Model save

README.md

@@ -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/jjh233/huggingface/runs/l9e79305) 
+[<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/jjh233/huggingface/runs/21iwqcw5) 
 
 
 This model was trained with SFT.

all_results.json

@@ -5,9 +5,9 @@
     "eval_samples_per_second": 4.944,
     "eval_steps_per_second": 1.265,
     "total_flos": 65635690217472.0,
-    "train_loss": 0.7333491018714284,
-    "train_runtime": 2770.9424,
-    "train_samples": 16610,
-    "train_samples_per_second": 31.21,
-    "train_steps_per_second": 0.244
+    "train_loss": 0.0,
+    "train_runtime": 0.9025,
+    "train_samples": 93733,
+    "train_samples_per_second": 38008.288,
+    "train_steps_per_second": 296.966
 }

train_results.json

@@ -1,8 +1,8 @@
 {
     "total_flos": 65635690217472.0,
-    "train_loss": 0.7333491018714284,
-    "train_runtime": 2770.9424,
-    "train_samples": 16610,
-    "train_samples_per_second": 31.21,
-    "train_steps_per_second": 0.244
+    "train_loss": 0.0,
+    "train_runtime": 0.9025,
+    "train_samples": 93733,
+    "train_samples_per_second": 38008.288,
+    "train_steps_per_second": 296.966
 }

trainer_state.json

@@ -2197,14 +2197,14 @@
       "epoch": 1.0,
       "step": 676,
       "total_flos": 65635690217472.0,
-      "train_loss": 0.7333491018714284,
-      "train_runtime": 2770.9424,
-      "train_samples_per_second": 31.21,
-      "train_steps_per_second": 0.244
+      "train_loss": 0.0,
+      "train_runtime": 0.9025,
+      "train_samples_per_second": 38008.288,
+      "train_steps_per_second": 296.966
     }
   ],
   "logging_steps": 5,
-  "max_steps": 676,
+  "max_steps": 268,
   "num_input_tokens_seen": 0,
   "num_train_epochs": 1,
   "save_steps": 100,

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:72587138b5bb06ff5f7542be9ffb8468fbbacb0edfdd36ce6dc92498595c3cdc
-size 7416
+oid sha256:0ffb7cda303708d9e1ceff69a1d310dcccc3fdd78818229bcf01013473d18deb
+size 7480