Model save

README.md

@@ -26,7 +26,7 @@ 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/6pz32z57) 
+[<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/mbsja32c) 
 
 
 This model was trained with SFT.

all_results.json

@@ -6,8 +6,8 @@
     "eval_steps_per_second": 5.352,
     "total_flos": 21958899171328.0,
     "train_loss": 0.0,
-    "train_runtime": 1.4645,
+    "train_runtime": 1.547,
     "train_samples": 821,
-    "train_samples_per_second": 1681.768,
-    "train_steps_per_second": 26.63
+    "train_samples_per_second": 1592.111,
+    "train_steps_per_second": 25.21
 }

train_results.json

@@ -1,8 +1,8 @@
 {
     "total_flos": 21958899171328.0,
     "train_loss": 0.0,
-    "train_runtime": 1.4645,
+    "train_runtime": 1.547,
     "train_samples": 821,
-    "train_samples_per_second": 1681.768,
-    "train_steps_per_second": 26.63
+    "train_samples_per_second": 1592.111,
+    "train_steps_per_second": 25.21
 }

trainer_state.json

@@ -353,9 +353,9 @@
       "step": 39,
       "total_flos": 21958899171328.0,
       "train_loss": 0.0,
-      "train_runtime": 1.4645,
-      "train_samples_per_second": 1681.768,
-      "train_steps_per_second": 26.63
+      "train_runtime": 1.547,
+      "train_samples_per_second": 1592.111,
+      "train_steps_per_second": 25.21
     }
   ],
   "logging_steps": 1,

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:7abbf6d93c48bb7a1e3846a182c4aed4be2d9d92eb50729f609e3a287e235b19
+oid sha256:d7eab563f645e08376b4e6e9b145632772a67fafadbd321da11aa3ed60d51052
 size 7889