初始化项目，由ModelHub XC社区提供模型

Model: SeongryongJung/Qwen3-8B-Chemistry-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,Chemistry / SciKnowEval chemistry,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-chemistry-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_014331-zs3h1j56,wandb
+Data,Train file,datasets/sciknoweval/chemistry/train.parquet,script override
+Data,Validation file,datasets/sciknoweval/chemistry/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/chemistry/qwen3gen-chemistry-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/chemistry/qwen3gen-chemistry-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-Chemistry-RLSD-TR",
+  "output_dir": "/mnt/mole/SDPO/L2T/hf_upload_tr/Qwen3-8B-Chemistry-RLSD-TR",
+  "experiment": "qwen3gen-chemistry-RLSD_TR-Qwen-Qwen3-8B-mbs8-decay0-tr0.1-train32-rollout8-lr1e-6-vllm0.8",
+  "best_step": 80,
+  "best_val_mean16": 0.7098214285714286,
+  "best_actor_dir": "/mnt/mole/SDPO/L2T/checkpoints/datasets/sciknoweval/chemistry/qwen3gen-chemistry-RLSD_TR-Qwen-Qwen3-8B-mbs8-decay0-tr0.1-train32-rollout8-lr1e-6-vllm0.8/global_step_80/actor",
+  "final_step": 100,
+  "final_val_mean16": 0.7065476190476191,
+  "final_actor_dir": "/mnt/mole/SDPO/L2T/checkpoints/datasets/sciknoweval/chemistry/qwen3gen-chemistry-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 sha256:99a2c52c9b441b1a71a06bbe04b74edcf2dc01c705cf0659e99e2fccc79e8062
+size 129165

results/training_scores.csv

@@ -0,0 +1,91 @@
+step,critic_score_mean
+1,0.46484375
+2,0.3671875
+3,0.4140625
+4,0.4453125
+5,0.42578125
+6,0.421875
+7,0.33203125
+8,0.46875
+9,0.359375
+11,0.48046875
+12,0.46484375
+13,0.375
+14,0.40625
+15,0.46875
+16,0.47265625
+17,0.41015625
+18,0.49609375
+19,0.5
+21,0.43359375
+22,0.4375
+23,0.69921875
+24,0.58984375
+25,0.5546875
+26,0.6328125
+27,0.51171875
+28,0.59765625
+29,0.48046875
+31,0.6015625
+32,0.55078125
+33,0.60546875
+34,0.59765625
+35,0.5078125
+36,0.5625
+37,0.54296875
+38,0.6484375
+39,0.6015625
+41,0.62890625
+42,0.6171875
+43,0.61328125
+44,0.66796875
+45,0.5390625
+46,0.5546875
+47,0.64453125
+48,0.671875
+49,0.640625
+51,0.5546875
+52,0.61328125
+53,0.62109375
+54,0.5546875
+55,0.640625
+56,0.6484375
+57,0.66796875
+58,0.703125
+59,0.5390625
+61,0.66015625
+62,0.62890625
+63,0.53125
+64,0.765625
+65,0.71484375
+66,0.6171875
+67,0.6796875
+68,0.46875
+69,0.6796875
+71,0.58984375
+72,0.671875
+73,0.640625
+74,0.703125
+75,0.71484375
+76,0.6171875
+77,0.7421875
+78,0.68359375
+79,0.609375
+81,0.63671875
+82,0.6015625
+83,0.49609375
+84,0.6875
+85,0.71484375
+86,0.609375
+87,0.62109375
+88,0.58984375
+89,0.59765625
+91,0.640625
+92,0.609375
+93,0.83203125
+94,0.76171875
+95,0.68359375
+96,0.67578125
+97,0.5625
+98,0.609375
+99,0.578125

results/validation_mean16.csv

@@ -0,0 +1,11 @@
+step,val_mean16
+10,0.4217261904761905
+20,0.5196428571428572
+30,0.606845238095238
+40,0.6455357142857143
+50,0.66875
+60,0.6875
+70,0.6967261904761904
+80,0.7098214285714286
+90,0.7092261904761905
+100,0.7065476190476191