End of training

README.md

@@ -3,10 +3,12 @@ library_name: peft
 license: llama3.1
 base_model: meta-llama/Meta-Llama-3.1-8B-Instruct
 tags:
+- alignment-handbook
 - trl
 - sft
-- alignment-handbook
 - generated_from_trainer
+datasets:
+- tttx/fake_dataset_prompt_3072_response_15360_data_size_1000
 model-index:
 - name: dummy_lora_ft_3k_1k
   results: []
@@ -17,7 +19,7 @@ should probably proofread and complete it, then remove this comment. -->
 
 # dummy_lora_ft_3k_1k
 
-This model is a fine-tuned version of [meta-llama/Meta-Llama-3.1-8B-Instruct](https://huggingface.co/meta-llama/Meta-Llama-3.1-8B-Instruct) on an unknown dataset.
+This model is a fine-tuned version of [meta-llama/Meta-Llama-3.1-8B-Instruct](https://huggingface.co/meta-llama/Meta-Llama-3.1-8B-Instruct) on the tttx/fake_dataset_prompt_3072_response_15360_data_size_1000 dataset.
 It achieves the following results on the evaluation set:
 - Loss: 12.0219
 

all_results.json

@@ -1,5 +1,10 @@
 {
     "epoch": 1.0,
+    "eval_loss": 12.021870613098145,
+    "eval_runtime": 2.9731,
+    "eval_samples": 1,
+    "eval_samples_per_second": 0.336,
+    "eval_steps_per_second": 0.336,
     "total_flos": 54979878453248.0,
     "train_loss": 12.068599019731794,
     "train_runtime": 98.6323,

eval_results.json

@@ -1,8 +1,8 @@
 {
     "epoch": 1.0,
-    "eval_loss": NaN,
-    "eval_runtime": 7.1546,
+    "eval_loss": 12.021870613098145,
+    "eval_runtime": 2.9731,
     "eval_samples": 1,
-    "eval_samples_per_second": 0.14,
-    "eval_steps_per_second": 0.14
+    "eval_samples_per_second": 0.336,
+    "eval_steps_per_second": 0.336
 }