初始化项目，由ModelHub XC社区提供模型

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

results/hyperparameters.csv

@@ -0,0 +1,33 @@
+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,GRPO,run_qwen3_generalization.sh
+Run identity,Config,baseline_grpo,run_qwen3_generalization.sh
+Run identity,Experiment,qwen3gen-physics-GRPO-Qwen-Qwen3-8B-mbs8-train32-rollout8-lr1e-6-vllm0.8,run_qwen3_generalization.sh
+Run identity,W&B run,run-20260703_032040-6ig3l55l,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,GRPO 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-GRPO-Qwen-Qwen3-8B-mbs8-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-GRPO-Qwen-Qwen3-8B-mbs8-train32-rollout8-lr1e-6-vllm0.8/_actor_archive,preserve_actor_checkpoints.py
+Checkpoint/Logging,Logger,"console, wandb",ppo_trainer.yaml
+PPO/GRPO,Policy loss mode,vanilla,method override
+PPO/GRPO,Actor KL loss coef,0.0,method override

results/summary.json

@@ -0,0 +1,29 @@
+{
+  "repo_id": "SeongryongJung/Qwen3-8B-Physics-GRPO-TR",
+  "output_dir": "/mnt/mole/SDPO/L2T/hf_upload_tr/Qwen3-8B-Physics-GRPO-TR",
+  "experiment": "qwen3gen-physics-GRPO-Qwen-Qwen3-8B-mbs8-train32-rollout8-lr1e-6-vllm0.8",
+  "best_step": 100,
+  "best_val_mean16": 0.7296875,
+  "best_actor_dir": "/mnt/mole/SDPO/L2T/checkpoints/datasets/sciknoweval/physics/qwen3gen-physics-GRPO-Qwen-Qwen3-8B-mbs8-train32-rollout8-lr1e-6-vllm0.8/global_step_100/actor",
+  "final_step": 100,
+  "final_val_mean16": 0.7296875,
+  "final_actor_dir": "/mnt/mole/SDPO/L2T/checkpoints/datasets/sciknoweval/physics/qwen3gen-physics-GRPO-Qwen-Qwen3-8B-mbs8-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 sha256:453f8ccb73768a3b3ae4b970f7f0b84df4b8d760e9ac177c35f17270c51db265
+size 129239

results/training_scores.csv

@@ -0,0 +1,91 @@
+step,critic_score_mean
+1,0.6015625
+2,0.5859375
+3,0.52734375
+4,0.63671875
+5,0.73046875
+6,0.65625
+7,0.75390625
+8,0.7265625
+9,0.58984375
+11,0.62890625
+12,0.5234375
+13,0.5703125
+14,0.64453125
+15,0.69140625
+16,0.765625
+17,0.70703125
+18,0.41015625
+19,0.71875
+21,0.6640625
+22,0.71484375
+23,0.5
+24,0.74609375
+25,0.71875
+26,0.7109375
+27,0.796875
+28,0.6171875
+29,0.70703125
+31,0.62890625
+32,0.73828125
+33,0.70703125
+34,0.6796875
+35,0.75390625
+36,0.578125
+37,0.6640625
+38,0.66015625
+39,0.87890625
+41,0.70703125
+42,0.6171875
+43,0.74609375
+44,0.66796875
+45,0.66796875
+46,0.7890625
+47,0.75
+48,0.765625
+49,0.859375
+51,0.7109375
+52,0.71484375
+53,0.73828125
+54,0.76171875
+55,0.84765625
+56,0.6484375
+57,0.734375
+58,0.64453125
+59,0.71875
+61,0.72265625
+62,0.65625
+63,0.7109375
+64,0.6875
+65,0.8671875
+66,0.765625
+67,0.8046875
+68,0.734375
+69,0.80859375
+71,0.82421875
+72,0.76953125
+73,0.66015625
+74,0.82421875
+75,0.85546875
+76,0.6953125
+77,0.72265625
+78,0.80859375
+79,0.82421875
+81,0.734375
+82,0.76953125
+83,0.6953125
+84,0.68359375
+85,0.73828125
+86,0.77734375
+87,0.76953125
+88,0.72265625
+89,0.7421875
+91,0.81640625
+92,0.73046875
+93,0.75
+94,0.8203125
+95,0.6328125
+96,0.7421875
+97,0.734375
+98,0.79296875
+99,0.84765625

results/validation_mean16.csv

@@ -0,0 +1,11 @@
+step,val_mean16
+10,0.58359375
+20,0.59921875
+30,0.6046875
+40,0.62109375
+50,0.65390625
+60,0.671875
+70,0.68125
+80,0.7125
+90,0.72265625
+100,0.7296875