初始化项目，由ModelHub XC社区提供模型

Model: Phantomcloak19/qwen2.5-3b-dpo
Source: Original Platform

lora_adapter/checkpoint-42/trainer_state.json

@@ -0,0 +1,70 @@
+{
+  "best_global_step": null,
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 2.0,
+  "eval_steps": 100,
+  "global_step": 42,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "entropy": 2.308793434500694,
+      "epoch": 0.963855421686747,
+      "grad_norm": 1.8339245319366455,
+      "learning_rate": 1.9000000000000002e-06,
+      "logits/chosen": 0.41484330565815475,
+      "logits/rejected": 0.5456838045995205,
+      "logps/chosen": -49.526001167297366,
+      "logps/rejected": -59.49296364784241,
+      "loss": 0.6891121387481689,
+      "mean_token_accuracy": 0.22993013681843877,
+      "num_tokens": 37803.0,
+      "rewards/accuracies": 0.759375,
+      "rewards/chosen": 7.2634290063433585e-06,
+      "rewards/margins": 0.008121937591931782,
+      "rewards/rejected": -0.008114674211174133,
+      "step": 20
+    },
+    {
+      "entropy": 2.277806546114668,
+      "epoch": 1.9156626506024095,
+      "grad_norm": 2.0258429050445557,
+      "learning_rate": 3.900000000000001e-06,
+      "logits/chosen": 0.4389253231148894,
+      "logits/rejected": 0.5264429102267841,
+      "logps/chosen": -48.951683044433594,
+      "logps/rejected": -61.126885426195365,
+      "loss": 0.61369309425354,
+      "mean_token_accuracy": 0.22803395109463342,
+      "num_tokens": 74911.0,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.004455907024219161,
+      "rewards/margins": 0.17089249783113034,
+      "rewards/rejected": -0.1664365911031071,
+      "step": 40
+    }
+  ],
+  "logging_steps": 20,
+  "max_steps": 63,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 3,
+  "save_steps": 500,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": false
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 1537373159227392.0,
+  "train_batch_size": 4,
+  "trial_name": null,
+  "trial_params": null
+}