End of training

README.md

@@ -18,9 +18,9 @@ should probably proofread and complete it, then remove this comment. -->
 
 # train_copa_42_1760623607
 
-This model is a fine-tuned version of [meta-llama/Meta-Llama-3-8B-Instruct](https://huggingface.co/meta-llama/Meta-Llama-3-8B-Instruct) on an unknown dataset.
+This model is a fine-tuned version of [meta-llama/Meta-Llama-3-8B-Instruct](https://huggingface.co/meta-llama/Meta-Llama-3-8B-Instruct) on the copa dataset.
 It achieves the following results on the evaluation set:
-- Loss: 0.1015
+- Loss: 0.0320
 - Num Input Tokens Seen: 564096
 
 ## Model description

all_results.json

@@ -0,0 +1,13 @@
+{
+    "epoch": 20.0,
+    "eval_loss": 0.031980086117982864,
+    "eval_runtime": 0.628,
+    "eval_samples_per_second": 63.695,
+    "eval_steps_per_second": 15.924,
+    "num_input_tokens_seen": 564096,
+    "total_flos": 2.5450673718951936e+16,
+    "train_loss": 0.018261103643923585,
+    "train_runtime": 415.2132,
+    "train_samples_per_second": 17.34,
+    "train_steps_per_second": 4.335
+}

eval_results.json

@@ -0,0 +1,8 @@
+{
+    "epoch": 20.0,
+    "eval_loss": 0.031980086117982864,
+    "eval_runtime": 0.628,
+    "eval_samples_per_second": 63.695,
+    "eval_steps_per_second": 15.924,
+    "num_input_tokens_seen": 564096
+}

train_results.json

@@ -0,0 +1,9 @@
+{
+    "epoch": 20.0,
+    "num_input_tokens_seen": 564096,
+    "total_flos": 2.5450673718951936e+16,
+    "train_loss": 0.018261103643923585,
+    "train_runtime": 415.2132,
+    "train_samples_per_second": 17.34,
+    "train_steps_per_second": 4.335
+}