初始化项目，由ModelHub XC社区提供模型

Model: SeongryongJung/Qwen3-8B-Physics-RLSD-TR
Source: Original Platform

results/hyperparameters.csv

@@ -0,0 +1,38 @@
+section,parameter,value,source
+Run identity,Base model,Qwen/Qwen3-8B,queue/script override
+Run identity,Dataset,Physics / SciKnowEval physics,run_qwen3_generalization.sh
+Run identity,Method,RLSD_TR,run_qwen3_generalization.sh
+Run identity,Config,rlsd,run_qwen3_generalization.sh
+Run identity,Experiment,qwen3gen-physics-RLSD_TR-Qwen-Qwen3-8B-mbs8-decay0-tr0.1-train32-rollout8-lr1e-6-vllm0.8,run_qwen3_generalization.sh
+Run identity,W&B run,run-20260703_055133-5wz4kyyd,wandb
+Data,Train file,datasets/sciknoweval/physics/train.parquet,script override
+Data,Validation file,datasets/sciknoweval/physics/test.parquet,script override
+Data,Train batch size,32,queue/script override
+Data,Train max samples,3200,queue/script override
+Schedule,Total training steps,100,queue/script override
+Schedule,Validation before train,False,queue/script override
+Schedule,Save frequency,10,queue/script override
+Schedule,Validation frequency,10,queue/script override
+Sequence,Max prompt length,2048,queue/script override
+Sequence,Max response length,8192,queue/script override
+Sequence,Max model length,10240,queue/script override
+Rollout,Train rollout n,8,queue/script override
+Rollout,Validation rollout n,16,queue/script override
+Rollout,vLLM GPU memory utilization,0.8,queue/script override
+Optimization,Learning rate,1e-6,RLSD_TR method override
+Optimization,Weight decay,0.01,script override
+PPO/GRPO,PPO mini batch size,8,queue/script override
+PPO/GRPO,Normalize GRPO advantages by std,False,baseline_grpo.yaml / script override
+Rollout correction,Importance sampling mode,token,script override
+Rollout correction,IS threshold,2.0,script override
+Checkpoint/Logging,Checkpoint root,checkpoints/datasets/sciknoweval/physics/qwen3gen-physics-RLSD_TR-Qwen-Qwen3-8B-mbs8-decay0-tr0.1-train32-rollout8-lr1e-6-vllm0.8,script override
+Checkpoint/Logging,Latest checkpointed iteration,100,latest_checkpointed_iteration.txt
+Checkpoint/Logging,External actor archive,checkpoints/datasets/sciknoweval/physics/qwen3gen-physics-RLSD_TR-Qwen-Qwen3-8B-mbs8-decay0-tr0.1-train32-rollout8-lr1e-6-vllm0.8/_actor_archive,preserve_actor_checkpoints.py
+Checkpoint/Logging,Logger,"console, wandb",ppo_trainer.yaml
+RLSD_TR,Policy loss mode,rlsd,method override
+RLSD_TR,Teacher regularization,trust-region,method override
+RLSD_TR,Trust-region mix / teacher update rate,0.1,queue/script override
+RLSD_TR,Token reweight lambda,0.5,queue/script override
+RLSD_TR,Token reweight eps_w,0.2,queue/script override
+RLSD_TR,Token reweight decay steps,0,queue/script override
+RLSD_TR,Fused kernels,False,method override

results/summary.json

@@ -0,0 +1,29 @@
+{
+  "repo_id": "SeongryongJung/Qwen3-8B-Physics-RLSD-TR",
+  "output_dir": "/mnt/mole/SDPO/L2T/hf_upload_tr/Qwen3-8B-Physics-RLSD-TR",
+  "experiment": "qwen3gen-physics-RLSD_TR-Qwen-Qwen3-8B-mbs8-decay0-tr0.1-train32-rollout8-lr1e-6-vllm0.8",
+  "best_step": 90,
+  "best_val_mean16": 0.68046875,
+  "best_actor_dir": "/mnt/mole/SDPO/L2T/checkpoints/datasets/sciknoweval/physics/qwen3gen-physics-RLSD_TR-Qwen-Qwen3-8B-mbs8-decay0-tr0.1-train32-rollout8-lr1e-6-vllm0.8/global_step_90/actor",
+  "final_step": 100,
+  "final_val_mean16": 0.6578125,
+  "final_actor_dir": "/mnt/mole/SDPO/L2T/checkpoints/datasets/sciknoweval/physics/qwen3gen-physics-RLSD_TR-Qwen-Qwen3-8B-mbs8-decay0-tr0.1-train32-rollout8-lr1e-6-vllm0.8/global_step_100/actor",
+  "train_rows": 90,
+  "val_rows": 10,
+  "hf_model_files": [
+    "added_tokens.json",
+    "chat_template.jinja",
+    "config.json",
+    "generation_config.json",
+    "merges.txt",
+    "model.safetensors.index.json",
+    "special_tokens_map.json",
+    "tokenizer.json",
+    "tokenizer_config.json",
+    "vocab.json",
+    "model-00001-of-00004.safetensors",
+    "model-00002-of-00004.safetensors",
+    "model-00003-of-00004.safetensors",
+    "model-00004-of-00004.safetensors"
+  ]
+}

results/training_score.png

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c48fb4344f9c26c21ce4d05db024b61d04ecf215c840b7dabedf68bd14e363bd
+size 133374

results/training_scores.csv

@@ -0,0 +1,91 @@
+step,critic_score_mean
+1,0.59765625
+2,0.59375
+3,0.546875
+4,0.6640625
+5,0.76171875
+6,0.609375
+7,0.74609375
+8,0.671875
+9,0.63671875
+11,0.640625
+12,0.5703125
+13,0.55078125
+14,0.6796875
+15,0.69140625
+16,0.76953125
+17,0.68359375
+18,0.421875
+19,0.73046875
+21,0.66015625
+22,0.640625
+23,0.4453125
+24,0.7109375
+25,0.67578125
+26,0.68359375
+27,0.765625
+28,0.6171875
+29,0.7421875
+31,0.625
+32,0.74609375
+33,0.74609375
+34,0.68359375
+35,0.71875
+36,0.55859375
+37,0.66796875
+38,0.6484375
+39,0.80859375
+41,0.7734375
+42,0.6171875
+43,0.671875
+44,0.640625
+45,0.69921875
+46,0.734375
+47,0.76171875
+48,0.76171875
+49,0.796875
+51,0.6328125
+52,0.6875
+53,0.75
+54,0.75
+55,0.80078125
+56,0.67578125
+57,0.68359375
+58,0.6484375
+59,0.71484375
+61,0.74609375
+62,0.703125
+63,0.6875
+64,0.63671875
+65,0.84375
+66,0.74609375
+67,0.81640625
+68,0.7578125
+69,0.80859375
+71,0.79296875
+72,0.734375
+73,0.65625
+74,0.81640625
+75,0.8515625
+76,0.70703125
+77,0.703125
+78,0.84765625
+79,0.8203125
+81,0.75
+82,0.7109375
+83,0.77734375
+84,0.76171875
+85,0.7421875
+86,0.671875
+87,0.75390625
+88,0.65625
+89,0.8671875
+91,0.859375
+92,0.7265625
+93,0.7890625
+94,0.76171875
+95,0.69140625
+96,0.75
+97,0.84765625
+98,0.74609375
+99,0.8828125

results/validation_mean16.csv

@@ -0,0 +1,11 @@
+step,val_mean16
+10,0.58671875
+20,0.6171875
+30,0.6140625
+40,0.61796875
+50,0.6421875
+60,0.6453125
+70,0.653125
+80,0.66640625
+90,0.68046875
+100,0.6578125