End of training

README.md

@@ -2,9 +2,10 @@
 license: apache-2.0
 library_name: peft
 tags:
+- llama-factory
+- lora
 - trl
 - dpo
-- llama-factory
 - generated_from_trainer
 base_model: mistralai/Mistral-7B-Instruct-v0.3
 model-index:
@@ -17,7 +18,7 @@ should probably proofread and complete it, then remove this comment. -->
 
 # Mistral-7B-Instruct-v0.3-ORPO
 
-This model is a fine-tuned version of [mistralai/Mistral-7B-Instruct-v0.3](https://huggingface.co/mistralai/Mistral-7B-Instruct-v0.3) on an unknown dataset.
+This model is a fine-tuned version of [mistralai/Mistral-7B-Instruct-v0.3](https://huggingface.co/mistralai/Mistral-7B-Instruct-v0.3) on the dpo_mix_en dataset.
 It achieves the following results on the evaluation set:
 - Loss: 0.8734
 - Rewards/chosen: -0.0810

all_results.json

@@ -0,0 +1,22 @@
+{
+    "epoch": 2.997999555456768,
+    "eval_logits/chosen": -3.1171207427978516,
+    "eval_logits/rejected": -3.145516872406006,
+    "eval_logps/chosen": -0.8098240494728088,
+    "eval_logps/rejected": -1.0174543857574463,
+    "eval_loss": 0.8734214901924133,
+    "eval_odds_ratio_loss": 0.6359757781028748,
+    "eval_rewards/accuracies": 0.5720000267028809,
+    "eval_rewards/chosen": -0.08098240196704865,
+    "eval_rewards/margins": 0.02076304331421852,
+    "eval_rewards/rejected": -0.10174543410539627,
+    "eval_runtime": 189.0411,
+    "eval_samples_per_second": 5.29,
+    "eval_sft_loss": 0.8098240494728088,
+    "eval_steps_per_second": 2.645,
+    "total_flos": 1.9814178520144282e+18,
+    "train_loss": 0.8985705958283811,
+    "train_runtime": 25618.6466,
+    "train_samples_per_second": 1.054,
+    "train_steps_per_second": 0.066
+}

eval_results.json

@@ -0,0 +1,17 @@
+{
+    "epoch": 2.997999555456768,
+    "eval_logits/chosen": -3.1171207427978516,
+    "eval_logits/rejected": -3.145516872406006,
+    "eval_logps/chosen": -0.8098240494728088,
+    "eval_logps/rejected": -1.0174543857574463,
+    "eval_loss": 0.8734214901924133,
+    "eval_odds_ratio_loss": 0.6359757781028748,
+    "eval_rewards/accuracies": 0.5720000267028809,
+    "eval_rewards/chosen": -0.08098240196704865,
+    "eval_rewards/margins": 0.02076304331421852,
+    "eval_rewards/rejected": -0.10174543410539627,
+    "eval_runtime": 189.0411,
+    "eval_samples_per_second": 5.29,
+    "eval_sft_loss": 0.8098240494728088,
+    "eval_steps_per_second": 2.645
+}

train_results.json

@@ -0,0 +1,8 @@
+{
+    "epoch": 2.997999555456768,
+    "total_flos": 1.9814178520144282e+18,
+    "train_loss": 0.8985705958283811,
+    "train_runtime": 25618.6466,
+    "train_samples_per_second": 1.054,
+    "train_steps_per_second": 0.066
+}