初始化项目，由ModelHub XC社区提供模型

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

results/hyperparameters.csv

@@ -0,0 +1,33 @@
+section,parameter,value,source
+Run identity,Base model,Qwen/Qwen3-4B,queue/script override
+Run identity,Dataset,Tool-use / tooluse,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-tooluse-GRPO-Qwen-Qwen3-4B-mbs8-train32-rollout8-lr1e-6-vllm0.8,run_qwen3_generalization.sh
+Run identity,W&B run,run-20260702_145459-dx38nnn7,wandb
+Data,Train file,datasets/tooluse/train.parquet,script override
+Data,Validation file,datasets/tooluse/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/tooluse/qwen3gen-tooluse-GRPO-Qwen-Qwen3-4B-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/tooluse/qwen3gen-tooluse-GRPO-Qwen-Qwen3-4B-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 +1,27 @@
+{
+  "repo_id": "SeongryongJung/Qwen3-4B-Tooluse-GRPO-TR",
+  "output_dir": "/mnt/mole/SDPO/L2T/hf_upload_tr/Qwen3-4B-Tooluse-GRPO-TR",
+  "experiment": "qwen3gen-tooluse-GRPO-Qwen-Qwen3-4B-mbs8-train32-rollout8-lr1e-6-vllm0.8",
+  "best_step": 40,
+  "best_val_mean16": 0.6029411764705882,
+  "best_actor_dir": "/mnt/mole/SDPO/L2T/checkpoints/datasets/tooluse/qwen3gen-tooluse-GRPO-Qwen-Qwen3-4B-mbs8-train32-rollout8-lr1e-6-vllm0.8/_actor_archive/global_step_40/actor",
+  "final_step": 100,
+  "final_val_mean16": 0.5689338235294118,
+  "final_actor_dir": "/mnt/mole/SDPO/L2T/checkpoints/datasets/tooluse/qwen3gen-tooluse-GRPO-Qwen-Qwen3-4B-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-00002.safetensors",
+    "model-00002-of-00002.safetensors"
+  ]
+}

results/training_score.png

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:34070b4fc5932ea614982c3e664464876ddf22215fe056a1804bc20fb7fac39f
+size 136037

results/training_scores.csv

@@ -0,0 +1,91 @@
+step,critic_score_mean
+1,0.3125
+2,0.24609375
+3,0.34375
+4,0.2734375
+5,0.375
+6,0.42578125
+7,0.3828125
+8,0.3046875
+9,0.28515625
+11,0.30078125
+12,0.359375
+13,0.26953125
+14,0.25390625
+15,0.5390625
+16,0.3125
+17,0.38671875
+18,0.3046875
+19,0.23828125
+21,0.20703125
+22,0.4453125
+23,0.20703125
+24,0.2890625
+25,0.2109375
+26,0.5234375
+27,0.375
+28,0.30078125
+29,0.171875
+31,0.421875
+32,0.28125
+33,0.2265625
+34,0.359375
+35,0.30078125
+36,0.4140625
+37,0.42578125
+38,0.3828125
+39,0.42578125
+41,0.30859375
+42,0.22265625
+43,0.3125
+44,0.265625
+45,0.25390625
+46,0.30859375
+47,0.30078125
+48,0.38671875
+49,0.2109375
+51,0.37109375
+52,0.4375
+53,0.36328125
+54,0.3359375
+55,0.4140625
+56,0.36328125
+57,0.4296875
+58,0.37109375
+59,0.4140625
+61,0.3828125
+62,0.33984375
+63,0.3671875
+64,0.38671875
+65,0.30859375
+66,0.421875
+67,0.421875
+68,0.46484375
+69,0.5390625
+71,0.3046875
+72,0.38671875
+73,0.33984375
+74,0.421875
+75,0.35546875
+76,0.25
+77,0.30078125
+78,0.375
+79,0.4140625
+81,0.40625
+82,0.44140625
+83,0.25390625
+84,0.46875
+85,0.390625
+86,0.40234375
+87,0.4375
+88,0.53515625
+89,0.30859375
+91,0.47265625
+92,0.41015625
+93,0.32421875
+94,0.3125
+95,0.30859375
+96,0.42578125
+97,0.33984375
+98,0.359375
+99,0.3828125

results/validation_mean16.csv

@@ -0,0 +1,11 @@
+step,val_mean16
+10,0.5965073529411765
+20,0.5790441176470589
+30,0.5909926470588235
+40,0.6029411764705882
+50,0.5946691176470589
+60,0.5698529411764706
+70,0.5496323529411765
+80,0.5597426470588235
+90,0.5643382352941176
+100,0.5689338235294118