Model save

README.md

@@ -1,5 +1,5 @@
 ---
-base_model: Qwen/Qwen3-4B
+base_model: Qwen/Qwen2.5-7B-Instruct
 library_name: transformers
 model_name: PreThink_MemAgent
 tags:
@@ -11,7 +11,7 @@ licence: license
 
 # Model Card for PreThink_MemAgent
 
-This model is a fine-tuned version of [Qwen/Qwen3-4B](https://huggingface.co/Qwen/Qwen3-4B).
+This model is a fine-tuned version of [Qwen/Qwen2.5-7B-Instruct](https://huggingface.co/Qwen/Qwen2.5-7B-Instruct).
 It has been trained using [TRL](https://github.com/huggingface/trl).
 
 ## Quick start
@@ -27,7 +27,7 @@ print(output["generated_text"])
 
 ## Training procedure
 
-[<img src="https://raw.githubusercontent.com/wandb/assets/main/wandb-github-badge-28.svg" alt="Visualize in Weights & Biases" width="150" height="24"/>](https://wandb.ai/mingzeli/PreThink_MemAgent/runs/pato1fcq) 
+[<img src="https://raw.githubusercontent.com/wandb/assets/main/wandb-github-badge-28.svg" alt="Visualize in Weights & Biases" width="150" height="24"/>](https://wandb.ai/mingzeli/PreThink_MemAgent/runs/bvl983ap) 
 
 
 This model was trained with SFT.

all_results.json

@@ -1,8 +1,8 @@
 {
-    "total_flos": 284515814506496.0,
-    "train_loss": 1.184776012692155,
-    "train_runtime": 2955.0783,
+    "total_flos": 1933127778631680.0,
+    "train_loss": 0.6745891945267265,
+    "train_runtime": 14540.5824,
     "train_samples": 26745,
-    "train_samples_per_second": 9.051,
-    "train_steps_per_second": 0.141
+    "train_samples_per_second": 7.357,
+    "train_steps_per_second": 0.115
 }

generation_config.json

@@ -3,8 +3,9 @@
   "do_sample": true,
   "eos_token_id": 151645,
   "pad_token_id": 151643,
-  "temperature": 0.6,
+  "repetition_penalty": 1.05,
+  "temperature": 0.7,
   "top_k": 20,
-  "top_p": 0.95,
+  "top_p": 0.8,
   "transformers_version": "4.52.3"
 }

train_results.json

@@ -1,8 +1,8 @@
 {
-    "total_flos": 284515814506496.0,
-    "train_loss": 1.184776012692155,
-    "train_runtime": 2955.0783,
+    "total_flos": 1933127778631680.0,
+    "train_loss": 0.6745891945267265,
+    "train_runtime": 14540.5824,
     "train_samples": 26745,
-    "train_samples_per_second": 9.051,
-    "train_steps_per_second": 0.141
+    "train_samples_per_second": 7.357,
+    "train_steps_per_second": 0.115
 }