Model save

README.md

@@ -1,4 +1,5 @@
 ---
+base_model: Qwen/Qwen3-4B
 library_name: transformers
 model_name: PreThink_MemAgent
 tags:
@@ -10,7 +11,7 @@ licence: license
 
 # Model Card for PreThink_MemAgent
 
-This model is a fine-tuned version of [None](https://huggingface.co/None).
+This model is a fine-tuned version of [Qwen/Qwen3-4B](https://huggingface.co/Qwen/Qwen3-4B).
 It has been trained using [TRL](https://github.com/huggingface/trl).
 
 ## Quick start
@@ -26,7 +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/9q53vlm5) 
+[<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/qe6p1l0a) 
 
 
 This model was trained with SFT.

all_results.json

@@ -1,8 +1,8 @@
 {
-    "total_flos": 3.064163325664297e+17,
-    "train_loss": 0.4129233885510468,
-    "train_runtime": 3035.4475,
-    "train_samples": 27456,
-    "train_samples_per_second": 27.135,
-    "train_steps_per_second": 0.107
+    "total_flos": 967966479089664.0,
+    "train_loss": 1.0463442883413772,
+    "train_runtime": 10114.1342,
+    "train_samples": 47071,
+    "train_samples_per_second": 13.962,
+    "train_steps_per_second": 0.055
 }

generation_config.json

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

train_results.json

@@ -1,8 +1,8 @@
 {
-    "total_flos": 3.064163325664297e+17,
-    "train_loss": 0.4129233885510468,
-    "train_runtime": 3035.4475,
-    "train_samples": 27456,
-    "train_samples_per_second": 27.135,
-    "train_steps_per_second": 0.107
+    "total_flos": 967966479089664.0,
+    "train_loss": 1.0463442883413772,
+    "train_runtime": 10114.1342,
+    "train_samples": 47071,
+    "train_samples_per_second": 13.962,
+    "train_steps_per_second": 0.055
 }