初始化项目，由ModelHub XC社区提供模型

Model: SeongryongJung/Qwen3-4B-Biology-GRPO-TR
Source: Original Platform

results/export_note.csv

@@ -0,0 +1,7 @@
+key,value
+root_checkpoint_step,100
+root_checkpoint_kind,final
+best_validation_step,90
+best_validation_mean16,0.467500000000
+best_actor_available,false
+reason,global_step_90 has validation metadata only; no actor model shards are present

results/hyperparameters.csv

@@ -0,0 +1,77 @@
+section,parameter,value,source
+Run identity,Base model,Qwen/Qwen3-4B,queue/script override
+Run identity,Dataset,Biology / SciKnowEval biology,run_qwen3_generalization.sh
+Run identity,Method,GRPO,run_qwen3_generalization.sh
+Run identity,Config,baseline_grpo,run_qwen3_generalization.sh
+Run identity,Experiment,qwen3gen-biology-GRPO-Qwen-Qwen3-4B-mbs8-train32-rollout8-lr1e-6-vllm0.8,run_qwen3_generalization.sh
+Run identity,W&B run,run-20260702_102415-94h2xs6b,wandb
+Data,Train file,datasets/sciknoweval/biology/train.parquet,script override
+Data,Validation file,datasets/sciknoweval/biology/test.parquet,script override
+Data,Train batch size,32,queue/script override
+Data,Train max samples,3200,queue/script override
+Data,Prompt key,prompt,legacy_data.yaml default
+Data,Reward key,data_source,legacy_data.yaml default
+Data,Shuffle train data,True,user.yaml / legacy_data.yaml
+Data,Validation shuffle,False,legacy_data.yaml default
+Data,Filter overlong prompts,True,user.yaml
+Data,Prompt truncation,error,legacy_data.yaml default
+Data,enable_thinking,false,script override
+Schedule,Total training steps,100,queue/script override
+Schedule,Total epochs,30,ppo_trainer/user.yaml default
+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
+Sequence,Actor max token length per GPU,10240,queue/script override
+Rollout,Rollout engine,vllm,user.yaml
+Rollout,Rollout dtype,bfloat16,rollout.yaml default
+Rollout,Train rollout n,8,queue/script override
+Rollout,Train rollout temperature,1.0,script override
+Rollout,Train rollout top_p,1.0,script override
+Rollout,Train rollout do_sample,True,rollout.yaml default
+Rollout,Calculate rollout log probs,True,baseline_grpo.yaml / script override
+Rollout,Max num batched tokens,10240,queue/script override
+Rollout,vLLM GPU memory utilization,0.8,queue/script override
+Rollout,Tensor model parallel size,2,rollout.yaml default
+Rollout,Free cache engine,True,rollout.yaml default
+Validation,Validation rollout n,16,queue/script override
+Validation,Validation temperature,0.6,queue/script override
+Validation,Validation top_p,0.95,queue/script override
+Validation,Validation do_sample,True,queue/script override
+Optimization,Optimizer,AdamW,fsdp optimizer config
+Optimization,Learning rate,1e-6,GRPO method override
+Optimization,LR scheduler,constant,W&B config
+Optimization,LR warmup steps,10,script override
+Optimization,Weight decay,0.01,script override
+Optimization,Betas,"(0.9, 0.999)",W&B config
+Optimization,Gradient clip,1.0,script override
+PPO/GRPO,Policy loss mode,vanilla,method override
+PPO/GRPO,Advantage estimator,grpo,baseline_grpo.yaml
+PPO/GRPO,Normalize GRPO advantages by std,False,baseline_grpo.yaml / script override
+PPO/GRPO,PPO epochs,1,W&B config
+PPO/GRPO,PPO mini batch size,8,queue/script override
+PPO/GRPO,PPO micro batch size per GPU,1,user.yaml
+PPO/GRPO,Clip ratio low,0.2,script override
+PPO/GRPO,Clip ratio high,0.28,script override
+PPO/GRPO,Gamma,1.0,ppo_trainer.yaml default
+PPO/GRPO,Lambda,1.0,ppo_trainer.yaml default
+PPO/GRPO,Actor KL loss coef,0.0,method override
+PPO/GRPO,Use KL in reward,False,ppo_trainer/user.yaml
+Rollout correction,Importance sampling mode,token,script override
+Rollout correction,IS threshold,2.0,script override
+FSDP/System,Actor strategy,fsdp,dp_actor.yaml
+FSDP/System,FSDP dtype,bfloat16,W&B config
+FSDP/System,FSDP model dtype,fp32,W&B config
+FSDP/System,Use torch compile,True,W&B config
+FSDP/System,GPUs per node,8,queue/script override
+FSDP/System,Nodes,1,user.yaml
+FSDP/System,GPU type,NVIDIA H200,wandb-metadata
+Checkpoint/Logging,Checkpoint root,checkpoints/datasets/sciknoweval/biology,script override
+Checkpoint/Logging,Latest checkpointed iteration,100,latest_checkpointed_iteration.txt
+Checkpoint/Logging,Max actor checkpoints to keep,1,user.yaml
+Checkpoint/Logging,Logger,"console, wandb",ppo_trainer.yaml
+Checkpoint/Logging,W&B entity,seongryongjung-chung-ang-university,environment
+Checkpoint/Logging,W&B project,SDPO-root,user.yaml project_name
+Checkpoint/Logging,W&B group,QWEN3-GRPO-generalization,method override

results/summary.json

@@ -0,0 +1,27 @@
+{
+  "repo_id": "SeongryongJung/Qwen3-4B-Biology-GRPO-TR",
+  "output_dir": "/mnt/mole/SDPO/L2T/hf_upload_tr/Qwen3-4B-Biology-GRPO-TR",
+  "root_checkpoint_step": 100,
+  "root_checkpoint_kind": "final",
+  "best_step": 90,
+  "best_val_mean16": 0.4675,
+  "best_actor_available": false,
+  "final_step": 100,
+  "final_val_mean16": 0.42625,
+  "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-00002.safetensors",
+    "model-00002-of-00002.safetensors"
+  ]
+}

results/training_score.png

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b2e33136ae641dfe1dee31897c5d1e3f4a15f89df7eb402f053cfd5cda8b5f14
+size 154942

results/training_scores.csv

@@ -0,0 +1,91 @@
+step,critic_score_mean
+1,0.3515625
+2,0.36328125
+3,0.37890625
+4,0.4140625
+5,0.4609375
+6,0.32421875
+7,0.29296875
+8,0.26953125
+9,0.3984375
+11,0.44140625
+12,0.45703125
+13,0.35546875
+14,0.26953125
+15,0.34765625
+16,0.546875
+17,0.28515625
+18,0.3359375
+19,0.40625
+21,0.40625
+22,0.49609375
+23,0.25390625
+24,0.36328125
+25,0.3828125
+26,0.41015625
+27,0.296875
+28,0.34765625
+29,0.296875
+31,0.375
+32,0.5078125
+33,0.453125
+34,0.3828125
+35,0.35546875
+36,0.34375
+37,0.359375
+38,0.43359375
+39,0.3828125
+41,0.48828125
+42,0.328125
+43,0.44140625
+44,0.4375
+45,0.31640625
+46,0.46484375
+47,0.39453125
+48,0.35546875
+49,0.34765625
+51,0.4140625
+52,0.42578125
+53,0.51953125
+54,0.48046875
+55,0.51171875
+56,0.42578125
+57,0.44140625
+58,0.59375
+59,0.5390625
+61,0.28125
+62,0.46484375
+63,0.39453125
+64,0.4921875
+65,0.4375
+66,0.4296875
+67,0.52734375
+68,0.47265625
+69,0.42578125
+71,0.44921875
+72,0.5
+73,0.66015625
+74,0.42578125
+75,0.3828125
+76,0.43359375
+77,0.6171875
+78,0.43359375
+79,0.41015625
+81,0.5390625
+82,0.59375
+83,0.5234375
+84,0.4296875
+85,0.5
+86,0.63671875
+87,0.5234375
+88,0.375
+89,0.58984375
+91,0.53515625
+92,0.65234375
+93,0.484375
+94,0.48046875
+95,0.53125
+96,0.61328125
+97,0.56640625
+98,0.4140625
+99,0.578125

results/validation_mean16.csv

@@ -0,0 +1,11 @@
+step,val_mean16
+10,0.32625
+20,0.3575
+30,0.3825
+40,0.35375
+50,0.41875
+60,0.3825
+70,0.395
+80,0.41625
+90,0.4675
+100,0.42625