005595714675735) - rollout_corr/k3_kl:np.float64(0.00047102890214034687) - rollout_corr/rollout_ppl:np.float64(1.1654559667222202) - rollout_corr/rollout_log_ppl:np.float64(0.15093792550760554) - rollout_corr/log_ppl_diff:np.float64(0.00030055966635700315) - rollout_corr/log_ppl_abs_diff:np.float64(0.0014758130128029734) - rollout_corr/log_ppl_diff_max:np.float64(0.006961077451705933) - rollout_corr/log_ppl_diff_min:np.float64(-0.006364375352859497) - rollout_corr/ppl_ratio:np.float64(1.0003026039339602) - rollout_corr/chi2_token:np.float64(0.001243996899574995) - rollout_corr/chi2_seq:np.float64(0.4573242797050625) - actor/pg_loss:np.float64(0.00016782758757472038) - actor/kl_loss:np.float64(0.0) - actor/pg_clipfrac:np.float64(1.8803054445015732e-05) - actor/ppo_kl:np.float64(-3.549655894374837e-05) - actor/pg_clipfrac_lower:np.float64(0.0) - self_distillation/token_reweight_lambda:np.float64(0.5) - self_distillation/token_reweight_eps_w:np.float64(0.20000000000000004) - self_distillation/token_reweight_active_token_fraction:np.float64(0.8125) - self_distillation/token_reweight_context_sample_fraction:np.float64(0.8125) - self_distillation/token_reweight_correct_sample_fraction:np.float64(0.80078125) - self_distillation/token_reweight_w_mean:np.float64(162.33343863044865) - self_distillation/token_reweight_w_std:np.float64(3453.4582545155426) - self_distillation/token_reweight_w_clipped_mean:np.float64(1.0000421507284045) - self_distillation/token_reweight_w_clip_frac:np.float64(0.004522646631812677) - self_distillation/empty_target_batch:np.float64(0.1875) - actor/grad_norm:np.float64(0.06182404700666666) - perf/mfu/actor:np.float64(0.05466950097703479) - perf/max_memory_allocated_gb:np.float64(130.98332166671753) - perf/max_memory_reserved_gb:np.float64(140.86328125) - perf/cpu_memory_used_gb:np.float64(195.71934127807617) - actor/lr:np.float64(1e-06) - training/global_step:67 - training/epoch:2 - critic/score/mean:0.80078125 - critic/score/max:1.0 - critic/score/min:0.0 - critic/rewards/mean:0.80078125 - critic/rewards/max:1.0 - critic/rewards/min:0.0 - critic/advantages/mean:-0.0018396541709080338 - critic/advantages/max:0.375 - critic/advantages/min:-0.625 - critic/returns/mean:-0.0018396541709080338 - critic/returns/max:0.375 - critic/returns/min:-0.625 - response_length/mean:374.5078125 - respo
|
||
step:68 - global_seqlen/min:16271 - global_seqlen/max:28326 - global_seqlen/minmax_diff:12055 - global_seqlen/balanced_min:23376 - global_seqlen/balanced_max:23384 - global_seqlen/mean:23380.5 - actor/entropy:0.18183209002017975 - perf/mfu/actor_infer:0 - training/rollout_probs_diff_valid:1 - training/rollout_probs_diff_max:0.2554054856300354 - training/rollout_probs_diff_mean:0.0028197066858410835 - training/rollout_probs_diff_std:0.009083811193704605 - training/rollout_actor_probs_pearson_corr:0.9990743398666382 - self_distillation/success_group_fraction:0.84375 - self_distillation/success_sample_fraction:0.83984375 - self_distillation/correct_sample_fraction:0.78125 - self_distillation/feedback_available_fraction:0.0 - self_distillation/feedback_used_fraction:0.0 - self_distillation/reprompt_sample_fraction:0.83984375 - rollout_corr/rollout_is_mean:0.9999996423721313 - rollout_corr/rollout_is_ratio_fraction_high:0.0 - rollout_corr/rollout_is_ratio_fraction_low:2.1522964743780904e-05 - rollout_corr/rollout_is_max:1.5750139951705933 - rollout_corr/rollout_is_min:0.4281967878341675 - rollout_corr/rollout_is_std:0.026538433507084846 - rollout_corr/rollout_is_eff_sample_size:0.9992963262320694 - rollout_corr/rollout_is_seq_mean:1.0001122951507568 - rollout_corr/rollout_is_seq_std:0.0016389826778322458 - rollout_corr/rollout_is_seq_max:1.0058777332305908 - rollout_corr/rollout_is_seq_min:0.9958250522613525 - rollout_corr/rollout_is_seq_max_deviation:0.00587773323059082 - rollout_corr/rollout_is_seq_fraction_high:0.0 - rollout_corr/rollout_is_seq_fraction_low:0.0 - rollout_corr/training_ppl:np.float64(1.2043992574326694) - rollout_corr/training_log_ppl:np.float64(0.17893013266075286) - rollout_corr/kl:np.float64(0.0003652134504061877) - rollout_corr/k3_kl:np.float64(0.00045952229044843307) - rollout_corr/rollout_ppl:np.float64(1.2039373209699988) - rollout_corr/rollout_log_ppl:np.float64(0.17856491890051984) - rollout_corr/log_ppl_diff:np.float64(0.00036521376023301855) - rollout_corr/log_ppl_abs_diff:np.float64(0.0012924449620186351) - rollout_corr/log_ppl_diff_max:np.float64(0.009212031960487366) - rollout_corr/log_ppl_diff_min:np.float64(-0.0048696547746658325) - rollout_corr/ppl_ratio:np.float64(1.0003667706623673) - rollout_corr/chi2_token:np.float64(0.0010798515286296606) - rollout_corr/chi2_seq:np.float64(0.2757666315883398) - actor/pg_loss:np.float64(0.0013036205200478435) - actor/kl_loss:np.float64(0.0) - actor/pg_clipfrac:np.float64(0.00010660139287210768) - actor/ppo_kl:np.float64(9.799990320047414e-05) - actor/pg_clipfrac_lower:np.float64(0.0) - self_distillation/token_reweight_lambda:np.float64(0.5) - self_distillation/token_reweight_eps_w:np.float64(0.20000000000000004) - self_distillation/token_reweight_active_token_fraction:np.float64(0.83984375) - self_distillation/token_reweight_context_sample_fraction:np.float64(0.83984375) - self_distillation/token_reweight_correct_sample_fraction:np.float64(0.78125) - self_distillation/token_reweight_w_mean:np.float64(6891.412626077421) - self_distillation/token_reweight_w_std:np.float64(127938.69730146386) - self_distillation/token_reweight_w_clipped_mean:np.float64(1.00010019633919) - self_distillation/token_reweight_w_clip_frac:np.float64(0.02797325953724794) - self_distillation/empty_target_batch:np.float64(0.16015625) - actor/grad_norm:np.float64(0.14852935820817947) - perf/mfu/actor:np.float64(0.053255143371093036) - perf/max_memory_allocated_gb:np.float64(130.98332166671753) - perf/max_memory_reserved_gb:np.float64(140.86328125) - perf/cpu_memory_used_gb:np.float64(195.65814971923828) - actor/lr:np.float64(1e-06) - training/global_step:68 - training/epoch:2 - critic/score/mean:0.78125 - critic/score/max:1.0 - critic/score/min:0.0 - critic/rewards/mean:0.78125 - critic/rewards/max:1.0 - critic/rewards/min:0.0 - critic/advantages/mean:0.00013586372369900346 - critic/advantages/max:0.875 - critic/advantages/min:-0.875 - critic/returns/mean:0.00013586372369900346 - critic/returns/max:0.875 - critic/returns/min:-0.875 - response_length/mean:362.984375 - response_length
|
||
step:69 - global_seqlen/min:18407 - global_seqlen/max:25750 - global_seqlen/minmax_diff:7343 - global_seqlen/balanced_min:23594 - global_seqlen/balanced_max:23599 - global_seqlen/mean:23596.875 - actor/entropy:0.16562606394290924 - perf/mfu/actor_infer:0 - training/rollout_probs_diff_valid:1 - training/rollout_probs_diff_max:0.24491426348686218 - training/rollout_probs_diff_mean:0.0027516221161931753 - training/rollout_probs_diff_std:0.009353713132441044 - training/rollout_actor_probs_pearson_corr:0.9988866448402405 - self_distillation/success_group_fraction:0.875 - self_distillation/success_sample_fraction:0.8671875 - self_distillation/correct_sample_fraction:0.796875 - self_distillation/feedback_available_fraction:0.0 - self_distillation/feedback_used_fraction:0.0 - self_distillation/reprompt_sample_fraction:0.8671875 - rollout_corr/rollout_is_mean:0.9999135136604309 - rollout_corr/rollout_is_ratio_fraction_high:0.0 - rollout_corr/rollout_is_ratio_fraction_low:0.0 - rollout_corr/rollout_is_max:1.6370512247085571 - rollout_corr/rollout_is_min:0.5175263285636902 - rollout_corr/rollout_is_std:0.027063293382525444 - rollout_corr/rollout_is_eff_sample_size:0.9992679951393448 - rollout_corr/rollout_is_seq_mean:0.999840259552002 - rollout_corr/rollout_is_seq_std:0.0015955602284520864 - rollout_corr/rollout_is_seq_max:1.0040080547332764 - rollout_corr/rollout_is_seq_min:0.9931057691574097 - rollout_corr/rollout_is_seq_max_deviation:0.006894230842590332 - rollout_corr/rollout_is_seq_fraction_high:0.0 - rollout_corr/rollout_is_seq_fraction_low:0.0 - rollout_corr/training_ppl:np.float64(1.1826541190966964) - rollout_corr/training_log_ppl:np.float64(0.16283819011005107) - rollout_corr/kl:np.float64(0.0006027205091285737) - rollout_corr/k3_kl:np.float64(0.00046253268788376545) - rollout_corr/rollout_ppl:np.float64(1.1819306765682995) - rollout_corr/rollout_log_ppl:np.float64(0.16223546834953595) - rollout_corr/log_ppl_diff:np.float64(0.0006027217605151236) - rollout_corr/log_ppl_abs_diff:np.float64(0.0013810389646096155) - rollout_corr/log_ppl_diff_max:np.float64(0.006462059915065765) - rollout_corr/log_ppl_diff_min:np.float64(-0.0063353851437568665) - rollout_corr/ppl_ratio:np.float64(1.0006045163609087) - rollout_corr/chi2_token:np.float64(0.00062129320576787) - rollout_corr/chi2_seq:np.float64(0.24325380730442703) - actor/pg_loss:np.float64(0.0005239989841356874) - actor/kl_loss:np.float64(0.0) - actor/pg_clipfrac:np.float64(7.407132443404407e-05) - actor/ppo_kl:np.float64(4.226569186727147e-05) - actor/pg_clipfrac_lower:np.float64(0.0) - self_distillation/token_reweight_lambda:np.float64(0.5) - self_distillation/token_reweight_eps_w:np.float64(0.20000000000000004) - self_distillation/token_reweight_active_token_fraction:np.float64(0.8671875) - self_distillation/token_reweight_context_sample_fraction:np.float64(0.8671875) - self_distillation/token_reweight_correct_sample_fraction:np.float64(0.796875) - self_distillation/token_reweight_w_mean:np.float64(40721.495245202444) - self_distillation/token_reweight_w_std:np.float64(783821.5592307698) - self_distillation/token_reweight_w_clipped_mean:np.float64(1.000015069032088) - self_distillation/token_reweight_w_clip_frac:np.float64(0.019517847162205726) - self_distillation/empty_target_batch:np.float64(0.1328125) - actor/grad_norm:np.float64(0.1003046054393053) - perf/mfu/actor:np.float64(0.05316328022214093) - perf/max_memory_allocated_gb:np.float64(130.98332166671753) - perf/max_memory_reserved_gb:np.float64(140.86328125) - perf/cpu_memory_used_gb:np.float64(195.77106618881226) - actor/lr:np.float64(1e-06) - training/global_step:69 - training/epoch:2 - critic/score/mean:0.796875 - critic/score/max:1.0 - critic/score/min:0.0 - critic/rewards/mean:0.796875 - critic/rewards/max:1.0 - critic/rewards/min:0.0 - critic/advantages/mean:-0.0013472663704305887 - critic/advantages/max:0.875 - critic/advantages/min:-0.875 - critic/returns/mean:-0.0013472663704305887 - critic/returns/max:0.875 - critic/returns/min:-0.875 - response_length/mean:374.74609375 - response_length/max:865.0 - response
|
||
test_gen_batch meta info: {'eos_token_id': 151645, 'pad_token_id': 151643, 'recompute_log_prob': False, 'do_sample': True, 'validate': True, 'global_steps': 70}
|
||
validation generation end
|
||
[prompt] system
|
||
|
||
Given a question and four options, please select the right answer. Respond in the following format:
|
||
<reasoning>
|
||
...
|
||
</reasoning>
|
||
<answer>
|
||
...
|
||
</answer>
|
||
|
||
For the answer, only output the letter corresponding to the correct option (A, B, C, or D), and nothing else. Do not restate the answer text. For example, if the answer is "A", just output:
|
||
<answer>
|
||
A
|
||
</answer>
|
||
|
||
user
|
||
Given the following crystal structure parameters for the material RbLa9(IrO6)4 (Material ID: mp-560657), calculate the volume of the unit cell (in ų).
|
||
|
||
Lattice Parameters:
|
||
|
||
a = 7.82 Å
|
||
b = 7.82 Å
|
||
c = 17.88 Å
|
||
α = 90.00º
|
||
β = 90.00º
|
||
γ = 90.00º
|
||
|
||
|
||
|
||
A: 1025.67 ų
|
||
|
||
B: 1094.31 ų
|
||
|
||
C: 1200.45 ų
|
||
|
||
D: 1150.78 ų
|
||
|
||
Please reason step by step.
|
||
assistant
|
||
<think>
|
||
|
||
</think>
|
||
|
||
|
||
[response] <reasoning>
|
||
The given crystal structure has orthorhombic symmetry, as all angles (α, β, γ) are 90.00º. For an orthorhombic system, the volume of the unit cell is calculated using the formula:
|
||
|
||
$$ V = a \times b \times c $$
|
||
|
||
Substituting the given lattice parameters:
|
||
|
||
$$ V = 7.82 \, \text{Å} \times 7.82 \, \text{Å} \times 17.88 \, \text{Å} $$
|
||
|
||
First, calculate $ 7.82 \times 7.82 $:
|
||
|
||
$$ 7.82 \times 7.82 = 61.1524 $$
|
||
|
||
Then multiply by 17.88:
|
||
|
||
$$ 61.1524 \times 17.88 = 1094.31 $$
|
||
|
||
Therefore, the volume of the unit cell is 1094.31 ų.
|
||
</reasoning>
|
||
<answer>
|
||
B
|
||
</answer>
|
||
[ground_truth] B
|
||
[score] 1.0
|
||
[acc] 1.0
|
||
[pred] B
|
||
[incorrect_format] 1
|
||
[feedback]
|
||
local_global_step_folder: /mnt/mole/SDPO/L2T/checkpoints/datasets/sciknoweval/material/qwen3gen-material-RLSD_TR-Qwen-Qwen3-8B-mbs8-decay0-tr0.1-train32-rollout8-lr1e-6-vllm0.8/global_step_70
|
||
step:70 - global_seqlen/min:21386 - global_seqlen/max:29266 - global_seqlen/minmax_diff:7880 - global_seqlen/balanced_min:24961 - global_seqlen/balanced_max:24973 - global_seqlen/mean:24969.375 - actor/entropy:0.15795306861400604 - perf/mfu/actor_infer:0 - training/rollout_probs_diff_valid:1 - training/rollout_probs_diff_max:0.17917734384536743 - training/rollout_probs_diff_mean:0.0028157206252217293 - training/rollout_probs_diff_std:0.009608049876987934 - training/rollout_actor_probs_pearson_corr:0.9988582730293274 - self_distillation/success_group_fraction:0.8125 - self_distillation/success_sample_fraction:0.8125 - self_distillation/correct_sample_fraction:0.7578125 - self_distillation/feedback_available_fraction:0.0 - self_distillation/feedback_used_fraction:0.0 - self_distillation/reprompt_sample_fraction:0.8125 - rollout_corr/rollout_is_mean:1.0001001358032227 - rollout_corr/rollout_is_ratio_fraction_high:0.0 - rollout_corr/rollout_is_ratio_fraction_low:1.9951517970184796e-05 - rollout_corr/rollout_is_max:1.5591782331466675 - rollout_corr/rollout_is_min:0.2789737582206726 - rollout_corr/rollout_is_std:0.027834173291921616 - rollout_corr/rollout_is_eff_sample_size:0.9992260966360754 - rollout_corr/rollout_is_seq_mean:1.0000709295272827 - rollout_corr/rollout_is_seq_std:0.0016937156906351447 - rollout_corr/rollout_is_seq_max:1.0067968368530273 - rollout_corr/rollout_is_seq_min:0.9945765137672424 - rollout_corr/rollout_is_seq_max_deviation:0.006796836853027344 - rollout_corr/rollout_is_seq_fraction_high:0.0 - rollout_corr/rollout_is_seq_fraction_low:0.0 - rollout_corr/training_ppl:np.float64(1.1822556359693408) - rollout_corr/training_log_ppl:np.float64(0.1616127286106348) - rollout_corr/kl:np.float64(0.0004549667964592885) - rollout_corr/k3_kl:np.float64(0.0005006182778970469) - rollout_corr/rollout_ppl:np.float64(1.1816930123604834) - rollout_corr/rollout_log_ppl:np.float64(0.16115776021615602) - rollout_corr/log_ppl_diff:np.float64(0.000454968394478783) - rollout_corr/log_ppl_abs_diff:np.float64(0.0013991295418236405) - rollout_corr/log_ppl_diff_max:np.float64(0.007021665573120117) - rollout_corr/log_ppl_diff_min:np.float64(-0.005086250603199005) - rollout_corr/ppl_ratio:np.float64(1.0004567354917526) - rollout_corr/chi2_token:np.float64(0.0010688593611121178) - rollout_corr/chi2_seq:np.float64(0.44107732083648443) - actor/pg_loss:np.float64(0.0014038580702617764) - actor/kl_loss:np.float64(0.0) - actor/pg_clipfrac:np.float64(0.00010900944607783458) - actor/ppo_kl:np.float64(0.00010926450992521097) - actor/pg_clipfrac_lower:np.float64(0.0) - self_distillation/token_reweight_lambda:np.float64(0.5) - self_distillation/token_reweight_eps_w:np.float64(0.20000000000000004) - self_distillation/token_reweight_active_token_fraction:np.float64(0.8125) - self_distillation/token_reweight_context_sample_fraction:np.float64(0.8125) - self_distillation/token_reweight_correct_sample_fraction:np.float64(0.7578125) - self_distillation/token_reweight_w_mean:np.float64(1203.3038881937973) - self_distillation/token_reweight_w_std:np.float64(20396.826950604853) - self_distillation/token_reweight_w_clipped_mean:np.float64(0.9995136295910925) - self_distillation/token_reweight_w_clip_frac:np.float64(0.03277971228817478) - self_distillation/empty_target_batch:np.float64(0.1875) - actor/grad_norm:np.float64(0.17008450627326965) - perf/mfu/actor:np.float64(0.05596456541974616) - perf/max_memory_allocated_gb:np.float64(130.98332166671753) - perf/max_memory_reserved_gb:np.float64(140.86328125) - perf/cpu_memory_used_gb:np.float64(195.7408905029297) - actor/lr:np.float64(1e-06) - val-aux/sciknoweval/reward/mean@16:np.float64(0.7519946808510638) - val-aux/sciknoweval/reward/std@16:np.float64(0.05755674390197442) - val-aux/sciknoweval/reward/best@2/mean:np.float64(0.7740212765957446) - val-aux/sciknoweval/reward/best@2/std:np.float64(0.03644780252328991) - val-aux/sciknoweval/reward/worst@2/mean:np.float64(0.7302765957446808) - val-aux/sciknoweval/reward/worst@2/std:np.float64(0.060077687206505734) - val-aux/sciknoweval/reward/maj@2/mean:np.float
|
||
step:71 - global_seqlen/min:21129 - global_seqlen/max:29382 - global_seqlen/minmax_diff:8253 - global_seqlen/balanced_min:25727 - global_seqlen/balanced_max:25731 - global_seqlen/mean:25729.5 - actor/entropy:0.14619044959545135 - perf/mfu/actor_infer:0 - training/rollout_probs_diff_valid:1 - training/rollout_probs_diff_max:0.4265592098236084 - training/rollout_probs_diff_mean:0.002610948635265231 - training/rollout_probs_diff_std:0.009499951265752316 - training/rollout_actor_probs_pearson_corr:0.9987483620643616 - self_distillation/success_group_fraction:0.8125 - self_distillation/success_sample_fraction:0.80859375 - self_distillation/correct_sample_fraction:0.7421875 - self_distillation/feedback_available_fraction:0.0 - self_distillation/feedback_used_fraction:0.0 - self_distillation/reprompt_sample_fraction:0.80859375 - rollout_corr/rollout_is_mean:0.9998709559440613 - rollout_corr/rollout_is_ratio_fraction_high:9.452867743675597e-06 - rollout_corr/rollout_is_ratio_fraction_low:2.835860323102679e-05 - rollout_corr/rollout_is_max:2.090681791305542 - rollout_corr/rollout_is_min:0.2555420696735382 - rollout_corr/rollout_is_std:0.02667061798274517 - rollout_corr/rollout_is_eff_sample_size:0.9992889457086402 - rollout_corr/rollout_is_seq_mean:0.9999041557312012 - rollout_corr/rollout_is_seq_std:0.0017140944255515933 - rollout_corr/rollout_is_seq_max:1.0067784786224365 - rollout_corr/rollout_is_seq_min:0.9939207434654236 - rollout_corr/rollout_is_seq_max_deviation:0.0067784786224365234 - rollout_corr/rollout_is_seq_fraction_high:0.0 - rollout_corr/rollout_is_seq_fraction_low:0.0 - rollout_corr/training_ppl:np.float64(1.157512275967747) - rollout_corr/training_log_ppl:np.float64(0.14429704349458916) - rollout_corr/kl:np.float64(0.0005933385952370251) - rollout_corr/k3_kl:np.float64(0.0004896153571110062) - rollout_corr/rollout_ppl:np.float64(1.156822511460632) - rollout_corr/rollout_log_ppl:np.float64(0.14370370344840921) - rollout_corr/log_ppl_diff:np.float64(0.0005933400461799465) - rollout_corr/log_ppl_abs_diff:np.float64(0.0014040681489859708) - rollout_corr/log_ppl_diff_max:np.float64(0.022363394498825073) - rollout_corr/log_ppl_diff_min:np.float64(-0.007897064089775085) - rollout_corr/ppl_ratio:np.float64(1.0005960373673588) - rollout_corr/chi2_token:np.float64(0.0006810310296714306) - rollout_corr/chi2_seq:np.float64(0.23998616635799408) - actor/pg_loss:np.float64(0.0012072301469743252) - actor/kl_loss:np.float64(0.0) - actor/pg_clipfrac:np.float64(9.832385194386006e-05) - actor/ppo_kl:np.float64(0.00010519745228787514) - actor/pg_clipfrac_lower:np.float64(0.0) - self_distillation/token_reweight_lambda:np.float64(0.5) - self_distillation/token_reweight_eps_w:np.float64(0.20000000000000004) - self_distillation/token_reweight_active_token_fraction:np.float64(0.80859375) - self_distillation/token_reweight_context_sample_fraction:np.float64(0.80859375) - self_distillation/token_reweight_correct_sample_fraction:np.float64(0.7421875) - self_distillation/token_reweight_w_mean:np.float64(22375.126596617047) - self_distillation/token_reweight_w_std:np.float64(455294.88350632007) - self_distillation/token_reweight_w_clipped_mean:np.float64(0.9999144962057471) - self_distillation/token_reweight_w_clip_frac:np.float64(0.023882172885350883) - self_distillation/empty_target_batch:np.float64(0.19140625) - actor/grad_norm:np.float64(0.12107829935848713) - perf/mfu/actor:np.float64(0.057396841037551355) - perf/max_memory_allocated_gb:np.float64(130.98332166671753) - perf/max_memory_reserved_gb:np.float64(140.86328125) - perf/cpu_memory_used_gb:np.float64(195.90234565734863) - actor/lr:np.float64(1e-06) - training/global_step:71 - training/epoch:2 - critic/score/mean:0.7421875 - critic/score/max:1.0 - critic/score/min:0.0 - critic/rewards/mean:0.7421875 - critic/rewards/max:1.0 - critic/rewards/min:0.0 - critic/advantages/mean:-0.0004301054868847132 - critic/advantages/max:0.875 - critic/advantages/min:-0.875 - critic/returns/mean:-0.0004301054868847132 - critic/returns/max:0.875 - critic/returns/min:-0.875 - response_length/mean:413.23
|
||
step:72 - global_seqlen/min:23755 - global_seqlen/max:27582 - global_seqlen/minmax_diff:3827 - global_seqlen/balanced_min:26033 - global_seqlen/balanced_max:26038 - global_seqlen/mean:26036.125 - actor/entropy:0.12893779575824738 - perf/mfu/actor_infer:0 - training/rollout_probs_diff_valid:1 - training/rollout_probs_diff_max:0.5466005206108093 - training/rollout_probs_diff_mean:0.002355077536776662 - training/rollout_probs_diff_std:0.008884367533028126 - training/rollout_actor_probs_pearson_corr:0.9988211989402771 - self_distillation/success_group_fraction:0.84375 - self_distillation/success_sample_fraction:0.8359375 - self_distillation/correct_sample_fraction:0.765625 - self_distillation/feedback_available_fraction:0.0 - self_distillation/feedback_used_fraction:0.0 - self_distillation/reprompt_sample_fraction:0.8359375 - rollout_corr/rollout_is_mean:0.9999760985374451 - rollout_corr/rollout_is_ratio_fraction_high:0.0 - rollout_corr/rollout_is_ratio_fraction_low:9.390641025674995e-06 - rollout_corr/rollout_is_max:1.6415457725524902 - rollout_corr/rollout_is_min:0.4085317850112915 - rollout_corr/rollout_is_std:0.025675423443317413 - rollout_corr/rollout_is_eff_sample_size:0.9993413259754993 - rollout_corr/rollout_is_seq_mean:0.9999563694000244 - rollout_corr/rollout_is_seq_std:0.001469815499149263 - rollout_corr/rollout_is_seq_max:1.0054359436035156 - rollout_corr/rollout_is_seq_min:0.9952075481414795 - rollout_corr/rollout_is_seq_max_deviation:0.005435943603515625 - rollout_corr/rollout_is_seq_fraction_high:0.0 - rollout_corr/rollout_is_seq_fraction_low:0.0 - rollout_corr/training_ppl:np.float64(1.142138805706054) - rollout_corr/training_log_ppl:np.float64(0.13091732577595394) - rollout_corr/kl:np.float64(0.0006154191354865723) - rollout_corr/k3_kl:np.float64(0.0004093494105674722) - rollout_corr/rollout_ppl:np.float64(1.1414216659031808) - rollout_corr/rollout_log_ppl:np.float64(0.13030190639256034) - rollout_corr/log_ppl_diff:np.float64(0.0006154193833936006) - rollout_corr/log_ppl_abs_diff:np.float64(0.001276002629310824) - rollout_corr/log_ppl_diff_max:np.float64(0.0072397589683532715) - rollout_corr/log_ppl_diff_min:np.float64(-0.0041563138365745544) - rollout_corr/ppl_ratio:np.float64(1.0006169544067234) - rollout_corr/chi2_token:np.float64(0.00042544701136648655) - rollout_corr/chi2_seq:np.float64(0.3312300390098244) - actor/pg_loss:np.float64(0.000465520890429616) - actor/kl_loss:np.float64(0.0) - actor/pg_clipfrac:np.float64(0.00012200302489873138) - actor/ppo_kl:np.float64(0.00021202303769474184) - actor/pg_clipfrac_lower:np.float64(0.0) - self_distillation/token_reweight_lambda:np.float64(0.5) - self_distillation/token_reweight_eps_w:np.float64(0.20000000000000004) - self_distillation/token_reweight_active_token_fraction:np.float64(0.8359375) - self_distillation/token_reweight_context_sample_fraction:np.float64(0.8359375) - self_distillation/token_reweight_correct_sample_fraction:np.float64(0.765625) - self_distillation/token_reweight_w_mean:np.float64(15657.842956456356) - self_distillation/token_reweight_w_std:np.float64(324596.33819807204) - self_distillation/token_reweight_w_clipped_mean:np.float64(1.0006162282079458) - self_distillation/token_reweight_w_clip_frac:np.float64(0.014747079403605312) - self_distillation/empty_target_batch:np.float64(0.1640625) - actor/grad_norm:np.float64(0.0753592997789383) - perf/mfu/actor:np.float64(0.05824163963292747) - perf/max_memory_allocated_gb:np.float64(130.98332166671753) - perf/max_memory_reserved_gb:np.float64(140.86328125) - perf/cpu_memory_used_gb:np.float64(195.82327270507812) - actor/lr:np.float64(1e-06) - training/global_step:72 - training/epoch:2 - critic/score/mean:0.765625 - critic/score/max:1.0 - critic/score/min:0.0 - critic/rewards/mean:0.765625 - critic/rewards/max:1.0 - critic/rewards/min:0.0 - critic/advantages/mean:0.002990919165313244 - critic/advantages/max:0.875 - critic/advantages/min:-0.875 - critic/returns/mean:0.002990919165313244 - critic/returns/max:0.875 - critic/returns/min:-0.875 - response_length/mean:415.97265625 - response_length/max
|
||
step:73 - global_seqlen/min:18636 - global_seqlen/max:29660 - global_seqlen/minmax_diff:11024 - global_seqlen/balanced_min:24241 - global_seqlen/balanced_max:24250 - global_seqlen/mean:24247.5 - actor/entropy:0.13270290195941925 - perf/mfu/actor_infer:0 - training/rollout_probs_diff_valid:1 - training/rollout_probs_diff_max:0.23105892539024353 - training/rollout_probs_diff_mean:0.0024152176920324564 - training/rollout_probs_diff_std:0.008945304900407791 - training/rollout_actor_probs_pearson_corr:0.9987912178039551 - self_distillation/success_group_fraction:0.75 - self_distillation/success_sample_fraction:0.74609375 - self_distillation/correct_sample_fraction:0.6796875 - self_distillation/feedback_available_fraction:0.0 - self_distillation/feedback_used_fraction:0.0 - self_distillation/reprompt_sample_fraction:0.74609375 - rollout_corr/rollout_is_mean:0.9999596476554871 - rollout_corr/rollout_is_ratio_fraction_high:0.0 - rollout_corr/rollout_is_ratio_fraction_low:0.0 - rollout_corr/rollout_is_max:1.7426339387893677 - rollout_corr/rollout_is_min:0.5378821492195129 - rollout_corr/rollout_is_std:0.025952504947781563 - rollout_corr/rollout_is_eff_sample_size:0.9993265637082287 - rollout_corr/rollout_is_seq_mean:0.9999666213989258 - rollout_corr/rollout_is_seq_std:0.001694069360382855 - rollout_corr/rollout_is_seq_max:1.009128451347351 - rollout_corr/rollout_is_seq_min:0.9934071898460388 - rollout_corr/rollout_is_seq_max_deviation:0.009128451347351074 - rollout_corr/rollout_is_seq_fraction_high:0.0 - rollout_corr/rollout_is_seq_fraction_low:0.0 - rollout_corr/training_ppl:np.float64(1.1533304639160633) - rollout_corr/training_log_ppl:np.float64(0.13871499759989092) - rollout_corr/kl:np.float64(0.0004179086189708414) - rollout_corr/k3_kl:np.float64(0.0004781059732508197) - rollout_corr/rollout_ppl:np.float64(1.1528317001648247) - rollout_corr/rollout_log_ppl:np.float64(0.13829708714911249) - rollout_corr/log_ppl_diff:np.float64(0.0004179104507784359) - rollout_corr/log_ppl_abs_diff:np.float64(0.0015569298338959925) - rollout_corr/log_ppl_diff_max:np.float64(0.007369577884674072) - rollout_corr/log_ppl_diff_min:np.float64(-0.008280247449874878) - rollout_corr/ppl_ratio:np.float64(1.0004202341660857) - rollout_corr/chi2_token:np.float64(0.0010818280279636383) - rollout_corr/chi2_seq:np.float64(0.5929345737677068) - actor/pg_loss:np.float64(0.0011989542981609702) - actor/kl_loss:np.float64(0.0) - actor/pg_clipfrac:np.float64(6.68577708893281e-05) - actor/ppo_kl:np.float64(-1.3919712605892087e-06) - actor/pg_clipfrac_lower:np.float64(0.0) - self_distillation/token_reweight_lambda:np.float64(0.5) - self_distillation/token_reweight_eps_w:np.float64(0.20000000000000004) - self_distillation/token_reweight_active_token_fraction:np.float64(0.74609375) - self_distillation/token_reweight_context_sample_fraction:np.float64(0.74609375) - self_distillation/token_reweight_correct_sample_fraction:np.float64(0.6796875) - self_distillation/token_reweight_w_mean:np.float64(12149.982365099946) - self_distillation/token_reweight_w_std:np.float64(255379.99605708098) - self_distillation/token_reweight_w_clipped_mean:np.float64(1.0005448078736663) - self_distillation/token_reweight_w_clip_frac:np.float64(0.02245678317558486) - self_distillation/empty_target_batch:np.float64(0.25390625) - actor/grad_norm:np.float64(0.12669192161411047) - perf/mfu/actor:np.float64(0.054887508358529194) - perf/max_memory_allocated_gb:np.float64(130.98332166671753) - perf/max_memory_reserved_gb:np.float64(140.86328125) - perf/cpu_memory_used_gb:np.float64(195.6793441772461) - actor/lr:np.float64(1e-06) - training/global_step:73 - training/epoch:2 - critic/score/mean:0.6796875 - critic/score/max:1.0 - critic/score/min:0.0 - critic/rewards/mean:0.6796875 - critic/rewards/max:1.0 - critic/rewards/min:0.0 - critic/advantages/mean:-0.00045284064253792167 - critic/advantages/max:0.875 - critic/advantages/min:-0.875 - critic/returns/mean:-0.00045284064253792167 - critic/returns/max:0.875 - critic/returns/min:-0.875 - response_length/mean:366.609375 - response_length/max:825.0 -
|
||
step:74 - global_seqlen/min:17738 - global_seqlen/max:28519 - global_seqlen/minmax_diff:10781 - global_seqlen/balanced_min:24352 - global_seqlen/balanced_max:24355 - global_seqlen/mean:24354.0 - actor/entropy:0.1518096625804901 - perf/mfu/actor_infer:0 - training/rollout_probs_diff_valid:1 - training/rollout_probs_diff_max:0.24477601051330566 - training/rollout_probs_diff_mean:0.0026547734159976244 - training/rollout_probs_diff_std:0.009086649864912033 - training/rollout_actor_probs_pearson_corr:0.9989684224128723 - self_distillation/success_group_fraction:0.875 - self_distillation/success_sample_fraction:0.87109375 - self_distillation/correct_sample_fraction:0.8359375 - self_distillation/feedback_available_fraction:0.0 - self_distillation/feedback_used_fraction:0.0 - self_distillation/reprompt_sample_fraction:0.87109375 - rollout_corr/rollout_is_mean:0.9999744296073914 - rollout_corr/rollout_is_ratio_fraction_high:0.0 - rollout_corr/rollout_is_ratio_fraction_low:0.0 - rollout_corr/rollout_is_max:1.9516299962997437 - rollout_corr/rollout_is_min:0.6066008806228638 - rollout_corr/rollout_is_std:0.026919756084680557 - rollout_corr/rollout_is_eff_sample_size:0.9992757324631958 - rollout_corr/rollout_is_seq_mean:1.0000312328338623 - rollout_corr/rollout_is_seq_std:0.0015839202096685767 - rollout_corr/rollout_is_seq_max:1.0063767433166504 - rollout_corr/rollout_is_seq_min:0.9951472282409668 - rollout_corr/rollout_is_seq_max_deviation:0.006376743316650391 - rollout_corr/rollout_is_seq_fraction_high:0.0 - rollout_corr/rollout_is_seq_fraction_low:0.0 - rollout_corr/training_ppl:np.float64(1.1675362568348646) - rollout_corr/training_log_ppl:np.float64(0.15320578197133727) - rollout_corr/kl:np.float64(0.0005907460380143448) - rollout_corr/k3_kl:np.float64(0.00047802696350629503) - rollout_corr/rollout_ppl:np.float64(1.1668550157919526) - rollout_corr/rollout_log_ppl:np.float64(0.15261503440706292) - rollout_corr/log_ppl_diff:np.float64(0.000590747564274352) - rollout_corr/log_ppl_abs_diff:np.float64(0.0013946343315183185) - rollout_corr/log_ppl_diff_max:np.float64(0.005996823310852051) - rollout_corr/log_ppl_diff_min:np.float64(-0.0064388662576675415) - rollout_corr/ppl_ratio:np.float64(1.0005924843717366) - rollout_corr/chi2_token:np.float64(0.000995079753920436) - rollout_corr/chi2_seq:np.float64(0.9846179345622659) - actor/pg_loss:np.float64(0.0004339113365858793) - actor/kl_loss:np.float64(0.0) - actor/pg_clipfrac:np.float64(7.072660491758143e-05) - actor/ppo_kl:np.float64(0.0002481599828243475) - actor/pg_clipfrac_lower:np.float64(0.0) - self_distillation/token_reweight_lambda:np.float64(0.5) - self_distillation/token_reweight_eps_w:np.float64(0.20000000000000004) - self_distillation/token_reweight_active_token_fraction:np.float64(0.87109375) - self_distillation/token_reweight_context_sample_fraction:np.float64(0.87109375) - self_distillation/token_reweight_correct_sample_fraction:np.float64(0.8359375) - self_distillation/token_reweight_w_mean:np.float64(25010.128267090302) - self_distillation/token_reweight_w_std:np.float64(403261.9257906423) - self_distillation/token_reweight_w_clipped_mean:np.float64(0.9998603481799364) - self_distillation/token_reweight_w_clip_frac:np.float64(0.015681787044741213) - self_distillation/empty_target_batch:np.float64(0.12890625) - actor/grad_norm:np.float64(0.07713901996612549) - perf/mfu/actor:np.float64(0.05541995222259979) - perf/max_memory_allocated_gb:np.float64(130.98332166671753) - perf/max_memory_reserved_gb:np.float64(140.86328125) - perf/cpu_memory_used_gb:np.float64(195.67742919921875) - actor/lr:np.float64(1e-06) - training/global_step:74 - training/epoch:2 - critic/score/mean:0.8359375 - critic/score/max:1.0 - critic/score/min:0.0 - critic/rewards/mean:0.8359375 - critic/rewards/max:1.0 - critic/rewards/min:0.0 - critic/advantages/mean:-9.123707059188746e-06 - critic/advantages/max:0.875 - critic/advantages/min:-0.875 - critic/returns/mean:-9.123707059188746e-06 - critic/returns/max:0.875 - critic/returns/min:-0.875 - response_length/mean:374.625 - response_length/max:708.0 - resp
|
||
step:75 - global_seqlen/min:19394 - global_seqlen/max:27755 - global_seqlen/minmax_diff:8361 - global_seqlen/balanced_min:24882 - global_seqlen/balanced_max:24895 - global_seqlen/mean:24891.625 - actor/entropy:0.14456571638584137 - perf/mfu/actor_infer:0 - training/rollout_probs_diff_valid:1 - training/rollout_probs_diff_max:0.41412925720214844 - training/rollout_probs_diff_mean:0.0024925703182816505 - training/rollout_probs_diff_std:0.008859826251864433 - training/rollout_actor_probs_pearson_corr:0.9989786744117737 - self_distillation/success_group_fraction:0.75 - self_distillation/success_sample_fraction:0.74609375 - self_distillation/correct_sample_fraction:0.70703125 - self_distillation/feedback_available_fraction:0.0 - self_distillation/feedback_used_fraction:0.0 - self_distillation/reprompt_sample_fraction:0.74609375 - rollout_corr/rollout_is_mean:1.0000722408294678 - rollout_corr/rollout_is_ratio_fraction_high:0.0 - rollout_corr/rollout_is_ratio_fraction_low:0.0 - rollout_corr/rollout_is_max:1.6829358339309692 - rollout_corr/rollout_is_min:0.5139058828353882 - rollout_corr/rollout_is_std:0.02535303682088852 - rollout_corr/rollout_is_eff_sample_size:0.9993578745193887 - rollout_corr/rollout_is_seq_mean:1.0000704526901245 - rollout_corr/rollout_is_seq_std:0.0013349363580346107 - rollout_corr/rollout_is_seq_max:1.0043613910675049 - rollout_corr/rollout_is_seq_min:0.9966480731964111 - rollout_corr/rollout_is_seq_max_deviation:0.004361391067504883 - rollout_corr/rollout_is_seq_fraction_high:0.0 - rollout_corr/rollout_is_seq_fraction_low:0.0 - rollout_corr/training_ppl:np.float64(1.1619159164838493) - rollout_corr/training_log_ppl:np.float64(0.14637414650496794) - rollout_corr/kl:np.float64(0.0002864610944490753) - rollout_corr/k3_kl:np.float64(0.0003769323916529288) - rollout_corr/rollout_ppl:np.float64(1.1615847297944129) - rollout_corr/rollout_log_ppl:np.float64(0.14608768460311694) - rollout_corr/log_ppl_diff:np.float64(0.0002864619018509984) - rollout_corr/log_ppl_abs_diff:np.float64(0.0011751155252568424) - rollout_corr/log_ppl_diff_max:np.float64(0.004867613315582275) - rollout_corr/log_ppl_diff_min:np.float64(-0.0032870694994926453) - rollout_corr/ppl_ratio:np.float64(1.0002875870559365) - rollout_corr/chi2_token:np.float64(0.0009425606112927198) - rollout_corr/chi2_seq:np.float64(0.3932009725831449) - actor/pg_loss:np.float64(0.0005540411220863461) - actor/kl_loss:np.float64(0.0) - actor/pg_clipfrac:np.float64(3.731543756657629e-05) - actor/ppo_kl:np.float64(2.2481496540827095e-05) - actor/pg_clipfrac_lower:np.float64(0.0) - self_distillation/token_reweight_lambda:np.float64(0.5) - self_distillation/token_reweight_eps_w:np.float64(0.20000000000000004) - self_distillation/token_reweight_active_token_fraction:np.float64(0.74609375) - self_distillation/token_reweight_context_sample_fraction:np.float64(0.74609375) - self_distillation/token_reweight_correct_sample_fraction:np.float64(0.70703125) - self_distillation/token_reweight_w_mean:np.float64(14579.109780266648) - self_distillation/token_reweight_w_std:np.float64(280755.2456506435) - self_distillation/token_reweight_w_clipped_mean:np.float64(1.0002813185565174) - self_distillation/token_reweight_w_clip_frac:np.float64(0.011456080566858873) - self_distillation/empty_target_batch:np.float64(0.25390625) - actor/grad_norm:np.float64(0.07376006245613098) - perf/mfu/actor:np.float64(0.05670890516263739) - perf/max_memory_allocated_gb:np.float64(130.98332166671753) - perf/max_memory_reserved_gb:np.float64(140.86328125) - perf/cpu_memory_used_gb:np.float64(195.6297607421875) - actor/lr:np.float64(1e-06) - training/global_step:75 - training/epoch:2 - critic/score/mean:0.70703125 - critic/score/max:1.0 - critic/score/min:0.0 - critic/rewards/mean:0.70703125 - critic/rewards/max:1.0 - critic/rewards/min:0.0 - critic/advantages/mean:-9.670490544522181e-05 - critic/advantages/max:0.875 - critic/advantages/min:-0.875 - critic/returns/mean:-9.670490544522181e-05 - critic/returns/max:0.875 - critic/returns/min:-0.875 - response_length/mean:383.73828125 - response_length/max:77
|
||
step:76 - global_seqlen/min:19023 - global_seqlen/max:25742 - global_seqlen/minmax_diff:6719 - global_seqlen/balanced_min:22863 - global_seqlen/balanced_max:22870 - global_seqlen/mean:22867.625 - actor/entropy:0.16983351111412048 - perf/mfu/actor_infer:0 - training/rollout_probs_diff_valid:1 - training/rollout_probs_diff_max:0.4560590386390686 - training/rollout_probs_diff_mean:0.0028075729496777058 - training/rollout_probs_diff_std:0.009388630278408527 - training/rollout_actor_probs_pearson_corr:0.9990025758743286 - self_distillation/success_group_fraction:0.78125 - self_distillation/success_sample_fraction:0.77734375 - self_distillation/correct_sample_fraction:0.703125 - self_distillation/feedback_available_fraction:0.0 - self_distillation/feedback_used_fraction:0.0 - self_distillation/reprompt_sample_fraction:0.77734375 - rollout_corr/rollout_is_mean:0.9999856948852539 - rollout_corr/rollout_is_ratio_fraction_high:0.0 - rollout_corr/rollout_is_ratio_fraction_low:2.306140049768146e-05 - rollout_corr/rollout_is_max:1.4818986654281616 - rollout_corr/rollout_is_min:0.3279767334461212 - rollout_corr/rollout_is_std:0.027030237019062042 - rollout_corr/rollout_is_eff_sample_size:0.9992695425945294 - rollout_corr/rollout_is_seq_mean:1.0000501871109009 - rollout_corr/rollout_is_seq_std:0.0017394665628671646 - rollout_corr/rollout_is_seq_max:1.0056047439575195 - rollout_corr/rollout_is_seq_min:0.9949859380722046 - rollout_corr/rollout_is_seq_max_deviation:0.005604743957519531 - rollout_corr/rollout_is_seq_fraction_high:0.0 - rollout_corr/rollout_is_seq_fraction_low:0.0 - rollout_corr/training_ppl:np.float64(1.1861248877830803) - rollout_corr/training_log_ppl:np.float64(0.16663943161984207) - rollout_corr/kl:np.float64(0.0005234517319987475) - rollout_corr/k3_kl:np.float64(0.00048701030445386095) - rollout_corr/rollout_ppl:np.float64(1.185497789643705) - rollout_corr/rollout_log_ppl:np.float64(0.1661159795585263) - rollout_corr/log_ppl_diff:np.float64(0.0005234520613157656) - rollout_corr/log_ppl_abs_diff:np.float64(0.0015148041384236421) - rollout_corr/log_ppl_diff_max:np.float64(0.009437471628189087) - rollout_corr/log_ppl_diff_min:np.float64(-0.004993893206119537) - rollout_corr/ppl_ratio:np.float64(1.0005255860742182) - rollout_corr/chi2_token:np.float64(0.0008696855511516333) - rollout_corr/chi2_seq:np.float64(0.4529962248634547) - actor/pg_loss:np.float64(0.0009462278103455901) - actor/kl_loss:np.float64(0.0) - actor/pg_clipfrac:np.float64(0.0001746799912325514) - actor/ppo_kl:np.float64(0.00016931543271425653) - actor/pg_clipfrac_lower:np.float64(0.0) - self_distillation/token_reweight_lambda:np.float64(0.5) - self_distillation/token_reweight_eps_w:np.float64(0.20000000000000004) - self_distillation/token_reweight_active_token_fraction:np.float64(0.77734375) - self_distillation/token_reweight_context_sample_fraction:np.float64(0.77734375) - self_distillation/token_reweight_correct_sample_fraction:np.float64(0.703125) - self_distillation/token_reweight_w_mean:np.float64(52772.14622266544) - self_distillation/token_reweight_w_std:np.float64(991781.8422375617) - self_distillation/token_reweight_w_clipped_mean:np.float64(1.0002462787088007) - self_distillation/token_reweight_w_clip_frac:np.float64(0.023418084252625704) - self_distillation/empty_target_batch:np.float64(0.22265625) - actor/grad_norm:np.float64(0.1341552957892418) - perf/mfu/actor:np.float64(0.052929615809226865) - perf/max_memory_allocated_gb:np.float64(130.98332166671753) - perf/max_memory_reserved_gb:np.float64(140.86328125) - perf/cpu_memory_used_gb:np.float64(195.67698287963867) - actor/lr:np.float64(1e-06) - training/global_step:76 - training/epoch:2 - critic/score/mean:0.703125 - critic/score/max:1.0 - critic/score/min:0.0 - critic/rewards/mean:0.703125 - critic/rewards/max:1.0 - critic/rewards/min:0.0 - critic/advantages/mean:0.002022196538746357 - critic/advantages/max:0.875 - critic/advantages/min:-0.875 - critic/returns/mean:0.002022196538746357 - critic/returns/max:0.875 - critic/returns/min:-0.875 - response_length/mean:338.76953125 - response_length
|
||
step:77 - global_seqlen/min:21140 - global_seqlen/max:29886 - global_seqlen/minmax_diff:8746 - global_seqlen/balanced_min:26058 - global_seqlen/balanced_max:26063 - global_seqlen/mean:26061.375 - actor/entropy:0.1314040571451187 - perf/mfu/actor_infer:0 - training/rollout_probs_diff_valid:1 - training/rollout_probs_diff_max:0.6118554472923279 - training/rollout_probs_diff_mean:0.0022847233340144157 - training/rollout_probs_diff_std:0.008869555778801441 - training/rollout_actor_probs_pearson_corr:0.9988621473312378 - self_distillation/success_group_fraction:0.625 - self_distillation/success_sample_fraction:0.62109375 - self_distillation/correct_sample_fraction:0.55078125 - self_distillation/feedback_available_fraction:0.0 - self_distillation/feedback_used_fraction:0.0 - self_distillation/reprompt_sample_fraction:0.62109375 - rollout_corr/rollout_is_mean:0.9998553991317749 - rollout_corr/rollout_is_ratio_fraction_high:0.0 - rollout_corr/rollout_is_ratio_fraction_low:9.608455002307892e-06 - rollout_corr/rollout_is_max:1.9780268669128418 - rollout_corr/rollout_is_min:0.16305549442768097 - rollout_corr/rollout_is_std:0.02495972253382206 - rollout_corr/rollout_is_eff_sample_size:0.999377281064019 - rollout_corr/rollout_is_seq_mean:0.9998329877853394 - rollout_corr/rollout_is_seq_std:0.0015425964957103133 - rollout_corr/rollout_is_seq_max:1.008661150932312 - rollout_corr/rollout_is_seq_min:0.9920440316200256 - rollout_corr/rollout_is_seq_max_deviation:0.008661150932312012 - rollout_corr/rollout_is_seq_fraction_high:0.0 - rollout_corr/rollout_is_seq_fraction_low:0.0 - rollout_corr/training_ppl:np.float64(1.1371952700428665) - rollout_corr/training_log_ppl:np.float64(0.12557044533605222) - rollout_corr/kl:np.float64(0.00041996561053991854) - rollout_corr/k3_kl:np.float64(0.0003773384485654674) - rollout_corr/rollout_ppl:np.float64(1.136701645795256) - rollout_corr/rollout_log_ppl:np.float64(0.1251504793690401) - rollout_corr/log_ppl_diff:np.float64(0.00041996596701210365) - rollout_corr/log_ppl_abs_diff:np.float64(0.0013194825223763473) - rollout_corr/log_ppl_diff_max:np.float64(0.008445695042610168) - rollout_corr/log_ppl_diff_min:np.float64(-0.007665909826755524) - rollout_corr/ppl_ratio:np.float64(1.000421627657488) - rollout_corr/chi2_token:np.float64(0.0006627803668379784) - rollout_corr/chi2_seq:np.float64(0.3250666738022119) - actor/pg_loss:np.float64(0.00032114051282405853) - actor/kl_loss:np.float64(0.0) - actor/pg_clipfrac:np.float64(0.00010906540319410851) - actor/ppo_kl:np.float64(-8.278402166705945e-05) - actor/pg_clipfrac_lower:np.float64(0.0) - self_distillation/token_reweight_lambda:np.float64(0.5) - self_distillation/token_reweight_eps_w:np.float64(0.20000000000000004) - self_distillation/token_reweight_active_token_fraction:np.float64(0.62109375) - self_distillation/token_reweight_context_sample_fraction:np.float64(0.62109375) - self_distillation/token_reweight_correct_sample_fraction:np.float64(0.55078125) - self_distillation/token_reweight_w_mean:np.float64(482.4049843470566) - self_distillation/token_reweight_w_std:np.float64(9469.058202488523) - self_distillation/token_reweight_w_clipped_mean:np.float64(1.0003158545587212) - self_distillation/token_reweight_w_clip_frac:np.float64(0.00614475453767227) - self_distillation/empty_target_batch:np.float64(0.37890625) - actor/grad_norm:np.float64(0.06303319334983826) - perf/mfu/actor:np.float64(0.05866944429011463) - perf/max_memory_allocated_gb:np.float64(130.98332166671753) - perf/max_memory_reserved_gb:np.float64(140.86328125) - perf/cpu_memory_used_gb:np.float64(195.73622131347656) - actor/lr:np.float64(1e-06) - training/global_step:77 - training/epoch:2 - critic/score/mean:0.55078125 - critic/score/max:1.0 - critic/score/min:0.0 - critic/rewards/mean:0.55078125 - critic/rewards/max:1.0 - critic/rewards/min:0.0 - critic/advantages/mean:-0.003772519761696458 - critic/advantages/max:0.875 - critic/advantages/min:-0.25 - critic/returns/mean:-0.003772519761696458 - critic/returns/max:0.875 - critic/returns/min:-0.25 - response_length/mean:406.54296875 - response_le
|
||
step:78 - global_seqlen/min:22408 - global_seqlen/max:26542 - global_seqlen/minmax_diff:4134 - global_seqlen/balanced_min:25162 - global_seqlen/balanced_max:25167 - global_seqlen/mean:25165.625 - actor/entropy:0.1404900997877121 - perf/mfu/actor_infer:0 - training/rollout_probs_diff_valid:1 - training/rollout_probs_diff_max:0.39975857734680176 - training/rollout_probs_diff_mean:0.0025989068672060966 - training/rollout_probs_diff_std:0.00935492292046547 - training/rollout_actor_probs_pearson_corr:0.9987855553627014 - self_distillation/success_group_fraction:0.6875 - self_distillation/success_sample_fraction:0.6875 - self_distillation/correct_sample_fraction:0.6640625 - self_distillation/feedback_available_fraction:0.0 - self_distillation/feedback_used_fraction:0.0 - self_distillation/reprompt_sample_fraction:0.6875 - rollout_corr/rollout_is_mean:1.0001541376113892 - rollout_corr/rollout_is_ratio_fraction_high:3.129792457912117e-05 - rollout_corr/rollout_is_ratio_fraction_low:1.0432641829538625e-05 - rollout_corr/rollout_is_max:2.244246006011963 - rollout_corr/rollout_is_min:0.48570868372917175 - rollout_corr/rollout_is_std:0.027534905821084976 - rollout_corr/rollout_is_eff_sample_size:0.9992425223351087 - rollout_corr/rollout_is_seq_mean:1.0001494884490967 - rollout_corr/rollout_is_seq_std:0.0017627255292609334 - rollout_corr/rollout_is_seq_max:1.009850263595581 - rollout_corr/rollout_is_seq_min:0.9946343302726746 - rollout_corr/rollout_is_seq_max_deviation:0.009850263595581055 - rollout_corr/rollout_is_seq_fraction_high:0.0 - rollout_corr/rollout_is_seq_fraction_low:0.0 - rollout_corr/training_ppl:np.float64(1.1456196890212595) - rollout_corr/training_log_ppl:np.float64(0.13408512135356432) - rollout_corr/kl:np.float64(0.0003313364118397999) - rollout_corr/k3_kl:np.float64(0.0004774892564824995) - rollout_corr/rollout_ppl:np.float64(1.1452268385328352) - rollout_corr/rollout_log_ppl:np.float64(0.13375378415366868) - rollout_corr/log_ppl_diff:np.float64(0.0003313371998956427) - rollout_corr/log_ppl_abs_diff:np.float64(0.0014909690798958763) - rollout_corr/log_ppl_diff_max:np.float64(0.009037353098392487) - rollout_corr/log_ppl_diff_min:np.float64(-0.009428516030311584) - rollout_corr/ppl_ratio:np.float64(1.000333433272317) - rollout_corr/chi2_token:np.float64(0.0012364955618977547) - rollout_corr/chi2_seq:np.float64(0.7343874129001051) - actor/pg_loss:np.float64(0.0005279415054246783) - actor/kl_loss:np.float64(0.0) - actor/pg_clipfrac:np.float64(7.377939436992165e-05) - actor/ppo_kl:np.float64(8.827176546333249e-05) - actor/pg_clipfrac_lower:np.float64(0.0) - self_distillation/token_reweight_lambda:np.float64(0.5) - self_distillation/token_reweight_eps_w:np.float64(0.20000000000000004) - self_distillation/token_reweight_active_token_fraction:np.float64(0.6875) - self_distillation/token_reweight_context_sample_fraction:np.float64(0.6875) - self_distillation/token_reweight_correct_sample_fraction:np.float64(0.6640625) - self_distillation/token_reweight_w_mean:np.float64(2963.168327776948) - self_distillation/token_reweight_w_std:np.float64(69523.39454561111) - self_distillation/token_reweight_w_clipped_mean:np.float64(0.9997723049018532) - self_distillation/token_reweight_w_clip_frac:np.float64(0.008337422099430114) - self_distillation/empty_target_batch:np.float64(0.3125) - actor/grad_norm:np.float64(0.05845398269593716) - perf/mfu/actor:np.float64(0.057231997315823215) - perf/max_memory_allocated_gb:np.float64(130.98332166671753) - perf/max_memory_reserved_gb:np.float64(140.86328125) - perf/cpu_memory_used_gb:np.float64(195.64541482925415) - actor/lr:np.float64(1e-06) - training/global_step:78 - training/epoch:2 - critic/score/mean:0.6640625 - critic/score/max:1.0 - critic/score/min:0.0 - critic/rewards/mean:0.6640625 - critic/rewards/max:1.0 - critic/rewards/min:0.0 - critic/advantages/mean:-0.0019535121973603964 - critic/advantages/max:0.625 - critic/advantages/min:-0.875 - critic/returns/mean:-0.0019535121973603964 - critic/returns/max:0.625 - critic/returns/min:-0.875 - response_length/mean:374.42578125 - response_le
|
||
step:79 - global_seqlen/min:17609 - global_seqlen/max:27080 - global_seqlen/minmax_diff:9471 - global_seqlen/balanced_min:24292 - global_seqlen/balanced_max:24296 - global_seqlen/mean:24294.125 - actor/entropy:0.11888319253921509 - perf/mfu/actor_infer:0 - training/rollout_probs_diff_valid:1 - training/rollout_probs_diff_max:0.6836330890655518 - training/rollout_probs_diff_mean:0.0023963868152350187 - training/rollout_probs_diff_std:0.009350299835205078 - training/rollout_actor_probs_pearson_corr:0.9986465573310852 - self_distillation/success_group_fraction:0.8125 - self_distillation/success_sample_fraction:0.8125 - self_distillation/correct_sample_fraction:0.8046875 - self_distillation/feedback_available_fraction:0.0 - self_distillation/feedback_used_fraction:0.0 - self_distillation/reprompt_sample_fraction:0.8125 - rollout_corr/rollout_is_mean:0.9999861717224121 - rollout_corr/rollout_is_ratio_fraction_high:1.0910241144301835e-05 - rollout_corr/rollout_is_ratio_fraction_low:1.0910241144301835e-05 - rollout_corr/rollout_is_max:2.0104925632476807 - rollout_corr/rollout_is_min:0.28233087062835693 - rollout_corr/rollout_is_std:0.02596168965101242 - rollout_corr/rollout_is_eff_sample_size:0.9993264446594591 - rollout_corr/rollout_is_seq_mean:0.9999976754188538 - rollout_corr/rollout_is_seq_std:0.0015455903485417366 - rollout_corr/rollout_is_seq_max:1.0057108402252197 - rollout_corr/rollout_is_seq_min:0.9947482943534851 - rollout_corr/rollout_is_seq_max_deviation:0.0057108402252197266 - rollout_corr/rollout_is_seq_fraction_high:0.0 - rollout_corr/rollout_is_seq_fraction_low:0.0 - rollout_corr/training_ppl:np.float64(1.1248688157647848) - rollout_corr/training_log_ppl:np.float64(0.11555000303633278) - rollout_corr/kl:np.float64(0.00035947103978273276) - rollout_corr/k3_kl:np.float64(0.00038126155526896355) - rollout_corr/rollout_ppl:np.float64(1.124436704441905) - rollout_corr/rollout_log_ppl:np.float64(0.11519053130905377) - rollout_corr/log_ppl_diff:np.float64(0.0003594717272790149) - rollout_corr/log_ppl_abs_diff:np.float64(0.0012188185646664351) - rollout_corr/log_ppl_diff_max:np.float64(0.004965603351593018) - rollout_corr/log_ppl_diff_min:np.float64(-0.003954470157623291) - rollout_corr/ppl_ratio:np.float64(1.000360760372132) - rollout_corr/chi2_token:np.float64(0.0008095051161944866) - rollout_corr/chi2_seq:np.float64(0.17021930450573564) - actor/pg_loss:np.float64(0.000181176932528615) - actor/kl_loss:np.float64(0.0) - actor/pg_clipfrac:np.float64(1.7104167454817798e-05) - actor/ppo_kl:np.float64(8.604286239943804e-06) - actor/pg_clipfrac_lower:np.float64(0.0) - self_distillation/token_reweight_lambda:np.float64(0.5) - self_distillation/token_reweight_eps_w:np.float64(0.20000000000000004) - self_distillation/token_reweight_active_token_fraction:np.float64(0.8125) - self_distillation/token_reweight_context_sample_fraction:np.float64(0.8125) - self_distillation/token_reweight_correct_sample_fraction:np.float64(0.8046875) - self_distillation/token_reweight_w_mean:np.float64(3.382794478908181) - self_distillation/token_reweight_w_std:np.float64(40.678365995234344) - self_distillation/token_reweight_w_clipped_mean:np.float64(0.9996934912633151) - self_distillation/token_reweight_w_clip_frac:np.float64(0.006391824339516461) - self_distillation/empty_target_batch:np.float64(0.1875) - actor/grad_norm:np.float64(0.04638487892225385) - perf/mfu/actor:np.float64(0.054837898195592656) - perf/max_memory_allocated_gb:np.float64(130.98332166671753) - perf/max_memory_reserved_gb:np.float64(140.86328125) - perf/cpu_memory_used_gb:np.float64(195.43430709838867) - actor/lr:np.float64(1e-06) - training/global_step:79 - training/epoch:3 - critic/score/mean:0.8046875 - critic/score/max:1.0 - critic/score/min:0.0 - critic/rewards/mean:0.8046875 - critic/rewards/max:1.0 - critic/rewards/min:0.0 - critic/advantages/mean:-0.0002863938279915601 - critic/advantages/max:0.125 - critic/advantages/min:-0.875 - critic/returns/mean:-0.0002863938279915601 - critic/returns/max:0.125 - critic/returns/min:-0.875 - response_length/mean:358.03515625 - resp
|
||
test_gen_batch meta info: {'eos_token_id': 151645, 'pad_token_id': 151643, 'recompute_log_prob': False, 'do_sample': True, 'validate': True, 'global_steps': 80}
|
||
validation generation end
|
||
[prompt] system
|
||
|
||
Given a question and four options, please select the right answer. Respond in the following format:
|
||
<reasoning>
|
||
...
|
||
</reasoning>
|
||
<answer>
|
||
...
|
||
</answer>
|
||
|
||
For the answer, only output the letter corresponding to the correct option (A, B, C, or D), and nothing else. Do not restate the answer text. For example, if the answer is "A", just output:
|
||
<answer>
|
||
A
|
||
</answer>
|
||
|
||
user
|
||
Given the following crystal structure parameters for the material RbLa9(IrO6)4 (Material ID: mp-560657), calculate the volume of the unit cell (in ų).
|
||
|
||
Lattice Parameters:
|
||
|
||
a = 7.82 Å
|
||
b = 7.82 Å
|
||
c = 17.88 Å
|
||
α = 90.00º
|
||
β = 90.00º
|
||
γ = 90.00º
|
||
|
||
|
||
|
||
A: 1025.67 ų
|
||
|
||
B: 1094.31 ų
|
||
|
||
C: 1200.45 ų
|
||
|
||
D: 1150.78 ų
|
||
|
||
Please reason step by step.
|
||
assistant
|
||
<think>
|
||
|
||
</think>
|
||
|
||
|
||
[response] <reasoning>
|
||
The given crystal structure has orthorhombic symmetry, as indicated by the right angles (α = β = γ = 90.00º). For an orthorhombic system, the volume of the unit cell is calculated using the formula:
|
||
|
||
$$ V = a \times b \times c $$
|
||
|
||
Substituting the given lattice parameters:
|
||
|
||
$$ V = 7.82 \times 7.82 \times 17.88 $$
|
||
|
||
First, calculate $ 7.82 \times 7.82 $:
|
||
|
||
$$ 7.82 \times 7.82 = 61.1524 $$
|
||
|
||
Next, multiply by 17.88:
|
||
|
||
$$ 61.1524 \times 17.88 = 1094.31 $$
|
||
|
||
Thus, the volume of the unit cell is 1094.31 ų.
|
||
</reasoning>
|
||
<answer>
|
||
B
|
||
</answer>
|
||
[ground_truth] B
|
||
[score] 1.0
|
||
[acc] 1.0
|
||
[pred] B
|
||
[incorrect_format] 1
|
||
[feedback]
|
||
local_global_step_folder: /mnt/mole/SDPO/L2T/checkpoints/datasets/sciknoweval/material/qwen3gen-material-RLSD_TR-Qwen-Qwen3-8B-mbs8-decay0-tr0.1-train32-rollout8-lr1e-6-vllm0.8/global_step_80
|
||
step:80 - global_seqlen/min:21209 - global_seqlen/max:28791 - global_seqlen/minmax_diff:7582 - global_seqlen/balanced_min:24840 - global_seqlen/balanced_max:24842 - global_seqlen/mean:24841.125 - actor/entropy:0.1547521948814392 - perf/mfu/actor_infer:0 - training/rollout_probs_diff_valid:1 - training/rollout_probs_diff_max:0.5170373916625977 - training/rollout_probs_diff_mean:0.002725405152887106 - training/rollout_probs_diff_std:0.009661387652158737 - training/rollout_actor_probs_pearson_corr:0.998755156993866 - self_distillation/success_group_fraction:0.78125 - self_distillation/success_sample_fraction:0.7734375 - self_distillation/correct_sample_fraction:0.68359375 - self_distillation/feedback_available_fraction:0.0 - self_distillation/feedback_used_fraction:0.0 - self_distillation/reprompt_sample_fraction:0.7734375 - rollout_corr/rollout_is_mean:1.0000289678573608 - rollout_corr/rollout_is_ratio_fraction_high:0.0 - rollout_corr/rollout_is_ratio_fraction_low:2.980181852763053e-05 - rollout_corr/rollout_is_max:1.6338125467300415 - rollout_corr/rollout_is_min:0.2606707215309143 - rollout_corr/rollout_is_std:0.027992188930511475 - rollout_corr/rollout_is_eff_sample_size:0.9992170508295175 - rollout_corr/rollout_is_seq_mean:0.9999718070030212 - rollout_corr/rollout_is_seq_std:0.0015326400753110647 - rollout_corr/rollout_is_seq_max:1.0061142444610596 - rollout_corr/rollout_is_seq_min:0.9952588081359863 - rollout_corr/rollout_is_seq_max_deviation:0.00611424446105957 - rollout_corr/rollout_is_seq_fraction_high:0.0 - rollout_corr/rollout_is_seq_fraction_low:0.0 - rollout_corr/training_ppl:np.float64(1.1599856242537498) - rollout_corr/training_log_ppl:np.float64(0.14582765987870516) - rollout_corr/kl:np.float64(0.0004167129488124033) - rollout_corr/k3_kl:np.float64(0.00045200509000409284) - rollout_corr/rollout_ppl:np.float64(1.159479976631701) - rollout_corr/rollout_log_ppl:np.float64(0.1454109457845334) - rollout_corr/log_ppl_diff:np.float64(0.0004167140941717662) - rollout_corr/log_ppl_abs_diff:np.float64(0.0013146949859219603) - rollout_corr/log_ppl_diff_max:np.float64(0.006044149398803711) - rollout_corr/log_ppl_diff_min:np.float64(-0.0035454928874969482) - rollout_corr/ppl_ratio:np.float64(1.0004181617405266) - rollout_corr/chi2_token:np.float64(0.0009744579438120127) - rollout_corr/chi2_seq:np.float64(0.3122515210416168) - actor/pg_loss:np.float64(0.0007777587743476033) - actor/kl_loss:np.float64(0.0) - actor/pg_clipfrac:np.float64(0.0001616802319404087) - actor/ppo_kl:np.float64(-1.446276943450897e-05) - actor/pg_clipfrac_lower:np.float64(0.0) - self_distillation/token_reweight_lambda:np.float64(0.5) - self_distillation/token_reweight_eps_w:np.float64(0.20000000000000004) - self_distillation/token_reweight_active_token_fraction:np.float64(0.7734375) - self_distillation/token_reweight_context_sample_fraction:np.float64(0.7734375) - self_distillation/token_reweight_correct_sample_fraction:np.float64(0.68359375) - self_distillation/token_reweight_w_mean:np.float64(1290.9808493084274) - self_distillation/token_reweight_w_std:np.float64(30826.316274836252) - self_distillation/token_reweight_w_clipped_mean:np.float64(1.0009112362749875) - self_distillation/token_reweight_w_clip_frac:np.float64(0.020249034860171378) - self_distillation/empty_target_batch:np.float64(0.2265625) - actor/grad_norm:np.float64(0.08906365092843771) - perf/mfu/actor:np.float64(0.056279375927647174) - perf/max_memory_allocated_gb:np.float64(130.98332166671753) - perf/max_memory_reserved_gb:np.float64(140.86328125) - perf/cpu_memory_used_gb:np.float64(195.7968292236328) - actor/lr:np.float64(1e-06) - val-aux/sciknoweval/reward/mean@16:np.float64(0.7446808510638298) - val-aux/sciknoweval/reward/std@16:np.float64(0.05538506140454391) - val-aux/sciknoweval/reward/best@2/mean:np.float64(0.7648191489361702) - val-aux/sciknoweval/reward/best@2/std:np.float64(0.041079114908916525) - val-aux/sciknoweval/reward/worst@2/mean:np.float64(0.7239255319148935) - val-aux/sciknoweval/reward/worst@2/std:np.float64(0.052388701912266876) - val-aux/sciknoweval/reward/maj@
|
||
step:81 - global_seqlen/min:17496 - global_seqlen/max:28019 - global_seqlen/minmax_diff:10523 - global_seqlen/balanced_min:22611 - global_seqlen/balanced_max:22618 - global_seqlen/mean:22615.375 - actor/entropy:0.12404749542474747 - perf/mfu/actor_infer:0 - training/rollout_probs_diff_valid:1 - training/rollout_probs_diff_max:0.9216691255569458 - training/rollout_probs_diff_mean:0.002631731564179063 - training/rollout_probs_diff_std:0.010278254747390747 - training/rollout_actor_probs_pearson_corr:0.9983770847320557 - self_distillation/success_group_fraction:0.78125 - self_distillation/success_sample_fraction:0.78125 - self_distillation/correct_sample_fraction:0.71875 - self_distillation/feedback_available_fraction:0.0 - self_distillation/feedback_used_fraction:0.0 - self_distillation/reprompt_sample_fraction:0.78125 - rollout_corr/rollout_is_mean:0.9999887347221375 - rollout_corr/rollout_is_ratio_fraction_high:0.0 - rollout_corr/rollout_is_ratio_fraction_low:1.1917956726392731e-05 - rollout_corr/rollout_is_max:1.6952898502349854 - rollout_corr/rollout_is_min:0.07604635506868362 - rollout_corr/rollout_is_std:0.027463378384709358 - rollout_corr/rollout_is_eff_sample_size:0.999246331269986 - rollout_corr/rollout_is_seq_mean:1.0000253915786743 - rollout_corr/rollout_is_seq_std:0.0017810006393119693 - rollout_corr/rollout_is_seq_max:1.007685899734497 - rollout_corr/rollout_is_seq_min:0.9950356483459473 - rollout_corr/rollout_is_seq_max_deviation:0.00768589973449707 - rollout_corr/rollout_is_seq_fraction_high:0.0 - rollout_corr/rollout_is_seq_fraction_low:0.0 - rollout_corr/training_ppl:np.float64(1.1329951421357691) - rollout_corr/training_log_ppl:np.float64(0.12217923940625042) - rollout_corr/kl:np.float64(0.00022328235970903165) - rollout_corr/k3_kl:np.float64(0.0005613560395403283) - rollout_corr/rollout_ppl:np.float64(1.13274285197258) - rollout_corr/rollout_log_ppl:np.float64(0.12195595684170257) - rollout_corr/log_ppl_diff:np.float64(0.00022328256454784423) - rollout_corr/log_ppl_abs_diff:np.float64(0.0015974190609995276) - rollout_corr/log_ppl_diff_max:np.float64(0.006190508604049683) - rollout_corr/log_ppl_diff_min:np.float64(-0.007130444049835205) - rollout_corr/ppl_ratio:np.float64(1.0002254962455481) - rollout_corr/chi2_token:np.float64(0.0018780133686959743) - rollout_corr/chi2_seq:np.float64(0.5290436635259539) - actor/pg_loss:np.float64(0.0008224116172641516) - actor/kl_loss:np.float64(0.0) - actor/pg_clipfrac:np.float64(0.00012829220668209018) - actor/ppo_kl:np.float64(-0.00017505270996531763) - actor/pg_clipfrac_lower:np.float64(0.0) - self_distillation/token_reweight_lambda:np.float64(0.5) - self_distillation/token_reweight_eps_w:np.float64(0.20000000000000004) - self_distillation/token_reweight_active_token_fraction:np.float64(0.78125) - self_distillation/token_reweight_context_sample_fraction:np.float64(0.78125) - self_distillation/token_reweight_correct_sample_fraction:np.float64(0.71875) - self_distillation/token_reweight_w_mean:np.float64(24058.840261494508) - self_distillation/token_reweight_w_std:np.float64(351943.8947785102) - self_distillation/token_reweight_w_clipped_mean:np.float64(0.9996452361810952) - self_distillation/token_reweight_w_clip_frac:np.float64(0.024172404606360942) - self_distillation/empty_target_batch:np.float64(0.21875) - actor/grad_norm:np.float64(0.14087015390396118) - perf/mfu/actor:np.float64(0.051926130920910266) - perf/max_memory_allocated_gb:np.float64(130.98332166671753) - perf/max_memory_reserved_gb:np.float64(140.86328125) - perf/cpu_memory_used_gb:np.float64(196.75435638427734) - actor/lr:np.float64(1e-06) - training/global_step:81 - training/epoch:3 - critic/score/mean:0.71875 - critic/score/max:1.0 - critic/score/min:0.0 - critic/rewards/mean:0.71875 - critic/rewards/max:1.0 - critic/rewards/min:0.0 - critic/advantages/mean:-0.002596624894067645 - critic/advantages/max:0.75 - critic/advantages/min:-0.875 - critic/returns/mean:-0.002596624894067645 - critic/returns/max:0.75 - critic/returns/min:-0.875 - response_length/mean:327.76171875 - response_length/max:711.0 - re
|
||
step:82 - global_seqlen/min:21426 - global_seqlen/max:26718 - global_seqlen/minmax_diff:5292 - global_seqlen/balanced_min:23464 - global_seqlen/balanced_max:23474 - global_seqlen/mean:23472.0 - actor/entropy:0.14597472548484802 - perf/mfu/actor_infer:0 - training/rollout_probs_diff_valid:1 - training/rollout_probs_diff_max:0.3772715628147125 - training/rollout_probs_diff_mean:0.0027662990614771843 - training/rollout_probs_diff_std:0.009847620502114296 - training/rollout_actor_probs_pearson_corr:0.9987053275108337 - self_distillation/success_group_fraction:0.8125 - self_distillation/success_sample_fraction:0.80859375 - self_distillation/correct_sample_fraction:0.734375 - self_distillation/feedback_available_fraction:0.0 - self_distillation/feedback_used_fraction:0.0 - self_distillation/reprompt_sample_fraction:0.80859375 - rollout_corr/rollout_is_mean:0.9999840259552002 - rollout_corr/rollout_is_ratio_fraction_high:0.0 - rollout_corr/rollout_is_ratio_fraction_low:2.2500225895782933e-05 - rollout_corr/rollout_is_max:1.758644700050354 - rollout_corr/rollout_is_min:0.39373767375946045 - rollout_corr/rollout_is_std:0.028364475816488266 - rollout_corr/rollout_is_eff_sample_size:0.9991961032747029 - rollout_corr/rollout_is_seq_mean:0.999990701675415 - rollout_corr/rollout_is_seq_std:0.0016806689091026783 - rollout_corr/rollout_is_seq_max:1.0044842958450317 - rollout_corr/rollout_is_seq_min:0.9935813546180725 - rollout_corr/rollout_is_seq_max_deviation:0.00641864538192749 - rollout_corr/rollout_is_seq_fraction_high:0.0 - rollout_corr/rollout_is_seq_fraction_low:0.0 - rollout_corr/training_ppl:np.float64(1.1481548310257494) - rollout_corr/training_log_ppl:np.float64(0.13636921173747396) - rollout_corr/kl:np.float64(0.0006211503117583561) - rollout_corr/k3_kl:np.float64(0.0004820507922715933) - rollout_corr/rollout_ppl:np.float64(1.1474357061088085) - rollout_corr/rollout_log_ppl:np.float64(0.1357480596198002) - rollout_corr/log_ppl_diff:np.float64(0.0006211521176737733) - rollout_corr/log_ppl_abs_diff:np.float64(0.0015310409144149162) - rollout_corr/log_ppl_diff_max:np.float64(0.00695037841796875) - rollout_corr/log_ppl_diff_min:np.float64(-0.008269622921943665) - rollout_corr/ppl_ratio:np.float64(1.0006232312880456) - rollout_corr/chi2_token:np.float64(0.0006954430136829615) - rollout_corr/chi2_seq:np.float64(0.27697336371056736) - actor/pg_loss:np.float64(0.0010938638588413596) - actor/kl_loss:np.float64(0.0) - actor/pg_clipfrac:np.float64(0.00014887547422404168) - actor/ppo_kl:np.float64(0.00020041486293642663) - actor/pg_clipfrac_lower:np.float64(0.0) - self_distillation/token_reweight_lambda:np.float64(0.5) - self_distillation/token_reweight_eps_w:np.float64(0.20000000000000004) - self_distillation/token_reweight_active_token_fraction:np.float64(0.80859375) - self_distillation/token_reweight_context_sample_fraction:np.float64(0.80859375) - self_distillation/token_reweight_correct_sample_fraction:np.float64(0.734375) - self_distillation/token_reweight_w_mean:np.float64(725.874184731394) - self_distillation/token_reweight_w_std:np.float64(16585.61214940023) - self_distillation/token_reweight_w_clipped_mean:np.float64(1.0002405804116279) - self_distillation/token_reweight_w_clip_frac:np.float64(0.02034169973921962) - self_distillation/empty_target_batch:np.float64(0.19140625) - actor/grad_norm:np.float64(0.10360464476980269) - perf/mfu/actor:np.float64(0.05363463253175198) - perf/max_memory_allocated_gb:np.float64(130.98332166671753) - perf/max_memory_reserved_gb:np.float64(140.86328125) - perf/cpu_memory_used_gb:np.float64(196.11970901489258) - actor/lr:np.float64(1e-06) - training/global_step:82 - training/epoch:3 - critic/score/mean:0.734375 - critic/score/max:1.0 - critic/score/min:0.0 - critic/rewards/mean:0.734375 - critic/rewards/max:1.0 - critic/rewards/min:0.0 - critic/advantages/mean:0.004921924322843552 - critic/advantages/max:0.875 - critic/advantages/min:-0.875 - critic/returns/mean:0.004921924322843552 - critic/returns/max:0.875 - critic/returns/min:-0.875 - response_length/mean:347.21875 - response_length/max:830
|
||
step:83 - global_seqlen/min:21063 - global_seqlen/max:24614 - global_seqlen/minmax_diff:3551 - global_seqlen/balanced_min:23076 - global_seqlen/balanced_max:23082 - global_seqlen/mean:23080.75 - actor/entropy:0.12527549266815186 - perf/mfu/actor_infer:0 - training/rollout_probs_diff_valid:1 - training/rollout_probs_diff_max:0.20161855220794678 - training/rollout_probs_diff_mean:0.0025815563276410103 - training/rollout_probs_diff_std:0.009686398319900036 - training/rollout_actor_probs_pearson_corr:0.9985157251358032 - self_distillation/success_group_fraction:0.71875 - self_distillation/success_sample_fraction:0.71484375 - self_distillation/correct_sample_fraction:0.68359375 - self_distillation/feedback_available_fraction:0.0 - self_distillation/feedback_used_fraction:0.0 - self_distillation/reprompt_sample_fraction:0.71484375 - rollout_corr/rollout_is_mean:0.999979555606842 - rollout_corr/rollout_is_ratio_fraction_high:0.0 - rollout_corr/rollout_is_ratio_fraction_low:1.1838522368634585e-05 - rollout_corr/rollout_is_max:1.6963958740234375 - rollout_corr/rollout_is_min:0.26981014013290405 - rollout_corr/rollout_is_std:0.027797745540738106 - rollout_corr/rollout_is_eff_sample_size:0.9992278820119855 - rollout_corr/rollout_is_seq_mean:0.9999013543128967 - rollout_corr/rollout_is_seq_std:0.001572471926920116 - rollout_corr/rollout_is_seq_max:1.0038666725158691 - rollout_corr/rollout_is_seq_min:0.9941545724868774 - rollout_corr/rollout_is_seq_max_deviation:0.005845427513122559 - rollout_corr/rollout_is_seq_fraction_high:0.0 - rollout_corr/rollout_is_seq_fraction_low:0.0 - rollout_corr/training_ppl:np.float64(1.129922913853079) - rollout_corr/training_log_ppl:np.float64(0.12065385213645641) - rollout_corr/kl:np.float64(0.0004963102201873681) - rollout_corr/k3_kl:np.float64(0.00042870396250194176) - rollout_corr/rollout_ppl:np.float64(1.1293713334016502) - rollout_corr/rollout_log_ppl:np.float64(0.12015754160529468) - rollout_corr/log_ppl_diff:np.float64(0.0004963105311617255) - rollout_corr/log_ppl_abs_diff:np.float64(0.0013216602674219757) - rollout_corr/log_ppl_diff_max:np.float64(0.0056051164865493774) - rollout_corr/log_ppl_diff_min:np.float64(-0.005131259560585022) - rollout_corr/ppl_ratio:np.float64(1.00049793231301) - rollout_corr/chi2_token:np.float64(0.0007408871315419674) - rollout_corr/chi2_seq:np.float64(0.5475682236719877) - actor/pg_loss:np.float64(0.0003311014734208584) - actor/kl_loss:np.float64(0.0) - actor/pg_clipfrac:np.float64(8.298924194605206e-05) - actor/ppo_kl:np.float64(-1.5865895953481868e-05) - actor/pg_clipfrac_lower:np.float64(0.0) - self_distillation/token_reweight_lambda:np.float64(0.5) - self_distillation/token_reweight_eps_w:np.float64(0.20000000000000004) - self_distillation/token_reweight_active_token_fraction:np.float64(0.71484375) - self_distillation/token_reweight_context_sample_fraction:np.float64(0.71484375) - self_distillation/token_reweight_correct_sample_fraction:np.float64(0.68359375) - self_distillation/token_reweight_w_mean:np.float64(5976.837657049531) - self_distillation/token_reweight_w_std:np.float64(105540.45544876461) - self_distillation/token_reweight_w_clipped_mean:np.float64(1.0001530186273158) - self_distillation/token_reweight_w_clip_frac:np.float64(0.008285001036711037) - self_distillation/empty_target_batch:np.float64(0.28515625) - actor/grad_norm:np.float64(0.05502156773582101) - perf/mfu/actor:np.float64(0.05255484396903881) - perf/max_memory_allocated_gb:np.float64(130.98332166671753) - perf/max_memory_reserved_gb:np.float64(140.86328125) - perf/cpu_memory_used_gb:np.float64(196.24004364013672) - actor/lr:np.float64(1e-06) - training/global_step:83 - training/epoch:3 - critic/score/mean:0.68359375 - critic/score/max:1.0 - critic/score/min:0.0 - critic/rewards/mean:0.68359375 - critic/rewards/max:1.0 - critic/rewards/min:0.0 - critic/advantages/mean:0.001291878754273057 - critic/advantages/max:0.875 - critic/advantages/min:-0.75 - critic/returns/mean:0.001291878754273057 - critic/returns/max:0.875 - critic/returns/min:-0.75 - response_length/mean:329.9609375 - response
|
||
step:84 - global_seqlen/min:19050 - global_seqlen/max:27314 - global_seqlen/minmax_diff:8264 - global_seqlen/balanced_min:23791 - global_seqlen/balanced_max:23802 - global_seqlen/mean:23799.125 - actor/entropy:0.11515828967094421 - perf/mfu/actor_infer:0 - training/rollout_probs_diff_valid:1 - training/rollout_probs_diff_max:0.1930849552154541 - training/rollout_probs_diff_mean:0.0024084453471004963 - training/rollout_probs_diff_std:0.009149638004601002 - training/rollout_actor_probs_pearson_corr:0.998576283454895 - self_distillation/success_group_fraction:0.84375 - self_distillation/success_sample_fraction:0.84375 - self_distillation/correct_sample_fraction:0.8203125 - self_distillation/feedback_available_fraction:0.0 - self_distillation/feedback_used_fraction:0.0 - self_distillation/reprompt_sample_fraction:0.84375 - rollout_corr/rollout_is_mean:1.000038743019104 - rollout_corr/rollout_is_ratio_fraction_high:0.0 - rollout_corr/rollout_is_ratio_fraction_low:0.0 - rollout_corr/rollout_is_max:1.6482936143875122 - rollout_corr/rollout_is_min:0.616061270236969 - rollout_corr/rollout_is_std:0.02596168965101242 - rollout_corr/rollout_is_eff_sample_size:0.9993265637082287 - rollout_corr/rollout_is_seq_mean:1.0000572204589844 - rollout_corr/rollout_is_seq_std:0.0016339186113327742 - rollout_corr/rollout_is_seq_max:1.0067466497421265 - rollout_corr/rollout_is_seq_min:0.9952927827835083 - rollout_corr/rollout_is_seq_max_deviation:0.006746649742126465 - rollout_corr/rollout_is_seq_fraction_high:0.0 - rollout_corr/rollout_is_seq_fraction_low:0.0 - rollout_corr/training_ppl:np.float64(1.1227850238792598) - rollout_corr/training_log_ppl:np.float64(0.11481493657265673) - rollout_corr/kl:np.float64(0.0003574394741372444) - rollout_corr/k3_kl:np.float64(0.0003687933111855557) - rollout_corr/rollout_ppl:np.float64(1.1223909854888916) - rollout_corr/rollout_log_ppl:np.float64(0.11445749627091573) - rollout_corr/log_ppl_diff:np.float64(0.0003574403017410077) - rollout_corr/log_ppl_abs_diff:np.float64(0.0012601374110090546) - rollout_corr/log_ppl_diff_max:np.float64(0.004564203321933746) - rollout_corr/log_ppl_diff_min:np.float64(-0.006322316825389862) - rollout_corr/ppl_ratio:np.float64(1.0003587694372982) - rollout_corr/chi2_token:np.float64(0.0007815023418515921) - rollout_corr/chi2_seq:np.float64(0.2785968375392258) - actor/pg_loss:np.float64(0.0006993752904236317) - actor/kl_loss:np.float64(0.0) - actor/pg_clipfrac:np.float64(5.200197028898401e-05) - actor/ppo_kl:np.float64(6.108405310456533e-05) - actor/pg_clipfrac_lower:np.float64(0.0) - self_distillation/token_reweight_lambda:np.float64(0.5) - self_distillation/token_reweight_eps_w:np.float64(0.20000000000000004) - self_distillation/token_reweight_active_token_fraction:np.float64(0.84375) - self_distillation/token_reweight_context_sample_fraction:np.float64(0.84375) - self_distillation/token_reweight_correct_sample_fraction:np.float64(0.8203125) - self_distillation/token_reweight_w_mean:np.float64(27.973567307693884) - self_distillation/token_reweight_w_std:np.float64(398.3262187499204) - self_distillation/token_reweight_w_clipped_mean:np.float64(0.999827864812687) - self_distillation/token_reweight_w_clip_frac:np.float64(0.009426361415535212) - self_distillation/empty_target_batch:np.float64(0.15625) - actor/grad_norm:np.float64(0.08874382358044386) - perf/mfu/actor:np.float64(0.05394679103021711) - perf/max_memory_allocated_gb:np.float64(130.98332166671753) - perf/max_memory_reserved_gb:np.float64(140.86328125) - perf/cpu_memory_used_gb:np.float64(196.16954803466797) - actor/lr:np.float64(1e-06) - training/global_step:84 - training/epoch:3 - critic/score/mean:0.8203125 - critic/score/max:1.0 - critic/score/min:0.0 - critic/rewards/mean:0.8203125 - critic/rewards/max:1.0 - critic/rewards/min:0.0 - critic/advantages/mean:0.002152180066332221 - critic/advantages/max:0.375 - critic/advantages/min:-0.875 - critic/returns/mean:0.002152180066332221 - critic/returns/max:0.375 - critic/returns/min:-0.875 - response_length/mean:351.66015625 - response_length/max:748.0 - response_length/min:
|
||
step:85 - global_seqlen/min:19911 - global_seqlen/max:26535 - global_seqlen/minmax_diff:6624 - global_seqlen/balanced_min:23181 - global_seqlen/balanced_max:23236 - global_seqlen/mean:23189.5 - actor/entropy:0.1488744616508484 - perf/mfu/actor_infer:0 - training/rollout_probs_diff_valid:1 - training/rollout_probs_diff_max:0.5948741436004639 - training/rollout_probs_diff_mean:0.0026586854364722967 - training/rollout_probs_diff_std:0.009458235464990139 - training/rollout_actor_probs_pearson_corr:0.9988400936126709 - self_distillation/success_group_fraction:0.75 - self_distillation/success_sample_fraction:0.75 - self_distillation/correct_sample_fraction:0.69140625 - self_distillation/feedback_available_fraction:0.0 - self_distillation/feedback_used_fraction:0.0 - self_distillation/reprompt_sample_fraction:0.75 - rollout_corr/rollout_is_mean:0.999997615814209 - rollout_corr/rollout_is_ratio_fraction_high:0.0 - rollout_corr/rollout_is_ratio_fraction_low:1.1743153663701378e-05 - rollout_corr/rollout_is_max:1.8089348077774048 - rollout_corr/rollout_is_min:0.2346915602684021 - rollout_corr/rollout_is_std:0.02704346552491188 - rollout_corr/rollout_is_eff_sample_size:0.999269304524189 - rollout_corr/rollout_is_seq_mean:0.999980628490448 - rollout_corr/rollout_is_seq_std:0.0015972725814208388 - rollout_corr/rollout_is_seq_max:1.004743218421936 - rollout_corr/rollout_is_seq_min:0.9943519830703735 - rollout_corr/rollout_is_seq_max_deviation:0.005648016929626465 - rollout_corr/rollout_is_seq_fraction_high:0.0 - rollout_corr/rollout_is_seq_fraction_low:0.0 - rollout_corr/training_ppl:np.float64(1.1522011863999069) - rollout_corr/training_log_ppl:np.float64(0.13857831715722568) - rollout_corr/kl:np.float64(0.0004144432389345809) - rollout_corr/k3_kl:np.float64(0.00039726549428564795) - rollout_corr/rollout_ppl:np.float64(1.1517116916365921) - rollout_corr/rollout_log_ppl:np.float64(0.13816387339466019) - rollout_corr/log_ppl_diff:np.float64(0.0004144437625654973) - rollout_corr/log_ppl_abs_diff:np.float64(0.001279399984923657) - rollout_corr/log_ppl_diff_max:np.float64(0.006227068603038788) - rollout_corr/log_ppl_diff_min:np.float64(-0.0034830868244171143) - rollout_corr/ppl_ratio:np.float64(1.0004158781375736) - rollout_corr/chi2_token:np.float64(0.0007748533971607685) - rollout_corr/chi2_seq:np.float64(0.20484365243464708) - actor/pg_loss:np.float64(0.0009036622941493988) - actor/kl_loss:np.float64(0.0) - actor/pg_clipfrac:np.float64(5.688434430339839e-05) - actor/ppo_kl:np.float64(1.6656271835113756e-05) - actor/pg_clipfrac_lower:np.float64(0.0) - self_distillation/token_reweight_lambda:np.float64(0.5) - self_distillation/token_reweight_eps_w:np.float64(0.20000000000000004) - self_distillation/token_reweight_active_token_fraction:np.float64(0.75) - self_distillation/token_reweight_context_sample_fraction:np.float64(0.75) - self_distillation/token_reweight_correct_sample_fraction:np.float64(0.69140625) - self_distillation/token_reweight_w_mean:np.float64(15106.58354095975) - self_distillation/token_reweight_w_std:np.float64(301363.08133950847) - self_distillation/token_reweight_w_clipped_mean:np.float64(0.9999013759661466) - self_distillation/token_reweight_w_clip_frac:np.float64(0.01797304963838542) - self_distillation/empty_target_batch:np.float64(0.25) - actor/grad_norm:np.float64(0.1237591989338398) - perf/mfu/actor:np.float64(0.05327105722159453) - perf/max_memory_allocated_gb:np.float64(130.98332166671753) - perf/max_memory_reserved_gb:np.float64(140.86328125) - perf/cpu_memory_used_gb:np.float64(196.0658302307129) - actor/lr:np.float64(1e-06) - training/global_step:85 - training/epoch:3 - critic/score/mean:0.69140625 - critic/score/max:1.0 - critic/score/min:0.0 - critic/rewards/mean:0.69140625 - critic/rewards/max:1.0 - critic/rewards/min:0.0 - critic/advantages/mean:0.004565150942653418 - critic/advantages/max:0.625 - critic/advantages/min:-0.875 - critic/returns/mean:0.004565150942653418 - critic/returns/max:0.625 - critic/returns/min:-0.875 - response_length/mean:332.640625 - response_length/max:959.0 - response_length/min:
|
||
step:86 - global_seqlen/min:19969 - global_seqlen/max:26446 - global_seqlen/minmax_diff:6477 - global_seqlen/balanced_min:23249 - global_seqlen/balanced_max:23252 - global_seqlen/mean:23250.375 - actor/entropy:0.15242978930473328 - perf/mfu/actor_infer:0 - training/rollout_probs_diff_valid:1 - training/rollout_probs_diff_max:0.23105868697166443 - training/rollout_probs_diff_mean:0.002931097289547324 - training/rollout_probs_diff_std:0.010023131035268307 - training/rollout_actor_probs_pearson_corr:0.998683750629425 - self_distillation/success_group_fraction:0.875 - self_distillation/success_sample_fraction:0.875 - self_distillation/correct_sample_fraction:0.83984375 - self_distillation/feedback_available_fraction:0.0 - self_distillation/feedback_used_fraction:0.0 - self_distillation/reprompt_sample_fraction:0.875 - rollout_corr/rollout_is_mean:1.0000308752059937 - rollout_corr/rollout_is_ratio_fraction_high:1.125986636907328e-05 - rollout_corr/rollout_is_ratio_fraction_low:1.125986636907328e-05 - rollout_corr/rollout_is_max:2.0322604179382324 - rollout_corr/rollout_is_min:0.3792487382888794 - rollout_corr/rollout_is_std:0.029225580394268036 - rollout_corr/rollout_is_eff_sample_size:0.9991467133695389 - rollout_corr/rollout_is_seq_mean:0.9999984502792358 - rollout_corr/rollout_is_seq_std:0.0019490608246996999 - rollout_corr/rollout_is_seq_max:1.010423183441162 - rollout_corr/rollout_is_seq_min:0.9918707013130188 - rollout_corr/rollout_is_seq_max_deviation:0.01042318344116211 - rollout_corr/rollout_is_seq_fraction_high:0.0 - rollout_corr/rollout_is_seq_fraction_low:0.0 - rollout_corr/training_ppl:np.float64(1.1619863198138773) - rollout_corr/training_log_ppl:np.float64(0.14921248353493866) - rollout_corr/kl:np.float64(0.0005324320977670993) - rollout_corr/k3_kl:np.float64(0.0004945815423127442) - rollout_corr/rollout_ppl:np.float64(1.161366907414049) - rollout_corr/rollout_log_ppl:np.float64(0.1486800509155728) - rollout_corr/log_ppl_diff:np.float64(0.0005324326193658635) - rollout_corr/log_ppl_abs_diff:np.float64(0.0015857149701332673) - rollout_corr/log_ppl_diff_max:np.float64(0.008885562419891357) - rollout_corr/log_ppl_diff_min:np.float64(-0.00977480411529541) - rollout_corr/ppl_ratio:np.float64(1.0005347630940378) - rollout_corr/chi2_token:np.float64(0.0009847534820437431) - rollout_corr/chi2_seq:np.float64(0.564132138621062) - actor/pg_loss:np.float64(0.0006826445460319519) - actor/kl_loss:np.float64(0.0) - actor/pg_clipfrac:np.float64(3.8786817640357185e-05) - actor/ppo_kl:np.float64(8.414072883056178e-05) - actor/pg_clipfrac_lower:np.float64(0.0) - self_distillation/token_reweight_lambda:np.float64(0.5) - self_distillation/token_reweight_eps_w:np.float64(0.20000000000000004) - self_distillation/token_reweight_active_token_fraction:np.float64(0.875) - self_distillation/token_reweight_context_sample_fraction:np.float64(0.875) - self_distillation/token_reweight_correct_sample_fraction:np.float64(0.83984375) - self_distillation/token_reweight_w_mean:np.float64(4939.58044633409) - self_distillation/token_reweight_w_std:np.float64(109938.30147719354) - self_distillation/token_reweight_w_clipped_mean:np.float64(0.9997702944092453) - self_distillation/token_reweight_w_clip_frac:np.float64(0.019916853663744405) - self_distillation/empty_target_batch:np.float64(0.125) - actor/grad_norm:np.float64(0.09240363747812808) - perf/mfu/actor:np.float64(0.05295501420496774) - perf/max_memory_allocated_gb:np.float64(130.98332166671753) - perf/max_memory_reserved_gb:np.float64(140.86328125) - perf/cpu_memory_used_gb:np.float64(196.0387420654297) - actor/lr:np.float64(1e-06) - training/global_step:86 - training/epoch:3 - critic/score/mean:0.83984375 - critic/score/max:1.0 - critic/score/min:0.0 - critic/rewards/mean:0.83984375 - critic/rewards/max:1.0 - critic/rewards/min:0.0 - critic/advantages/mean:-0.0016115683829411864 - critic/advantages/max:0.625 - critic/advantages/min:-0.875 - critic/returns/mean:-0.0016115683829411864 - critic/returns/max:0.625 - critic/returns/min:-0.875 - response_length/mean:346.91796875 - response_length/max
|
||
step:87 - global_seqlen/min:21085 - global_seqlen/max:26247 - global_seqlen/minmax_diff:5162 - global_seqlen/balanced_min:24087 - global_seqlen/balanced_max:24090 - global_seqlen/mean:24088.375 - actor/entropy:0.1323355734348297 - perf/mfu/actor_infer:0 - training/rollout_probs_diff_valid:1 - training/rollout_probs_diff_max:0.17917847633361816 - training/rollout_probs_diff_mean:0.0025879396125674248 - training/rollout_probs_diff_std:0.009271741844713688 - training/rollout_actor_probs_pearson_corr:0.9986913204193115 - self_distillation/success_group_fraction:0.65625 - self_distillation/success_sample_fraction:0.65625 - self_distillation/correct_sample_fraction:0.609375 - self_distillation/feedback_available_fraction:0.0 - self_distillation/feedback_used_fraction:0.0 - self_distillation/reprompt_sample_fraction:0.65625 - rollout_corr/rollout_is_mean:0.9999597668647766 - rollout_corr/rollout_is_ratio_fraction_high:1.149068702943623e-05 - rollout_corr/rollout_is_ratio_fraction_low:0.0 - rollout_corr/rollout_is_max:2.032074213027954 - rollout_corr/rollout_is_min:0.5740495324134827 - rollout_corr/rollout_is_std:0.02641010284423828 - rollout_corr/rollout_is_eff_sample_size:0.9993027545187237 - rollout_corr/rollout_is_seq_mean:0.9999773502349854 - rollout_corr/rollout_is_seq_std:0.0016537748742848635 - rollout_corr/rollout_is_seq_max:1.0089139938354492 - rollout_corr/rollout_is_seq_min:0.9930402040481567 - rollout_corr/rollout_is_seq_max_deviation:0.008913993835449219 - rollout_corr/rollout_is_seq_fraction_high:0.0 - rollout_corr/rollout_is_seq_fraction_low:0.0 - rollout_corr/training_ppl:np.float64(1.1391075374558568) - rollout_corr/training_log_ppl:np.float64(0.12840614165907027) - rollout_corr/kl:np.float64(0.0004205593369874805) - rollout_corr/k3_kl:np.float64(0.0003839448123414968) - rollout_corr/rollout_ppl:np.float64(1.1386172864586115) - rollout_corr/rollout_log_ppl:np.float64(0.1279855821048841) - rollout_corr/log_ppl_diff:np.float64(0.00042055955418618396) - rollout_corr/log_ppl_abs_diff:np.float64(0.0011943888894165866) - rollout_corr/log_ppl_diff_max:np.float64(0.004606947302818298) - rollout_corr/log_ppl_diff_min:np.float64(-0.007427416741847992) - rollout_corr/ppl_ratio:np.float64(1.0004217766691) - rollout_corr/chi2_token:np.float64(0.0006975787691771984) - rollout_corr/chi2_seq:np.float64(0.2279121340252459) - actor/pg_loss:np.float64(0.0009529207600280643) - actor/kl_loss:np.float64(0.0) - actor/pg_clipfrac:np.float64(0.00015217112922982778) - actor/ppo_kl:np.float64(4.016830045949771e-05) - actor/pg_clipfrac_lower:np.float64(0.0) - self_distillation/token_reweight_lambda:np.float64(0.5) - self_distillation/token_reweight_eps_w:np.float64(0.20000000000000004) - self_distillation/token_reweight_active_token_fraction:np.float64(0.65625) - self_distillation/token_reweight_context_sample_fraction:np.float64(0.65625) - self_distillation/token_reweight_correct_sample_fraction:np.float64(0.609375) - self_distillation/token_reweight_w_mean:np.float64(204.87072951882146) - self_distillation/token_reweight_w_std:np.float64(4660.289743357804) - self_distillation/token_reweight_w_clipped_mean:np.float64(0.9996334998868406) - self_distillation/token_reweight_w_clip_frac:np.float64(0.020627167890779674) - self_distillation/empty_target_batch:np.float64(0.34375) - actor/grad_norm:np.float64(0.11101096495985985) - perf/mfu/actor:np.float64(0.054963622205453096) - perf/max_memory_allocated_gb:np.float64(130.98332166671753) - perf/max_memory_reserved_gb:np.float64(140.86328125) - perf/cpu_memory_used_gb:np.float64(195.97815322875977) - actor/lr:np.float64(1e-06) - training/global_step:87 - training/epoch:3 - critic/score/mean:0.609375 - critic/score/max:1.0 - critic/score/min:0.0 - critic/rewards/mean:0.609375 - critic/rewards/max:1.0 - critic/rewards/min:0.0 - critic/advantages/mean:-0.0018227101536467671 - critic/advantages/max:0.625 - critic/advantages/min:-0.875 - critic/returns/mean:-0.0018227101536467671 - critic/returns/max:0.625 - critic/returns/min:-0.875 - response_length/mean:339.94921875 - response_length/max:602.0 -
|
||
step:88 - global_seqlen/min:18837 - global_seqlen/max:25671 - global_seqlen/minmax_diff:6834 - global_seqlen/balanced_min:23089 - global_seqlen/balanced_max:23094 - global_seqlen/mean:23092.0 - actor/entropy:0.14068984985351562 - perf/mfu/actor_infer:0 - training/rollout_probs_diff_valid:1 - training/rollout_probs_diff_max:0.19462502002716064 - training/rollout_probs_diff_mean:0.002681329380720854 - training/rollout_probs_diff_std:0.009461375884711742 - training/rollout_actor_probs_pearson_corr:0.9987494349479675 - self_distillation/success_group_fraction:0.71875 - self_distillation/success_sample_fraction:0.71484375 - self_distillation/correct_sample_fraction:0.671875 - self_distillation/feedback_available_fraction:0.0 - self_distillation/feedback_used_fraction:0.0 - self_distillation/reprompt_sample_fraction:0.71484375 - rollout_corr/rollout_is_mean:0.9999996423721313 - rollout_corr/rollout_is_ratio_fraction_high:0.0 - rollout_corr/rollout_is_ratio_fraction_low:2.3253651306731626e-05 - rollout_corr/rollout_is_max:1.644294023513794 - rollout_corr/rollout_is_min:0.4757203161716461 - rollout_corr/rollout_is_std:0.027249855920672417 - rollout_corr/rollout_is_eff_sample_size:0.9992577582488479 - rollout_corr/rollout_is_seq_mean:0.9998669624328613 - rollout_corr/rollout_is_seq_std:0.001749269780702889 - rollout_corr/rollout_is_seq_max:1.005174160003662 - rollout_corr/rollout_is_seq_min:0.9924665093421936 - rollout_corr/rollout_is_seq_max_deviation:0.0075334906578063965 - rollout_corr/rollout_is_seq_fraction_high:0.0 - rollout_corr/rollout_is_seq_fraction_low:0.0 - rollout_corr/training_ppl:np.float64(1.1600256217643619) - rollout_corr/training_log_ppl:np.float64(0.14285353356899577) - rollout_corr/kl:np.float64(0.0006477979862360428) - rollout_corr/k3_kl:np.float64(0.00043804328623764377) - rollout_corr/rollout_ppl:np.float64(1.1592506598681211) - rollout_corr/rollout_log_ppl:np.float64(0.1422057347954251) - rollout_corr/log_ppl_diff:np.float64(0.0006477987735706847) - rollout_corr/log_ppl_abs_diff:np.float64(0.001455325818824349) - rollout_corr/log_ppl_diff_max:np.float64(0.008176639676094055) - rollout_corr/log_ppl_diff_min:np.float64(-0.0035071074962615967) - rollout_corr/ppl_ratio:np.float64(1.0006498480215669) - rollout_corr/chi2_token:np.float64(0.0004831866826862097) - rollout_corr/chi2_seq:np.float64(0.22203046013601124) - actor/pg_loss:np.float64(0.0006858368869870901) - actor/kl_loss:np.float64(0.0) - actor/pg_clipfrac:np.float64(0.00016782948659965768) - actor/ppo_kl:np.float64(0.0001333341535927346) - actor/pg_clipfrac_lower:np.float64(0.0) - self_distillation/token_reweight_lambda:np.float64(0.5) - self_distillation/token_reweight_eps_w:np.float64(0.20000000000000004) - self_distillation/token_reweight_active_token_fraction:np.float64(0.71484375) - self_distillation/token_reweight_context_sample_fraction:np.float64(0.71484375) - self_distillation/token_reweight_correct_sample_fraction:np.float64(0.671875) - self_distillation/token_reweight_w_mean:np.float64(570.608905545203) - self_distillation/token_reweight_w_std:np.float64(12229.07915653108) - self_distillation/token_reweight_w_clipped_mean:np.float64(1.0001349451486021) - self_distillation/token_reweight_w_clip_frac:np.float64(0.01699924370041117) - self_distillation/empty_target_batch:np.float64(0.28515625) - actor/grad_norm:np.float64(0.1187331285327673) - perf/mfu/actor:np.float64(0.05280560311060975) - perf/max_memory_allocated_gb:np.float64(130.98332166671753) - perf/max_memory_reserved_gb:np.float64(140.86328125) - perf/cpu_memory_used_gb:np.float64(196.03455352783203) - actor/lr:np.float64(1e-06) - training/global_step:88 - training/epoch:3 - critic/score/mean:0.671875 - critic/score/max:1.0 - critic/score/min:0.0 - critic/rewards/mean:0.671875 - critic/rewards/max:1.0 - critic/rewards/min:0.0 - critic/advantages/mean:0.0006365687004290521 - critic/advantages/max:0.875 - critic/advantages/min:-0.875 - critic/returns/mean:0.0006365687004290521 - critic/returns/max:0.875 - critic/returns/min:-0.875 - response_length/mean:335.96875 - response_length/max:
|
||
step:89 - global_seqlen/min:22363 - global_seqlen/max:26624 - global_seqlen/minmax_diff:4261 - global_seqlen/balanced_min:24749 - global_seqlen/balanced_max:24753 - global_seqlen/mean:24751.0 - actor/entropy:0.1324988603591919 - perf/mfu/actor_infer:0 - training/rollout_probs_diff_valid:1 - training/rollout_probs_diff_max:0.27730056643486023 - training/rollout_probs_diff_mean:0.0024554075207561255 - training/rollout_probs_diff_std:0.00897773914039135 - training/rollout_actor_probs_pearson_corr:0.9988168478012085 - self_distillation/success_group_fraction:0.78125 - self_distillation/success_sample_fraction:0.78125 - self_distillation/correct_sample_fraction:0.71484375 - self_distillation/feedback_available_fraction:0.0 - self_distillation/feedback_used_fraction:0.0 - self_distillation/reprompt_sample_fraction:0.78125 - rollout_corr/rollout_is_mean:1.000014305114746 - rollout_corr/rollout_is_ratio_fraction_high:1.1008366527676117e-05 - rollout_corr/rollout_is_ratio_fraction_low:2.2016733055352233e-05 - rollout_corr/rollout_is_max:2.2451746463775635 - rollout_corr/rollout_is_min:0.3002515733242035 - rollout_corr/rollout_is_std:0.02629021555185318 - rollout_corr/rollout_is_eff_sample_size:0.9993093019327395 - rollout_corr/rollout_is_seq_mean:1.0000624656677246 - rollout_corr/rollout_is_seq_std:0.0015305746346712112 - rollout_corr/rollout_is_seq_max:1.0051894187927246 - rollout_corr/rollout_is_seq_min:0.9954579472541809 - rollout_corr/rollout_is_seq_max_deviation:0.005189418792724609 - rollout_corr/rollout_is_seq_fraction_high:0.0 - rollout_corr/rollout_is_seq_fraction_low:0.0 - rollout_corr/training_ppl:np.float64(1.1442256579175591) - rollout_corr/training_log_ppl:np.float64(0.1322182439325843) - rollout_corr/kl:np.float64(0.00022351242467166799) - rollout_corr/k3_kl:np.float64(0.00042760402516250906) - rollout_corr/rollout_ppl:np.float64(1.1439585289917886) - rollout_corr/rollout_log_ppl:np.float64(0.1319947297743056) - rollout_corr/log_ppl_diff:np.float64(0.0002235141582787037) - rollout_corr/log_ppl_abs_diff:np.float64(0.0011873634211951867) - rollout_corr/log_ppl_diff_max:np.float64(0.005702301859855652) - rollout_corr/log_ppl_diff_min:np.float64(-0.004311807453632355) - rollout_corr/ppl_ratio:np.float64(1.0002248019445688) - rollout_corr/chi2_token:np.float64(0.0013164845295250416) - rollout_corr/chi2_seq:np.float64(0.43028471851721406) - actor/pg_loss:np.float64(0.0009632922010496259) - actor/kl_loss:np.float64(0.0) - actor/pg_clipfrac:np.float64(0.0002576569099801418) - actor/ppo_kl:np.float64(-7.108147247625318e-05) - actor/pg_clipfrac_lower:np.float64(0.0) - self_distillation/token_reweight_lambda:np.float64(0.5) - self_distillation/token_reweight_eps_w:np.float64(0.20000000000000004) - self_distillation/token_reweight_active_token_fraction:np.float64(0.78125) - self_distillation/token_reweight_context_sample_fraction:np.float64(0.78125) - self_distillation/token_reweight_correct_sample_fraction:np.float64(0.71484375) - self_distillation/token_reweight_w_mean:np.float64(36720.438478566706) - self_distillation/token_reweight_w_std:np.float64(389112.48932122736) - self_distillation/token_reweight_w_clipped_mean:np.float64(0.9997167198453099) - self_distillation/token_reweight_w_clip_frac:np.float64(0.026357669281424023) - self_distillation/empty_target_batch:np.float64(0.21875) - actor/grad_norm:np.float64(0.17487578466534615) - perf/mfu/actor:np.float64(0.05634677546245617) - perf/max_memory_allocated_gb:np.float64(130.98332166671753) - perf/max_memory_reserved_gb:np.float64(140.86328125) - perf/cpu_memory_used_gb:np.float64(196.1344871520996) - actor/lr:np.float64(1e-06) - training/global_step:89 - training/epoch:3 - critic/score/mean:0.71484375 - critic/score/max:1.0 - critic/score/min:0.0 - critic/rewards/mean:0.71484375 - critic/rewards/max:1.0 - critic/rewards/min:0.0 - critic/advantages/mean:0.002399823861196637 - critic/advantages/max:0.625 - critic/advantages/min:-0.875 - critic/returns/mean:0.002399823861196637 - critic/returns/max:0.625 - critic/returns/min:-0.875 - response_length/mean:354.84375 - respons
|
||
test_gen_batch meta info: {'eos_token_id': 151645, 'pad_token_id': 151643, 'recompute_log_prob': False, 'do_sample': True, 'validate': True, 'global_steps': 90}
|
||
validation generation end
|
||
[prompt] system
|
||
|
||
Given a question and four options, please select the right answer. Respond in the following format:
|
||
<reasoning>
|
||
...
|
||
</reasoning>
|
||
<answer>
|
||
...
|
||
</answer>
|
||
|
||
For the answer, only output the letter corresponding to the correct option (A, B, C, or D), and nothing else. Do not restate the answer text. For example, if the answer is "A", just output:
|
||
<answer>
|
||
A
|
||
</answer>
|
||
|
||
user
|
||
Given the following crystal structure parameters for the material RbLa9(IrO6)4 (Material ID: mp-560657), calculate the volume of the unit cell (in ų).
|
||
|
||
Lattice Parameters:
|
||
|
||
a = 7.82 Å
|
||
b = 7.82 Å
|
||
c = 17.88 Å
|
||
α = 90.00º
|
||
β = 90.00º
|
||
γ = 90.00º
|
||
|
||
|
||
|
||
A: 1025.67 ų
|
||
|
||
B: 1094.31 ų
|
||
|
||
C: 1200.45 ų
|
||
|
||
D: 1150.78 ų
|
||
|
||
Please reason step by step.
|
||
assistant
|
||
<think>
|
||
|
||
</think>
|
||
|
||
|
||
[response] <reasoning>
|
||
The given crystal structure has orthorhombic symmetry, as all angles (α, β, γ) are 90.00º. For an orthorhombic system, the volume of the unit cell is calculated using the formula:
|
||
|
||
$$ V = a \times b \times c $$
|
||
|
||
Substituting the given lattice parameters:
|
||
|
||
$$ V = 7.82 \, \text{Å} \times 7.82 \, \text{Å} \times 17.88 \, \text{Å} $$
|
||
|
||
First, calculate $ 7.82 \times 7.82 $:
|
||
|
||
$$ 7.82 \times 7.82 = 61.1524 $$
|
||
|
||
Then multiply by 17.88:
|
||
|
||
$$ 61.1524 \times 17.88 = 1094.31 $$
|
||
|
||
Thus, the volume of the unit cell is 1094.31 ų.
|
||
</reasoning>
|
||
<answer>
|
||
B
|
||
</answer>
|
||
[ground_truth] B
|
||
[score] 1.0
|
||
[acc] 1.0
|
||
[pred] B
|
||
[incorrect_format] 1
|
||
[feedback]
|
||
local_global_step_folder: /mnt/mole/SDPO/L2T/checkpoints/datasets/sciknoweval/material/qwen3gen-material-RLSD_TR-Qwen-Qwen3-8B-mbs8-decay0-tr0.1-train32-rollout8-lr1e-6-vllm0.8/global_step_90
|
||
step:90 - global_seqlen/min:19604 - global_seqlen/max:25499 - global_seqlen/minmax_diff:5895 - global_seqlen/balanced_min:23067 - global_seqlen/balanced_max:23073 - global_seqlen/mean:23071.25 - actor/entropy:0.14256960153579712 - perf/mfu/actor_infer:0 - training/rollout_probs_diff_valid:1 - training/rollout_probs_diff_max:0.35351723432540894 - training/rollout_probs_diff_mean:0.002670942572876811 - training/rollout_probs_diff_std:0.009400460869073868 - training/rollout_actor_probs_pearson_corr:0.9987446665763855 - self_distillation/success_group_fraction:0.8125 - self_distillation/success_sample_fraction:0.8125 - self_distillation/correct_sample_fraction:0.80859375 - self_distillation/feedback_available_fraction:0.0 - self_distillation/feedback_used_fraction:0.0 - self_distillation/reprompt_sample_fraction:0.8125 - rollout_corr/rollout_is_mean:0.9999788403511047 - rollout_corr/rollout_is_ratio_fraction_high:0.0 - rollout_corr/rollout_is_ratio_fraction_low:0.0 - rollout_corr/rollout_is_max:1.5973143577575684 - rollout_corr/rollout_is_min:0.5164310336112976 - rollout_corr/rollout_is_std:0.026839928701519966 - rollout_corr/rollout_is_eff_sample_size:0.99928001780181 - rollout_corr/rollout_is_seq_mean:1.0000172853469849 - rollout_corr/rollout_is_seq_std:0.001670500379987061 - rollout_corr/rollout_is_seq_max:1.0047962665557861 - rollout_corr/rollout_is_seq_min:0.9956794381141663 - rollout_corr/rollout_is_seq_max_deviation:0.004796266555786133 - rollout_corr/rollout_is_seq_fraction_high:0.0 - rollout_corr/rollout_is_seq_fraction_low:0.0 - rollout_corr/training_ppl:np.float64(1.1533320206217468) - rollout_corr/training_log_ppl:np.float64(0.14133044816117035) - rollout_corr/kl:np.float64(0.0003802113450959155) - rollout_corr/k3_kl:np.float64(0.0005155056666126256) - rollout_corr/rollout_ppl:np.float64(1.152886132709682) - rollout_corr/rollout_log_ppl:np.float64(0.14095023600384593) - rollout_corr/log_ppl_diff:np.float64(0.00038021215732442215) - rollout_corr/log_ppl_abs_diff:np.float64(0.0015572865449939854) - rollout_corr/log_ppl_diff_max:np.float64(0.008414089679718018) - rollout_corr/log_ppl_diff_min:np.float64(-0.0054509639739990234) - rollout_corr/ppl_ratio:np.float64(1.000382244354114) - rollout_corr/chi2_token:np.float64(0.0013116379268467426) - rollout_corr/chi2_seq:np.float64(0.4403390462975949) - actor/pg_loss:np.float64(7.652048952877522e-05) - actor/kl_loss:np.float64(0.0) - actor/pg_clipfrac:np.float64(1.100352073990507e-05) - actor/ppo_kl:np.float64(2.3073779324533916e-05) - actor/pg_clipfrac_lower:np.float64(0.0) - self_distillation/token_reweight_lambda:np.float64(0.5) - self_distillation/token_reweight_eps_w:np.float64(0.20000000000000004) - self_distillation/token_reweight_active_token_fraction:np.float64(0.8125) - self_distillation/token_reweight_context_sample_fraction:np.float64(0.8125) - self_distillation/token_reweight_correct_sample_fraction:np.float64(0.80859375) - self_distillation/token_reweight_w_mean:np.float64(1.160266701830551) - self_distillation/token_reweight_w_std:np.float64(3.191360288124997) - self_distillation/token_reweight_w_clipped_mean:np.float64(0.9997768390458077) - self_distillation/token_reweight_w_clip_frac:np.float64(0.004193740169284865) - self_distillation/empty_target_batch:np.float64(0.1875) - actor/grad_norm:np.float64(0.028098714305087924) - perf/mfu/actor:np.float64(0.05342099023043444) - perf/max_memory_allocated_gb:np.float64(130.98332166671753) - perf/max_memory_reserved_gb:np.float64(140.86328125) - perf/cpu_memory_used_gb:np.float64(196.17328262329102) - actor/lr:np.float64(1e-06) - val-aux/sciknoweval/reward/mean@16:np.float64(0.7539893617021277) - val-aux/sciknoweval/reward/std@16:np.float64(0.047999869452259585) - val-aux/sciknoweval/reward/best@2/mean:np.float64(0.7715425531914895) - val-aux/sciknoweval/reward/best@2/std:np.float64(0.031585664717190674) - val-aux/sciknoweval/reward/worst@2/mean:np.float64(0.7356595744680852) - val-aux/sciknoweval/reward/worst@2/std:np.float64(0.04820199399455401) - val-aux/sciknoweval/reward/maj@2/mean:np.float64(0.753776595744680
|
||
step:91 - global_seqlen/min:20497 - global_seqlen/max:25268 - global_seqlen/minmax_diff:4771 - global_seqlen/balanced_min:22088 - global_seqlen/balanced_max:22093 - global_seqlen/mean:22091.25 - actor/entropy:0.14610257744789124 - perf/mfu/actor_infer:0 - training/rollout_probs_diff_valid:1 - training/rollout_probs_diff_max:0.5466005206108093 - training/rollout_probs_diff_mean:0.002629169262945652 - training/rollout_probs_diff_std:0.009485256858170033 - training/rollout_actor_probs_pearson_corr:0.9987938404083252 - self_distillation/success_group_fraction:0.90625 - self_distillation/success_sample_fraction:0.90625 - self_distillation/correct_sample_fraction:0.84765625 - self_distillation/feedback_available_fraction:0.0 - self_distillation/feedback_used_fraction:0.0 - self_distillation/reprompt_sample_fraction:0.90625 - rollout_corr/rollout_is_mean:1.0000755786895752 - rollout_corr/rollout_is_ratio_fraction_high:0.0 - rollout_corr/rollout_is_ratio_fraction_low:3.5055738408118486e-05 - rollout_corr/rollout_is_max:1.9108070135116577 - rollout_corr/rollout_is_min:0.3648514449596405 - rollout_corr/rollout_is_std:0.027369895949959755 - rollout_corr/rollout_is_eff_sample_size:0.9992515686028562 - rollout_corr/rollout_is_seq_mean:1.0000547170639038 - rollout_corr/rollout_is_seq_std:0.001723990892060101 - rollout_corr/rollout_is_seq_max:1.0056302547454834 - rollout_corr/rollout_is_seq_min:0.9936156272888184 - rollout_corr/rollout_is_seq_max_deviation:0.006384372711181641 - rollout_corr/rollout_is_seq_fraction_high:0.0 - rollout_corr/rollout_is_seq_fraction_low:0.0 - rollout_corr/training_ppl:np.float64(1.1725259199738503) - rollout_corr/training_log_ppl:np.float64(0.15398886370530818) - rollout_corr/kl:np.float64(0.0005335942853097464) - rollout_corr/k3_kl:np.float64(0.00044532014708309475) - rollout_corr/rollout_ppl:np.float64(1.1718511590734124) - rollout_corr/rollout_log_ppl:np.float64(0.15345526859891834) - rollout_corr/log_ppl_diff:np.float64(0.0005335951063898392) - rollout_corr/log_ppl_abs_diff:np.float64(0.0014636352643719874) - rollout_corr/log_ppl_diff_max:np.float64(0.010127514600753784) - rollout_corr/log_ppl_diff_min:np.float64(-0.004548102617263794) - rollout_corr/ppl_ratio:np.float64(1.0005357707850635) - rollout_corr/chi2_token:np.float64(0.0006994069553911686) - rollout_corr/chi2_seq:np.float64(0.23433439433574677) - actor/pg_loss:np.float64(0.0007798228180035949) - actor/kl_loss:np.float64(0.0) - actor/pg_clipfrac:np.float64(0.00021067558282084065) - actor/ppo_kl:np.float64(0.00019650942880389266) - actor/pg_clipfrac_lower:np.float64(0.0) - self_distillation/token_reweight_lambda:np.float64(0.5) - self_distillation/token_reweight_eps_w:np.float64(0.20000000000000004) - self_distillation/token_reweight_active_token_fraction:np.float64(0.90625) - self_distillation/token_reweight_context_sample_fraction:np.float64(0.90625) - self_distillation/token_reweight_correct_sample_fraction:np.float64(0.84765625) - self_distillation/token_reweight_w_mean:np.float64(25743.910056033637) - self_distillation/token_reweight_w_std:np.float64(337349.557794077) - self_distillation/token_reweight_w_clipped_mean:np.float64(0.9999736596364528) - self_distillation/token_reweight_w_clip_frac:np.float64(0.021167089085793123) - self_distillation/empty_target_batch:np.float64(0.09375) - actor/grad_norm:np.float64(0.1266098888590932) - perf/mfu/actor:np.float64(0.05037939348765556) - perf/max_memory_allocated_gb:np.float64(130.98332166671753) - perf/max_memory_reserved_gb:np.float64(140.86328125) - perf/cpu_memory_used_gb:np.float64(196.7395133972168) - actor/lr:np.float64(1e-06) - training/global_step:91 - training/epoch:3 - critic/score/mean:0.84765625 - critic/score/max:1.0 - critic/score/min:0.0 - critic/rewards/mean:0.84765625 - critic/rewards/max:1.0 - critic/rewards/min:0.0 - critic/advantages/mean:-0.0007186426664702594 - critic/advantages/max:0.75 - critic/advantages/min:-0.875 - critic/returns/mean:-0.0007186426664702594 - critic/returns/max:0.75 - critic/returns/min:-0.875 - response_length/mean:334.2890625 - response_length/max:7
|
||
step:92 - global_seqlen/min:21156 - global_seqlen/max:25601 - global_seqlen/minmax_diff:4445 - global_seqlen/balanced_min:23914 - global_seqlen/balanced_max:23917 - global_seqlen/mean:23915.875 - actor/entropy:0.15240398049354553 - perf/mfu/actor_infer:0 - training/rollout_probs_diff_valid:1 - training/rollout_probs_diff_max:0.4429726004600525 - training/rollout_probs_diff_mean:0.0027351249009370804 - training/rollout_probs_diff_std:0.00975241232663393 - training/rollout_actor_probs_pearson_corr:0.9988147020339966 - self_distillation/success_group_fraction:0.78125 - self_distillation/success_sample_fraction:0.77734375 - self_distillation/correct_sample_fraction:0.71875 - self_distillation/feedback_available_fraction:0.0 - self_distillation/feedback_used_fraction:0.0 - self_distillation/reprompt_sample_fraction:0.77734375 - rollout_corr/rollout_is_mean:1.0001146793365479 - rollout_corr/rollout_is_ratio_fraction_high:0.0 - rollout_corr/rollout_is_ratio_fraction_low:3.320383984828368e-05 - rollout_corr/rollout_is_max:1.9662312269210815 - rollout_corr/rollout_is_min:0.47732794284820557 - rollout_corr/rollout_is_std:0.0285780131816864 - rollout_corr/rollout_is_eff_sample_size:0.9991842016461998 - rollout_corr/rollout_is_seq_mean:1.0001089572906494 - rollout_corr/rollout_is_seq_std:0.001473374548368156 - rollout_corr/rollout_is_seq_max:1.0044150352478027 - rollout_corr/rollout_is_seq_min:0.9953008890151978 - rollout_corr/rollout_is_seq_max_deviation:0.004699110984802246 - rollout_corr/rollout_is_seq_fraction_high:0.0 - rollout_corr/rollout_is_seq_fraction_low:0.0 - rollout_corr/training_ppl:np.float64(1.159420844167471) - rollout_corr/training_log_ppl:np.float64(0.14538506651297212) - rollout_corr/kl:np.float64(0.0005853978563141737) - rollout_corr/k3_kl:np.float64(0.0004851882544585351) - rollout_corr/rollout_ppl:np.float64(1.1587399523705244) - rollout_corr/rollout_log_ppl:np.float64(0.144799667839834) - rollout_corr/log_ppl_diff:np.float64(0.0005853986731381156) - rollout_corr/log_ppl_abs_diff:np.float64(0.0014114797013462521) - rollout_corr/log_ppl_diff_max:np.float64(0.006640970706939697) - rollout_corr/log_ppl_diff_min:np.float64(-0.004197612404823303) - rollout_corr/ppl_ratio:np.float64(1.0005870771128684) - rollout_corr/chi2_token:np.float64(0.0007457071915268898) - rollout_corr/chi2_seq:np.float64(0.1930299310479313) - actor/pg_loss:np.float64(0.001052756910212338) - actor/kl_loss:np.float64(0.0) - actor/pg_clipfrac:np.float64(0.0001924268062793999) - actor/ppo_kl:np.float64(0.00029038446403717444) - actor/pg_clipfrac_lower:np.float64(0.0) - self_distillation/token_reweight_lambda:np.float64(0.5) - self_distillation/token_reweight_eps_w:np.float64(0.20000000000000004) - self_distillation/token_reweight_active_token_fraction:np.float64(0.77734375) - self_distillation/token_reweight_context_sample_fraction:np.float64(0.77734375) - self_distillation/token_reweight_correct_sample_fraction:np.float64(0.71875) - self_distillation/token_reweight_w_mean:np.float64(22084.149476626422) - self_distillation/token_reweight_w_std:np.float64(376728.0488451377) - self_distillation/token_reweight_w_clipped_mean:np.float64(0.999960406916216) - self_distillation/token_reweight_w_clip_frac:np.float64(0.02320667250023689) - self_distillation/empty_target_batch:np.float64(0.22265625) - actor/grad_norm:np.float64(0.1322405505925417) - perf/mfu/actor:np.float64(0.05412615204124727) - perf/max_memory_allocated_gb:np.float64(130.98332166671753) - perf/max_memory_reserved_gb:np.float64(140.86328125) - perf/cpu_memory_used_gb:np.float64(196.32924270629883) - actor/lr:np.float64(1e-06) - training/global_step:92 - training/epoch:3 - critic/score/mean:0.71875 - critic/score/max:1.0 - critic/score/min:0.0 - critic/rewards/mean:0.71875 - critic/rewards/max:1.0 - critic/rewards/min:0.0 - critic/advantages/mean:-0.0007000475889071822 - critic/advantages/max:0.875 - critic/advantages/min:-0.875 - critic/returns/mean:-0.0007000475889071822 - critic/returns/max:0.875 - critic/returns/min:-0.875 - response_length/mean:352.93359375 - response_length/max:609
|
||
step:93 - global_seqlen/min:17767 - global_seqlen/max:25876 - global_seqlen/minmax_diff:8109 - global_seqlen/balanced_min:21939 - global_seqlen/balanced_max:21946 - global_seqlen/mean:21943.25 - actor/entropy:0.12090399861335754 - perf/mfu/actor_infer:0 - training/rollout_probs_diff_valid:1 - training/rollout_probs_diff_max:0.15149319171905518 - training/rollout_probs_diff_mean:0.002356545068323612 - training/rollout_probs_diff_std:0.009056123904883862 - training/rollout_actor_probs_pearson_corr:0.998685896396637 - self_distillation/success_group_fraction:0.84375 - self_distillation/success_sample_fraction:0.83984375 - self_distillation/correct_sample_fraction:0.8046875 - self_distillation/feedback_available_fraction:0.0 - self_distillation/feedback_used_fraction:0.0 - self_distillation/reprompt_sample_fraction:0.83984375 - rollout_corr/rollout_is_mean:0.9999880194664001 - rollout_corr/rollout_is_ratio_fraction_high:1.2775633877026848e-05 - rollout_corr/rollout_is_ratio_fraction_low:0.0 - rollout_corr/rollout_is_max:2.128607749938965 - rollout_corr/rollout_is_min:0.5685504674911499 - rollout_corr/rollout_is_std:0.025369489565491676 - rollout_corr/rollout_is_eff_sample_size:0.9993565649024203 - rollout_corr/rollout_is_seq_mean:1.0000287294387817 - rollout_corr/rollout_is_seq_std:0.0017659437144175172 - rollout_corr/rollout_is_seq_max:1.0067225694656372 - rollout_corr/rollout_is_seq_min:0.9932870864868164 - rollout_corr/rollout_is_seq_max_deviation:0.006722569465637207 - rollout_corr/rollout_is_seq_fraction_high:0.0 - rollout_corr/rollout_is_seq_fraction_low:0.0 - rollout_corr/training_ppl:np.float64(1.1301220706664026) - rollout_corr/training_log_ppl:np.float64(0.11933962863258785) - rollout_corr/kl:np.float64(0.0003626701826924261) - rollout_corr/k3_kl:np.float64(0.00046709808577105605) - rollout_corr/rollout_ppl:np.float64(1.129688723012805) - rollout_corr/rollout_log_ppl:np.float64(0.11897695804509567) - rollout_corr/log_ppl_diff:np.float64(0.0003626705874921754) - rollout_corr/log_ppl_abs_diff:np.float64(0.0015627569082425907) - rollout_corr/log_ppl_diff_max:np.float64(0.008959487080574036) - rollout_corr/log_ppl_diff_min:np.float64(-0.007843941450119019) - rollout_corr/ppl_ratio:np.float64(1.0003649715799838) - rollout_corr/chi2_token:np.float64(0.0011762918438762426) - rollout_corr/chi2_seq:np.float64(0.34349790937267244) - actor/pg_loss:np.float64(0.0003992551937699318) - actor/kl_loss:np.float64(0.0) - actor/pg_clipfrac:np.float64(6.536668161061243e-05) - actor/ppo_kl:np.float64(3.433879705161402e-05) - actor/pg_clipfrac_lower:np.float64(0.0) - self_distillation/token_reweight_lambda:np.float64(0.5) - self_distillation/token_reweight_eps_w:np.float64(0.20000000000000004) - self_distillation/token_reweight_active_token_fraction:np.float64(0.83984375) - self_distillation/token_reweight_context_sample_fraction:np.float64(0.83984375) - self_distillation/token_reweight_correct_sample_fraction:np.float64(0.8046875) - self_distillation/token_reweight_w_mean:np.float64(1485.541782564018) - self_distillation/token_reweight_w_std:np.float64(32076.792206338432) - self_distillation/token_reweight_w_clipped_mean:np.float64(1.0001281290315092) - self_distillation/token_reweight_w_clip_frac:np.float64(0.01311015960527584) - self_distillation/empty_target_batch:np.float64(0.16015625) - actor/grad_norm:np.float64(0.08833481976762414) - perf/mfu/actor:np.float64(0.049890876456477364) - perf/max_memory_allocated_gb:np.float64(130.98332166671753) - perf/max_memory_reserved_gb:np.float64(140.86328125) - perf/cpu_memory_used_gb:np.float64(196.14807891845703) - actor/lr:np.float64(1e-06) - training/global_step:93 - training/epoch:3 - critic/score/mean:0.8046875 - critic/score/max:1.0 - critic/score/min:0.0 - critic/rewards/mean:0.8046875 - critic/rewards/max:1.0 - critic/rewards/min:0.0 - critic/advantages/mean:-0.004495426546782255 - critic/advantages/max:0.875 - critic/advantages/min:-0.875 - critic/returns/mean:-0.004495426546782255 - critic/returns/max:0.875 - critic/returns/min:-0.875 - response_length/mean:305.7578125 - response_l
|
||
step:94 - global_seqlen/min:17725 - global_seqlen/max:26194 - global_seqlen/minmax_diff:8469 - global_seqlen/balanced_min:23294 - global_seqlen/balanced_max:23300 - global_seqlen/mean:23296.375 - actor/entropy:0.12870950996875763 - perf/mfu/actor_infer:0 - training/rollout_probs_diff_valid:1 - training/rollout_probs_diff_max:0.1791786253452301 - training/rollout_probs_diff_mean:0.0024543788749724627 - training/rollout_probs_diff_std:0.009046412073075771 - training/rollout_actor_probs_pearson_corr:0.9987512230873108 - self_distillation/success_group_fraction:0.75 - self_distillation/success_sample_fraction:0.74609375 - self_distillation/correct_sample_fraction:0.69921875 - self_distillation/feedback_available_fraction:0.0 - self_distillation/feedback_used_fraction:0.0 - self_distillation/reprompt_sample_fraction:0.74609375 - rollout_corr/rollout_is_mean:1.0000104904174805 - rollout_corr/rollout_is_ratio_fraction_high:1.1564035958144814e-05 - rollout_corr/rollout_is_ratio_fraction_low:0.0 - rollout_corr/rollout_is_max:2.084258556365967 - rollout_corr/rollout_is_min:0.5468906760215759 - rollout_corr/rollout_is_std:0.026822155341506004 - rollout_corr/rollout_is_eff_sample_size:0.9992809701042712 - rollout_corr/rollout_is_seq_mean:0.9999830722808838 - rollout_corr/rollout_is_seq_std:0.001727128168568015 - rollout_corr/rollout_is_seq_max:1.0052071809768677 - rollout_corr/rollout_is_seq_min:0.9946647882461548 - rollout_corr/rollout_is_seq_max_deviation:0.005335211753845215 - rollout_corr/rollout_is_seq_fraction_high:0.0 - rollout_corr/rollout_is_seq_fraction_low:0.0 - rollout_corr/training_ppl:np.float64(1.1428441368043423) - rollout_corr/training_log_ppl:np.float64(0.13037367451761384) - rollout_corr/kl:np.float64(0.0003746576256418166) - rollout_corr/k3_kl:np.float64(0.00045740963861362616) - rollout_corr/rollout_ppl:np.float64(1.1423959978856146) - rollout_corr/rollout_log_ppl:np.float64(0.12999901587318163) - rollout_corr/log_ppl_diff:np.float64(0.000374658644432202) - rollout_corr/log_ppl_abs_diff:np.float64(0.0015223244263324887) - rollout_corr/log_ppl_diff_max:np.float64(0.0070696622133255005) - rollout_corr/log_ppl_diff_min:np.float64(-0.006661802530288696) - rollout_corr/ppl_ratio:np.float64(1.0003767299931496) - rollout_corr/chi2_token:np.float64(0.0011147765908390284) - rollout_corr/chi2_seq:np.float64(0.7085019601508975) - actor/pg_loss:np.float64(0.000595554243773222) - actor/kl_loss:np.float64(0.0) - actor/pg_clipfrac:np.float64(0.00016299049775625463) - actor/ppo_kl:np.float64(-1.9219931974845395e-05) - actor/pg_clipfrac_lower:np.float64(0.0) - self_distillation/token_reweight_lambda:np.float64(0.5) - self_distillation/token_reweight_eps_w:np.float64(0.20000000000000004) - self_distillation/token_reweight_active_token_fraction:np.float64(0.74609375) - self_distillation/token_reweight_context_sample_fraction:np.float64(0.74609375) - self_distillation/token_reweight_correct_sample_fraction:np.float64(0.69921875) - self_distillation/token_reweight_w_mean:np.float64(5829.977993725799) - self_distillation/token_reweight_w_std:np.float64(131522.8458643295) - self_distillation/token_reweight_w_clipped_mean:np.float64(0.9996835256461054) - self_distillation/token_reweight_w_clip_frac:np.float64(0.01764957682462409) - self_distillation/empty_target_batch:np.float64(0.25390625) - actor/grad_norm:np.float64(0.09286441234871745) - perf/mfu/actor:np.float64(0.05320265824096174) - perf/max_memory_allocated_gb:np.float64(130.98332166671753) - perf/max_memory_reserved_gb:np.float64(140.86328125) - perf/cpu_memory_used_gb:np.float64(196.21632719039917) - actor/lr:np.float64(1e-06) - training/global_step:94 - training/epoch:3 - critic/score/mean:0.69921875 - critic/score/max:1.0 - critic/score/min:0.0 - critic/rewards/mean:0.69921875 - critic/rewards/max:1.0 - critic/rewards/min:0.0 - critic/advantages/mean:0.002400982892140746 - critic/advantages/max:0.875 - critic/advantages/min:-0.875 - critic/returns/mean:0.002400982892140746 - critic/returns/max:0.875 - critic/returns/min:-0.875 - response_length/mean:337.79296875 - response_
|
||
step:95 - global_seqlen/min:20448 - global_seqlen/max:27302 - global_seqlen/minmax_diff:6854 - global_seqlen/balanced_min:23397 - global_seqlen/balanced_max:23404 - global_seqlen/mean:23402.125 - actor/entropy:0.11967595666646957 - perf/mfu/actor_infer:0 - training/rollout_probs_diff_valid:1 - training/rollout_probs_diff_max:0.1548396348953247 - training/rollout_probs_diff_mean:0.00242202403023839 - training/rollout_probs_diff_std:0.00918981060385704 - training/rollout_actor_probs_pearson_corr:0.9986111521720886 - self_distillation/success_group_fraction:0.75 - self_distillation/success_sample_fraction:0.74609375 - self_distillation/correct_sample_fraction:0.7109375 - self_distillation/feedback_available_fraction:0.0 - self_distillation/feedback_used_fraction:0.0 - self_distillation/reprompt_sample_fraction:0.74609375 - rollout_corr/rollout_is_mean:1.0000094175338745 - rollout_corr/rollout_is_ratio_fraction_high:0.0 - rollout_corr/rollout_is_ratio_fraction_low:0.0 - rollout_corr/rollout_is_max:1.624280571937561 - rollout_corr/rollout_is_min:0.5480358004570007 - rollout_corr/rollout_is_std:0.026753177866339684 - rollout_corr/rollout_is_eff_sample_size:0.9992848983711098 - rollout_corr/rollout_is_seq_mean:0.999933123588562 - rollout_corr/rollout_is_seq_std:0.001586279133334756 - rollout_corr/rollout_is_seq_max:1.005149245262146 - rollout_corr/rollout_is_seq_min:0.9930873513221741 - rollout_corr/rollout_is_seq_max_deviation:0.006912648677825928 - rollout_corr/rollout_is_seq_fraction_high:0.0 - rollout_corr/rollout_is_seq_fraction_low:0.0 - rollout_corr/training_ppl:np.float64(1.1280051534995437) - rollout_corr/training_log_ppl:np.float64(0.11896156305010663) - rollout_corr/kl:np.float64(0.0003956216399711998) - rollout_corr/k3_kl:np.float64(0.0004248532313830822) - rollout_corr/rollout_ppl:np.float64(1.1275437236763537) - rollout_corr/rollout_log_ppl:np.float64(0.11856594059645431) - rollout_corr/log_ppl_diff:np.float64(0.0003956224536523223) - rollout_corr/log_ppl_abs_diff:np.float64(0.0012849519116571173) - rollout_corr/log_ppl_diff_max:np.float64(0.0077836960554122925) - rollout_corr/log_ppl_diff_min:np.float64(-0.005988866090774536) - rollout_corr/ppl_ratio:np.float64(1.0003972640261054) - rollout_corr/chi2_token:np.float64(0.0009248631540685892) - rollout_corr/chi2_seq:np.float64(0.2403997452929616) - actor/pg_loss:np.float64(0.00034071970731019974) - actor/kl_loss:np.float64(0.0) - actor/pg_clipfrac:np.float64(0.00010108559263244388) - actor/ppo_kl:np.float64(-4.2658330890787965e-05) - actor/pg_clipfrac_lower:np.float64(0.0) - self_distillation/token_reweight_lambda:np.float64(0.5) - self_distillation/token_reweight_eps_w:np.float64(0.20000000000000004) - self_distillation/token_reweight_active_token_fraction:np.float64(0.74609375) - self_distillation/token_reweight_context_sample_fraction:np.float64(0.74609375) - self_distillation/token_reweight_correct_sample_fraction:np.float64(0.7109375) - self_distillation/token_reweight_w_mean:np.float64(18573.2146856694) - self_distillation/token_reweight_w_std:np.float64(329012.32793184) - self_distillation/token_reweight_w_clipped_mean:np.float64(1.0002196375280619) - self_distillation/token_reweight_w_clip_frac:np.float64(0.006969294132431969) - self_distillation/empty_target_batch:np.float64(0.25390625) - actor/grad_norm:np.float64(0.05942569300532341) - perf/mfu/actor:np.float64(0.05214527055013675) - perf/max_memory_allocated_gb:np.float64(130.98332166671753) - perf/max_memory_reserved_gb:np.float64(140.86328125) - perf/cpu_memory_used_gb:np.float64(196.1522102355957) - actor/lr:np.float64(1e-06) - training/global_step:95 - training/epoch:3 - critic/score/mean:0.7109375 - critic/score/max:1.0 - critic/score/min:0.0 - critic/rewards/mean:0.7109375 - critic/rewards/max:1.0 - critic/rewards/min:0.0 - critic/advantages/mean:0.003893822431564331 - critic/advantages/max:0.875 - critic/advantages/min:-0.625 - critic/returns/mean:0.003893822431564331 - critic/returns/max:0.875 - critic/returns/min:-0.625 - response_length/mean:335.19140625 - response_length/max:741.0 - response_l
|
||
step:96 - global_seqlen/min:20160 - global_seqlen/max:27007 - global_seqlen/minmax_diff:6847 - global_seqlen/balanced_min:24376 - global_seqlen/balanced_max:24378 - global_seqlen/mean:24377.25 - actor/entropy:0.11729797720909119 - perf/mfu/actor_infer:0 - training/rollout_probs_diff_valid:1 - training/rollout_probs_diff_max:0.2773001194000244 - training/rollout_probs_diff_mean:0.002411847235634923 - training/rollout_probs_diff_std:0.009429632686078548 - training/rollout_actor_probs_pearson_corr:0.9985706210136414 - self_distillation/success_group_fraction:0.625 - self_distillation/success_sample_fraction:0.625 - self_distillation/correct_sample_fraction:0.58984375 - self_distillation/feedback_available_fraction:0.0 - self_distillation/feedback_used_fraction:0.0 - self_distillation/reprompt_sample_fraction:0.625 - rollout_corr/rollout_is_mean:1.0000203847885132 - rollout_corr/rollout_is_ratio_fraction_high:1.1028520020772703e-05 - rollout_corr/rollout_is_ratio_fraction_low:1.1028520020772703e-05 - rollout_corr/rollout_is_max:2.404789447784424 - rollout_corr/rollout_is_min:0.44539976119995117 - rollout_corr/rollout_is_std:0.02721264585852623 - rollout_corr/rollout_is_eff_sample_size:0.9992601389023397 - rollout_corr/rollout_is_seq_mean:0.999981701374054 - rollout_corr/rollout_is_seq_std:0.001592422486282885 - rollout_corr/rollout_is_seq_max:1.0050090551376343 - rollout_corr/rollout_is_seq_min:0.9955622553825378 - rollout_corr/rollout_is_seq_max_deviation:0.005009055137634277 - rollout_corr/rollout_is_seq_fraction_high:0.0 - rollout_corr/rollout_is_seq_fraction_low:0.0 - rollout_corr/training_ppl:np.float64(1.1281056944280863) - rollout_corr/training_log_ppl:np.float64(0.11904539880924858) - rollout_corr/kl:np.float64(0.00041123665303999246) - rollout_corr/k3_kl:np.float64(0.0004748523673612226) - rollout_corr/rollout_ppl:np.float64(1.1276400778442621) - rollout_corr/rollout_log_ppl:np.float64(0.11863416097185109) - rollout_corr/log_ppl_diff:np.float64(0.0004112378373974934) - rollout_corr/log_ppl_abs_diff:np.float64(0.0012866570759797469) - rollout_corr/log_ppl_diff_max:np.float64(0.007371693849563599) - rollout_corr/log_ppl_diff_min:np.float64(-0.0064626410603523254) - rollout_corr/ppl_ratio:np.float64(1.0004128504078835) - rollout_corr/chi2_token:np.float64(0.0010835977736860514) - rollout_corr/chi2_seq:np.float64(0.24060776853002608) - actor/pg_loss:np.float64(0.0005744848167523742) - actor/kl_loss:np.float64(0.0) - actor/pg_clipfrac:np.float64(0.000119695831926947) - actor/ppo_kl:np.float64(1.3767832875544173e-05) - actor/pg_clipfrac_lower:np.float64(0.0) - self_distillation/token_reweight_lambda:np.float64(0.5) - self_distillation/token_reweight_eps_w:np.float64(0.20000000000000004) - self_distillation/token_reweight_active_token_fraction:np.float64(0.625) - self_distillation/token_reweight_context_sample_fraction:np.float64(0.625) - self_distillation/token_reweight_correct_sample_fraction:np.float64(0.58984375) - self_distillation/token_reweight_w_mean:np.float64(2840.6781995082274) - self_distillation/token_reweight_w_std:np.float64(53555.554571515124) - self_distillation/token_reweight_w_clipped_mean:np.float64(0.9997150534763932) - self_distillation/token_reweight_w_clip_frac:np.float64(0.015691433451138437) - self_distillation/empty_target_batch:np.float64(0.375) - actor/grad_norm:np.float64(0.10368238762021065) - perf/mfu/actor:np.float64(0.0549931270347762) - perf/max_memory_allocated_gb:np.float64(130.98332166671753) - perf/max_memory_reserved_gb:np.float64(140.86328125) - perf/cpu_memory_used_gb:np.float64(196.1422119140625) - actor/lr:np.float64(1e-06) - training/global_step:96 - training/epoch:3 - critic/score/mean:0.58984375 - critic/score/max:1.0 - critic/score/min:0.0 - critic/rewards/mean:0.58984375 - critic/rewards/max:1.0 - critic/rewards/min:0.0 - critic/advantages/mean:-9.925667836796492e-05 - critic/advantages/max:0.75 - critic/advantages/min:-0.875 - critic/returns/mean:-9.925667836796492e-05 - critic/returns/max:0.75 - critic/returns/min:-0.875 - response_length/mean:354.1953125 - response_length
|
||
step:97 - global_seqlen/min:20013 - global_seqlen/max:25357 - global_seqlen/minmax_diff:5344 - global_seqlen/balanced_min:22630 - global_seqlen/balanced_max:22638 - global_seqlen/mean:22635.25 - actor/entropy:0.1280536651611328 - perf/mfu/actor_infer:0 - training/rollout_probs_diff_valid:1 - training/rollout_probs_diff_max:0.18463575839996338 - training/rollout_probs_diff_mean:0.0024065831676125526 - training/rollout_probs_diff_std:0.009107198566198349 - training/rollout_actor_probs_pearson_corr:0.9987413883209229 - self_distillation/success_group_fraction:0.90625 - self_distillation/success_sample_fraction:0.90625 - self_distillation/correct_sample_fraction:0.84375 - self_distillation/feedback_available_fraction:0.0 - self_distillation/feedback_used_fraction:0.0 - self_distillation/reprompt_sample_fraction:0.90625 - rollout_corr/rollout_is_mean:1.0000966787338257 - rollout_corr/rollout_is_ratio_fraction_high:1.1470783647382632e-05 - rollout_corr/rollout_is_ratio_fraction_low:1.1470783647382632e-05 - rollout_corr/rollout_is_max:2.5134615898132324 - rollout_corr/rollout_is_min:0.4878324866294861 - rollout_corr/rollout_is_std:0.025915730744600296 - rollout_corr/rollout_is_eff_sample_size:0.9993290637389448 - rollout_corr/rollout_is_seq_mean:1.000105857849121 - rollout_corr/rollout_is_seq_std:0.00157264550216496 - rollout_corr/rollout_is_seq_max:1.00606107711792 - rollout_corr/rollout_is_seq_min:0.994608998298645 - rollout_corr/rollout_is_seq_max_deviation:0.006061077117919922 - rollout_corr/rollout_is_seq_fraction_high:0.0 - rollout_corr/rollout_is_seq_fraction_low:0.0 - rollout_corr/training_ppl:np.float64(1.1334166121669114) - rollout_corr/training_log_ppl:np.float64(0.12309159211145015) - rollout_corr/kl:np.float64(0.000282021749916872) - rollout_corr/k3_kl:np.float64(0.00044198074789392194) - rollout_corr/rollout_ppl:np.float64(1.1330834380351007) - rollout_corr/rollout_log_ppl:np.float64(0.1228095693004434) - rollout_corr/log_ppl_diff:np.float64(0.00028202281100675464) - rollout_corr/log_ppl_abs_diff:np.float64(0.0012496994895627722) - rollout_corr/log_ppl_diff_max:np.float64(0.007074952125549316) - rollout_corr/log_ppl_diff_min:np.float64(-0.00588814914226532) - rollout_corr/ppl_ratio:np.float64(1.000283483415842) - rollout_corr/chi2_token:np.float64(0.0012291779275983572) - rollout_corr/chi2_seq:np.float64(0.3651922077406198) - actor/pg_loss:np.float64(0.000943023944273591) - actor/kl_loss:np.float64(0.0) - actor/pg_clipfrac:np.float64(0.0001242990306309366) - actor/ppo_kl:np.float64(4.270520436566727e-05) - actor/pg_clipfrac_lower:np.float64(0.0) - self_distillation/token_reweight_lambda:np.float64(0.5) - self_distillation/token_reweight_eps_w:np.float64(0.20000000000000004) - self_distillation/token_reweight_active_token_fraction:np.float64(0.90625) - self_distillation/token_reweight_context_sample_fraction:np.float64(0.90625) - self_distillation/token_reweight_correct_sample_fraction:np.float64(0.84375) - self_distillation/token_reweight_w_mean:np.float64(30499.929090753663) - self_distillation/token_reweight_w_std:np.float64(639412.5114061498) - self_distillation/token_reweight_w_clipped_mean:np.float64(1.0000444219913334) - self_distillation/token_reweight_w_clip_frac:np.float64(0.015111897548194975) - self_distillation/empty_target_batch:np.float64(0.09375) - actor/grad_norm:np.float64(0.09439769387245178) - perf/mfu/actor:np.float64(0.051329233916159775) - perf/max_memory_allocated_gb:np.float64(130.98332166671753) - perf/max_memory_reserved_gb:np.float64(140.86328125) - perf/cpu_memory_used_gb:np.float64(196.2738265991211) - actor/lr:np.float64(1e-06) - training/global_step:97 - training/epoch:3 - critic/score/mean:0.84375 - critic/score/max:1.0 - critic/score/min:0.0 - critic/rewards/mean:0.84375 - critic/rewards/max:1.0 - critic/rewards/min:0.0 - critic/advantages/mean:0.00036132970126345754 - critic/advantages/max:0.75 - critic/advantages/min:-0.75 - critic/returns/mean:0.00036132970126345754 - critic/returns/max:0.75 - critic/returns/min:-0.75 - response_length/mean:340.5390625 - response_length/max:751.
|
||
step:98 - global_seqlen/min:21562 - global_seqlen/max:27701 - global_seqlen/minmax_diff:6139 - global_seqlen/balanced_min:24212 - global_seqlen/balanced_max:24218 - global_seqlen/mean:24216.375 - actor/entropy:0.1430254727602005 - perf/mfu/actor_infer:0 - training/rollout_probs_diff_valid:1 - training/rollout_probs_diff_max:0.20211440324783325 - training/rollout_probs_diff_mean:0.002688319655135274 - training/rollout_probs_diff_std:0.009453684091567993 - training/rollout_actor_probs_pearson_corr:0.9988110065460205 - self_distillation/success_group_fraction:0.9375 - self_distillation/success_sample_fraction:0.93359375 - self_distillation/correct_sample_fraction:0.89453125 - self_distillation/feedback_available_fraction:0.0 - self_distillation/feedback_used_fraction:0.0 - self_distillation/reprompt_sample_fraction:0.93359375 - rollout_corr/rollout_is_mean:0.9998697638511658 - rollout_corr/rollout_is_ratio_fraction_high:0.0 - rollout_corr/rollout_is_ratio_fraction_low:0.0 - rollout_corr/rollout_is_max:1.6097447872161865 - rollout_corr/rollout_is_min:0.5124151706695557 - rollout_corr/rollout_is_std:0.027273906394839287 - rollout_corr/rollout_is_eff_sample_size:0.9992565679263558 - rollout_corr/rollout_is_seq_mean:0.9998807907104492 - rollout_corr/rollout_is_seq_std:0.0015214976156130433 - rollout_corr/rollout_is_seq_max:1.0076751708984375 - rollout_corr/rollout_is_seq_min:0.9955283999443054 - rollout_corr/rollout_is_seq_max_deviation:0.0076751708984375 - rollout_corr/rollout_is_seq_fraction_high:0.0 - rollout_corr/rollout_is_seq_fraction_low:0.0 - rollout_corr/training_ppl:np.float64(1.151344778481871) - rollout_corr/training_log_ppl:np.float64(0.13949581736596883) - rollout_corr/kl:np.float64(0.0008577270308975926) - rollout_corr/k3_kl:np.float64(0.0005512973715582348) - rollout_corr/rollout_ppl:np.float64(1.1503589190542698) - rollout_corr/rollout_log_ppl:np.float64(0.13863808915994014) - rollout_corr/log_ppl_diff:np.float64(0.0008577282060286961) - rollout_corr/log_ppl_abs_diff:np.float64(0.001595133486262057) - rollout_corr/log_ppl_diff_max:np.float64(0.008243545889854431) - rollout_corr/log_ppl_diff_min:np.float64(-0.006423532962799072) - rollout_corr/ppl_ratio:np.float64(1.0008599534630775) - rollout_corr/chi2_token:np.float64(0.00048058945685625076) - rollout_corr/chi2_seq:np.float64(0.11350936791859567) - actor/pg_loss:np.float64(0.0004699843702837825) - actor/kl_loss:np.float64(0.0) - actor/pg_clipfrac:np.float64(0.00018328063106309855) - actor/ppo_kl:np.float64(0.0003591571183854114) - actor/pg_clipfrac_lower:np.float64(0.0) - self_distillation/token_reweight_lambda:np.float64(0.5) - self_distillation/token_reweight_eps_w:np.float64(0.20000000000000004) - self_distillation/token_reweight_active_token_fraction:np.float64(0.93359375) - self_distillation/token_reweight_context_sample_fraction:np.float64(0.93359375) - self_distillation/token_reweight_correct_sample_fraction:np.float64(0.89453125) - self_distillation/token_reweight_w_mean:np.float64(452.343374283053) - self_distillation/token_reweight_w_std:np.float64(8862.90448090632) - self_distillation/token_reweight_w_clipped_mean:np.float64(0.9999787393026054) - self_distillation/token_reweight_w_clip_frac:np.float64(0.011719158239429817) - self_distillation/empty_target_batch:np.float64(0.06640625) - actor/grad_norm:np.float64(0.08010996598750353) - perf/mfu/actor:np.float64(0.05323706476936674) - perf/max_memory_allocated_gb:np.float64(130.98332166671753) - perf/max_memory_reserved_gb:np.float64(140.86328125) - perf/cpu_memory_used_gb:np.float64(196.16266250610352) - actor/lr:np.float64(1e-06) - training/global_step:98 - training/epoch:3 - critic/score/mean:0.89453125 - critic/score/max:1.0 - critic/score/min:0.0 - critic/rewards/mean:0.89453125 - critic/rewards/max:1.0 - critic/rewards/min:0.0 - critic/advantages/mean:-0.0006149417022243142 - critic/advantages/max:0.875 - critic/advantages/min:-0.875 - critic/returns/mean:-0.0006149417022243142 - critic/returns/max:0.875 - critic/returns/min:-0.875 - response_length/mean:354.13671875 - response_length/max:725.0
|
||
step:99 - global_seqlen/min:20094 - global_seqlen/max:26591 - global_seqlen/minmax_diff:6497 - global_seqlen/balanced_min:23799 - global_seqlen/balanced_max:23806 - global_seqlen/mean:23804.125 - actor/entropy:0.15504607558250427 - perf/mfu/actor_infer:0 - training/rollout_probs_diff_valid:1 - training/rollout_probs_diff_max:0.4744115471839905 - training/rollout_probs_diff_mean:0.002842691959813237 - training/rollout_probs_diff_std:0.009904769249260426 - training/rollout_actor_probs_pearson_corr:0.9986906051635742 - self_distillation/success_group_fraction:0.8125 - self_distillation/success_sample_fraction:0.8125 - self_distillation/correct_sample_fraction:0.765625 - self_distillation/feedback_available_fraction:0.0 - self_distillation/feedback_used_fraction:0.0 - self_distillation/reprompt_sample_fraction:0.8125 - rollout_corr/rollout_is_mean:1.000064492225647 - rollout_corr/rollout_is_ratio_fraction_high:0.0 - rollout_corr/rollout_is_ratio_fraction_low:1.117555711971363e-05 - rollout_corr/rollout_is_max:1.6279256343841553 - rollout_corr/rollout_is_min:0.44314810633659363 - rollout_corr/rollout_is_std:0.0281895212829113 - rollout_corr/rollout_is_eff_sample_size:0.9992061008617588 - rollout_corr/rollout_is_seq_mean:1.0000643730163574 - rollout_corr/rollout_is_seq_std:0.0016942165093496442 - rollout_corr/rollout_is_seq_max:1.006766438484192 - rollout_corr/rollout_is_seq_min:0.9935619235038757 - rollout_corr/rollout_is_seq_max_deviation:0.0067664384841918945 - rollout_corr/rollout_is_seq_fraction_high:0.0 - rollout_corr/rollout_is_seq_fraction_low:0.0 - rollout_corr/training_ppl:np.float64(1.1785350814461708) - rollout_corr/training_log_ppl:np.float64(0.15894308649149025) - rollout_corr/kl:np.float64(0.0013515585573031785) - rollout_corr/k3_kl:np.float64(0.0012069832163774663) - rollout_corr/rollout_ppl:np.float64(1.1769968876615167) - rollout_corr/rollout_log_ppl:np.float64(0.1575915266002994) - rollout_corr/log_ppl_diff:np.float64(0.0013515599057427607) - rollout_corr/log_ppl_abs_diff:np.float64(0.002168246872315649) - rollout_corr/log_ppl_diff_max:np.float64(0.06749674677848816) - rollout_corr/log_ppl_diff_min:np.float64(-0.004221692681312561) - rollout_corr/ppl_ratio:np.float64(1.0013778761494905) - rollout_corr/chi2_token:np.float64(0.0007215156219899654) - rollout_corr/chi2_seq:np.float64(0.20517127867788076) - actor/pg_loss:np.float64(0.0009915356058627367) - actor/kl_loss:np.float64(0.0) - actor/pg_clipfrac:np.float64(0.00020690295605163556) - actor/ppo_kl:np.float64(0.000960174699570826) - actor/pg_clipfrac_lower:np.float64(0.0) - self_distillation/token_reweight_lambda:np.float64(0.5) - self_distillation/token_reweight_eps_w:np.float64(0.20000000000000004) - self_distillation/token_reweight_active_token_fraction:np.float64(0.8125) - self_distillation/token_reweight_context_sample_fraction:np.float64(0.8125) - self_distillation/token_reweight_correct_sample_fraction:np.float64(0.765625) - self_distillation/token_reweight_w_mean:np.float64(7099.986027709674) - self_distillation/token_reweight_w_std:np.float64(116633.08640508627) - self_distillation/token_reweight_w_clipped_mean:np.float64(0.9997013744432479) - self_distillation/token_reweight_w_clip_frac:np.float64(0.021721423909184523) - self_distillation/empty_target_batch:np.float64(0.1875) - actor/grad_norm:np.float64(0.15946919191628695) - perf/mfu/actor:np.float64(0.05383568726534263) - perf/max_memory_allocated_gb:np.float64(130.98332166671753) - perf/max_memory_reserved_gb:np.float64(140.86328125) - perf/cpu_memory_used_gb:np.float64(196.18372344970703) - actor/lr:np.float64(1e-06) - training/global_step:99 - training/epoch:3 - critic/score/mean:0.765625 - critic/score/max:1.0 - critic/score/min:0.0 - critic/rewards/mean:0.765625 - critic/rewards/max:1.0 - critic/rewards/min:0.0 - critic/advantages/mean:-0.0029405683744698763 - critic/advantages/max:0.625 - critic/advantages/min:-0.875 - critic/returns/mean:-0.0029405683744698763 - critic/returns/max:0.625 - critic/returns/min:-0.875 - response_length/mean:349.53515625 - response_length/max:706.0 - response
|
||
test_gen_batch meta info: {'eos_token_id': 151645, 'pad_token_id': 151643, 'recompute_log_prob': False, 'do_sample': True, 'validate': True, 'global_steps': 100}
|
||
validation generation end
|
||
[prompt] system
|
||
|
||
Given a question and four options, please select the right answer. Respond in the following format:
|
||
<reasoning>
|
||
...
|
||
</reasoning>
|
||
<answer>
|
||
...
|
||
</answer>
|
||
|
||
For the answer, only output the letter corresponding to the correct option (A, B, C, or D), and nothing else. Do not restate the answer text. For example, if the answer is "A", just output:
|
||
<answer>
|
||
A
|
||
</answer>
|
||
|
||
user
|
||
Given the following crystal structure parameters for the material RbLa9(IrO6)4 (Material ID: mp-560657), calculate the volume of the unit cell (in ų).
|
||
|
||
Lattice Parameters:
|
||
|
||
a = 7.82 Å
|
||
b = 7.82 Å
|
||
c = 17.88 Å
|
||
α = 90.00º
|
||
β = 90.00º
|
||
γ = 90.00º
|
||
|
||
|
||
|
||
A: 1025.67 ų
|
||
|
||
B: 1094.31 ų
|
||
|
||
C: 1200.45 ų
|
||
|
||
D: 1150.78 ų
|
||
|
||
Please reason step by step.
|
||
assistant
|
||
<think>
|
||
|
||
</think>
|
||
|
||
|
||
[response] <reasoning>
|
||
The given crystal structure has lattice parameters a = 7.82 Å, b = 7.82 Å, c = 17.88 Å, and all angles are 90 degrees, indicating a orthorhombic unit cell. The volume of a unit cell in an orthorhombic system is calculated using the formula:
|
||
|
||
V = a × b × c
|
||
|
||
Substituting the given values:
|
||
|
||
V = 7.82 × 7.82 × 17.88
|
||
|
||
First, calculate 7.82 × 7.82:
|
||
|
||
7.82 × 7.82 = 61.1524
|
||
|
||
Then multiply by 17.88:
|
||
|
||
61.1524 × 17.88 ≈ 1094.31
|
||
|
||
Therefore, the volume of the unit cell is approximately 1094.31 ų.
|
||
</reasoning>
|
||
<answer>
|
||
B
|
||
</answer>
|
||
[ground_truth] B
|
||
[score] 1.0
|
||
[acc] 1.0
|
||
[pred] B
|
||
[incorrect_format] 1
|
||
[feedback]
|
||
local_global_step_folder: /mnt/mole/SDPO/L2T/checkpoints/datasets/sciknoweval/material/qwen3gen-material-RLSD_TR-Qwen-Qwen3-8B-mbs8-decay0-tr0.1-train32-rollout8-lr1e-6-vllm0.8/global_step_100
|
||
step:100 - global_seqlen/min:20320 - global_seqlen/max:28053 - global_seqlen/minmax_diff:7733 - global_seqlen/balanced_min:23678 - global_seqlen/balanced_max:23681 - global_seqlen/mean:23679.5 - actor/entropy:0.15241530537605286 - perf/mfu/actor_infer:0 - training/rollout_probs_diff_valid:1 - training/rollout_probs_diff_max:0.25462448596954346 - training/rollout_probs_diff_mean:0.002963797887787223 - training/rollout_probs_diff_std:0.010244213975965977 - training/rollout_actor_probs_pearson_corr:0.9986335039138794 - self_distillation/success_group_fraction:0.6875 - self_distillation/success_sample_fraction:0.6875 - self_distillation/correct_sample_fraction:0.67578125 - self_distillation/feedback_available_fraction:0.0 - self_distillation/feedback_used_fraction:0.0 - self_distillation/reprompt_sample_fraction:0.6875 - rollout_corr/rollout_is_mean:1.0000261068344116 - rollout_corr/rollout_is_ratio_fraction_high:1.119720491260523e-05 - rollout_corr/rollout_is_ratio_fraction_low:0.0 - rollout_corr/rollout_is_max:2.0033769607543945 - rollout_corr/rollout_is_min:0.567776620388031 - rollout_corr/rollout_is_std:0.029670871794223785 - rollout_corr/rollout_is_eff_sample_size:0.9991205327494357 - rollout_corr/rollout_is_seq_mean:1.0000578165054321 - rollout_corr/rollout_is_seq_std:0.0018322683172300458 - rollout_corr/rollout_is_seq_max:1.0058804750442505 - rollout_corr/rollout_is_seq_min:0.9950834512710571 - rollout_corr/rollout_is_seq_max_deviation:0.005880475044250488 - rollout_corr/rollout_is_seq_fraction_high:0.0 - rollout_corr/rollout_is_seq_fraction_low:0.0 - rollout_corr/training_ppl:np.float64(1.1711667915806174) - rollout_corr/training_log_ppl:np.float64(0.15226043093571207) - rollout_corr/kl:np.float64(0.00048633673415032774) - rollout_corr/k3_kl:np.float64(0.0005004820688583322) - rollout_corr/rollout_ppl:np.float64(1.170548481401056) - rollout_corr/rollout_log_ppl:np.float64(0.15177409329044167) - rollout_corr/log_ppl_diff:np.float64(0.000486337645270396) - rollout_corr/log_ppl_abs_diff:np.float64(0.0015185297597781755) - rollout_corr/log_ppl_diff_max:np.float64(0.008212178945541382) - rollout_corr/log_ppl_diff_min:np.float64(-0.004179961979389191) - rollout_corr/ppl_ratio:np.float64(1.0004883385263383) - rollout_corr/chi2_token:np.float64(0.0010083636734634638) - rollout_corr/chi2_seq:np.float64(0.34577386314049363) - actor/pg_loss:np.float64(0.0005346338730305433) - actor/kl_loss:np.float64(0.0) - actor/pg_clipfrac:np.float64(5.5688376050966326e-05) - actor/ppo_kl:np.float64(0.00010043691768402141) - actor/pg_clipfrac_lower:np.float64(0.0) - self_distillation/token_reweight_lambda:np.float64(0.5) - self_distillation/token_reweight_eps_w:np.float64(0.20000000000000004) - self_distillation/token_reweight_active_token_fraction:np.float64(0.6875) - self_distillation/token_reweight_context_sample_fraction:np.float64(0.6875) - self_distillation/token_reweight_correct_sample_fraction:np.float64(0.67578125) - self_distillation/token_reweight_w_mean:np.float64(1.0763714532367885) - self_distillation/token_reweight_w_std:np.float64(0.7491657065111212) - self_distillation/token_reweight_w_clipped_mean:np.float64(0.9991688379086554) - self_distillation/token_reweight_w_clip_frac:np.float64(0.016922423528740183) - self_distillation/empty_target_batch:np.float64(0.3125) - actor/grad_norm:np.float64(0.08581123873591423) - perf/mfu/actor:np.float64(0.05333022759720618) - perf/max_memory_allocated_gb:np.float64(130.98332166671753) - perf/max_memory_reserved_gb:np.float64(140.86328125) - perf/cpu_memory_used_gb:np.float64(196.12500381469727) - actor/lr:np.float64(1e-06) - val-aux/sciknoweval/reward/mean@16:np.float64(0.7646276595744681) - val-aux/sciknoweval/reward/std@16:np.float64(0.0401737640202366) - val-aux/sciknoweval/reward/best@2/mean:np.float64(0.7809255319148936) - val-aux/sciknoweval/reward/best@2/std:np.float64(0.03462769102759328) - val-aux/sciknoweval/reward/worst@2/mean:np.float64(0.7494468085106384) - val-aux/sciknoweval/reward/worst@2/std:np.float64(0.03367836377537821) - val-aux/sciknoweval/reward/maj@2/mean:np.float6
|
||
("Final validation metrics: {'val-aux/sciknoweval/reward/mean@16': "
|
||
"np.float64(0.7646276595744681), 'val-aux/sciknoweval/reward/std@16': "
|
||
"np.float64(0.0401737640202366), 'val-aux/sciknoweval/reward/best@2/mean': "
|
||
"np.float64(0.7809255319148936), 'val-aux/sciknoweval/reward/best@2/std': "
|
||
"np.float64(0.03462769102759328), 'val-aux/sciknoweval/reward/worst@2/mean': "
|
||
"np.float64(0.7494468085106384), 'val-aux/sciknoweval/reward/worst@2/std': "
|
||
"np.float64(0.03367836377537821), 'val-aux/sciknoweval/reward/maj@2/mean': "
|
||
"np.float64(0.7651063829787235), 'val-aux/sciknoweval/reward/maj@2/std': "
|
||
"np.float64(0.040366939474519616), 'val-aux/sciknoweval/reward/best@4/mean': "
|
||
"np.float64(0.7944468085106383), 'val-aux/sciknoweval/reward/best@4/std': "
|
||
"np.float64(0.027759320424956824), 'val-aux/sciknoweval/reward/worst@4/mean': "
|
||
"np.float64(0.7366170212765957), 'val-aux/sciknoweval/reward/worst@4/std': "
|
||
"np.float64(0.025172389625263896), 'val-aux/sciknoweval/reward/maj@4/mean': "
|
||
"np.float64(0.764436170212766), 'val-aux/sciknoweval/reward/maj@4/std': "
|
||
"np.float64(0.031638291528035434), 'val-aux/sciknoweval/reward/best@8/mean': "
|
||
"np.float64(0.8061702127659575), 'val-aux/sciknoweval/reward/best@8/std': "
|
||
"np.float64(0.019636408581214976), 'val-aux/sciknoweval/reward/worst@8/mean': "
|
||
"np.float64(0.726340425531915), 'val-aux/sciknoweval/reward/worst@8/std': "
|
||
"np.float64(0.018341796156463988), 'val-aux/sciknoweval/reward/maj@8/mean': "
|
||
"np.float64(0.764223404255319), 'val-aux/sciknoweval/reward/maj@8/std': "
|
||
"np.float64(0.020384144360752216), 'val-aux/sciknoweval/reward/best@16/mean': "
|
||
"np.float64(0.813808510638298), 'val-aux/sciknoweval/reward/best@16/std': "
|
||
'np.float64(0.011791035148083585), '
|
||
"'val-aux/sciknoweval/reward/worst@16/mean': np.float64(0.7182659574468084), "
|
||
"'val-aux/sciknoweval/reward/worst@16/std': np.float64(0.01148574534900323), "
|
||
"'val-aux/sciknoweval/reward/maj@16/mean': np.float64(0.7633617021276595), "
|
||
"'val-aux/sciknoweval/reward/maj@16/std': np.float64(0.013016076917955125), "
|
||
"'val-aux/sciknoweval/score/mean@16': np.float64(0.7646276595744681), "
|
||
"'val-aux/sciknoweval/score/std@16': np.float64(0.0401737640202366), "
|
||
"'val-aux/sciknoweval/score/best@2/mean': np.float64(0.7809255319148936), "
|
||
"'val-aux/sciknoweval/score/best@2/std': np.float64(0.03462769102759328), "
|
||
"'val-aux/sciknoweval/score/worst@2/mean': np.float64(0.7494468085106384), "
|
||
"'val-aux/sciknoweval/score/worst@2/std': np.float64(0.03367836377537821), "
|
||
"'val-aux/sciknoweval/score/maj@2/mean': np.float64(0.7651063829787235), "
|
||
"'val-aux/sciknoweval/score/maj@2/std': np.float64(0.040366939474519616), "
|
||
"'val-aux/sciknoweval/score/best@4/mean': np.float64(0.7944468085106383), "
|
||
"'val-aux/sciknoweval/score/best@4/std': np.float64(0.027759320424956824), "
|
||
"'val-aux/sciknoweval/score/worst@4/mean': np.float64(0.7366170212765957), "
|
||
"'val-aux/sciknoweval/score/worst@4/std': np.float64(0.025172389625263896), "
|
||
"'val-aux/sciknoweval/score/maj@4/mean': np.float64(0.764436170212766), "
|
||
"'val-aux/sciknoweval/score/maj@4/std': np.float64(0.031638291528035434), "
|
||
"'val-aux/sciknoweval/score/best@8/mean': np.float64(0.8061702127659575), "
|
||
"'val-aux/sciknoweval/score/best@8/std': np.float64(0.019636408581214976), "
|
||
"'val-aux/sciknoweval/score/worst@8/mean': np.float64(0.726340425531915), "
|
||
"'val-aux/sciknoweval/score/worst@8/std': np.float64(0.018341796156463988), "
|
||
"'val-aux/sciknoweval/score/maj@8/mean': np.float64(0.764223404255319), "
|
||
"'val-aux/sciknoweval/score/maj@8/std': np.float64(0.020384144360752216), "
|
||
"'val-aux/sciknoweval/score/best@16/mean': np.float64(0.813808510638298), "
|
||
"'val-aux/sciknoweval/score/best@16/std': np.float64(0.011791035148083585), "
|
||
"'val-aux/sciknoweval/score/worst@16/mean': np.float64(0.7182659574468084), "
|
||
"'val-aux/sciknoweval/score/worst@16/std': np.float64(0.01148574534900323), "
|
||
"'val-aux/sciknoweval/score/maj@16/mean': np.float64(0.7633617021276595), "
|
||
"'val-aux/sciknoweval/score/maj@16/std': np.float64(0.013016076917955125), "
|
||
"'val-core/sciknoweval/acc/mean@16': np.float64(0.7646276595744681), "
|
||
"'val-aux/sciknoweval/acc/std@16': np.float64(0.0401737640202366), "
|
||
"'val-aux/sciknoweval/acc/best@2/mean': np.float64(0.7809255319148936), "
|
||
"'val-aux/sciknoweval/acc/best@2/std': np.float64(0.03462769102759328), "
|
||
"'val-aux/sciknoweval/acc/worst@2/mean': np.float64(0.7494468085106384), "
|
||
"'val-aux/sciknoweval/acc/worst@2/std': np.float64(0.03367836377537821), "
|
||
"'val-aux/sciknoweval/acc/maj@2/mean': np.float64(0.7651063829787235), "
|
||
"'val-aux/sciknoweval/acc/maj@2/std': np.float64(0.040366939474519616), "
|
||
"'val-aux/sciknoweval/acc/best@4/mean': np.float64(0.7944468085106383), "
|
||
"'val-aux/sciknoweval/acc/best@4/std': np.float64(0.027759320424956824), "
|
||
"'val-aux/sciknoweval/acc/worst@4/mean': np.float64(0.7366170212765957), "
|
||
"'val-aux/sciknoweval/acc/worst@4/std': np.float64(0.025172389625263896), "
|
||
"'val-aux/sciknoweval/acc/maj@4/mean': np.float64(0.764436170212766), "
|
||
"'val-aux/sciknoweval/acc/maj@4/std': np.float64(0.031638291528035434), "
|
||
"'val-aux/sciknoweval/acc/best@8/mean': np.float64(0.8061702127659575), "
|
||
"'val-aux/sciknoweval/acc/best@8/std': np.float64(0.019636408581214976), "
|
||
"'val-aux/sciknoweval/acc/worst@8/mean': np.float64(0.726340425531915), "
|
||
"'val-aux/sciknoweval/acc/worst@8/std': np.float64(0.018341796156463988), "
|
||
"'val-aux/sciknoweval/acc/maj@8/mean': np.float64(0.764223404255319), "
|
||
"'val-aux/sciknoweval/acc/maj@8/std': np.float64(0.020384144360752216), "
|
||
"'val-core/sciknoweval/acc/best@16/mean': np.float64(0.813808510638298), "
|
||
"'val-core/sciknoweval/acc/best@16/std': np.float64(0.011791035148083585), "
|
||
"'val-aux/sciknoweval/acc/worst@16/mean': np.float64(0.7182659574468084), "
|
||
"'val-aux/sciknoweval/acc/worst@16/std': np.float64(0.01148574534900323), "
|
||
"'val-core/sciknoweval/acc/maj@16/mean': np.float64(0.7633617021276595), "
|
||
"'val-core/sciknoweval/acc/maj@16/std': np.float64(0.013016076917955125), "
|
||
"'val-aux/sciknoweval/incorrect_format/mean@16': np.float64(1.0), "
|
||
"'val-aux/sciknoweval/incorrect_format/std@16': np.float64(0.0), "
|
||
"'val-aux/sciknoweval/incorrect_format/best@2/mean': np.float64(1.0), "
|
||
"'val-aux/sciknoweval/incorrect_format/best@2/std': np.float64(0.0), "
|
||
"'val-aux/sciknoweval/incorrect_format/worst@2/mean': np.float64(1.0), "
|
||
"'val-aux/sciknoweval/incorrect_format/worst@2/std': np.float64(0.0), "
|
||
"'val-aux/sciknoweval/incorrect_format/maj@2/mean': np.float64(1.0), "
|
||
"'val-aux/sciknoweval/incorrect_format/maj@2/std': np.float64(0.0), "
|
||
"'val-aux/sciknoweval/incorrect_format/best@4/mean': np.float64(1.0), "
|
||
"'val-aux/sciknoweval/incorrect_format/best@4/std': np.float64(0.0), "
|
||
"'val-aux/sciknoweval/incorrect_format/worst@4/mean': np.float64(1.0), "
|
||
"'val-aux/sciknoweval/incorrect_format/worst@4/std': np.float64(0.0), "
|
||
"'val-aux/sciknoweval/incorrect_format/maj@4/mean': np.float64(1.0), "
|
||
"'val-aux/sciknoweval/incorrect_format/maj@4/std': np.float64(0.0), "
|
||
"'val-aux/sciknoweval/incorrect_format/best@8/mean': np.float64(1.0), "
|
||
"'val-aux/sciknoweval/incorrect_format/best@8/std': np.float64(0.0), "
|
||
"'val-aux/sciknoweval/incorrect_format/worst@8/mean': np.float64(1.0), "
|
||
"'val-aux/sciknoweval/incorrect_format/worst@8/std': np.float64(0.0), "
|
||
"'val-aux/sciknoweval/incorrect_format/maj@8/mean': np.float64(1.0), "
|
||
"'val-aux/sciknoweval/incorrect_format/maj@8/std': np.float64(0.0), "
|
||
"'val-aux/sciknoweval/incorrect_format/best@16/mean': np.float64(1.0), "
|
||
"'val-aux/sciknoweval/incorrect_format/best@16/std': np.float64(0.0), "
|
||
"'val-aux/sciknoweval/incorrect_format/worst@16/mean': np.float64(1.0), "
|
||
"'val-aux/sciknoweval/incorrect_format/worst@16/std': np.float64(0.0), "
|
||
"'val-aux/sciknoweval/incorrect_format/maj@16/mean': np.float64(1.0), "
|
||
"'val-aux/sciknoweval/incorrect_format/maj@16/std': np.float64(0.0), "
|
||
"'val-aux/num_turns/min': np.int32(2), 'val-aux/num_turns/max': np.int32(2), "
|
||
"'val-aux/num_turns/mean': np.float64(2.0)}")
|