End of training

README.md

@@ -2,9 +2,10 @@
 license: llama2
 library_name: peft
 tags:
+- llama-factory
+- lora
 - trl
 - dpo
-- llama-factory
 - generated_from_trainer
 base_model: lmsys/vicuna-7b-v1.5
 model-index:
@@ -17,7 +18,7 @@ should probably proofread and complete it, then remove this comment. -->
 
 # Vicuna-7B-v1.5-ORPO-SALT
 
-This model is a fine-tuned version of [lmsys/vicuna-7b-v1.5](https://huggingface.co/lmsys/vicuna-7b-v1.5) on the None dataset.
+This model is a fine-tuned version of [lmsys/vicuna-7b-v1.5](https://huggingface.co/lmsys/vicuna-7b-v1.5) on the dpo_mix_en and the bct_non_cot_dpo_1000 datasets.
 It achieves the following results on the evaluation set:
 - Loss: 0.9497
 - Rewards/chosen: -0.0879

all_results.json

@@ -0,0 +1,22 @@
+{
+    "epoch": 2.9969690846635686,
+    "eval_logits/chosen": -0.3775150775909424,
+    "eval_logits/rejected": -0.3581116795539856,
+    "eval_logps/chosen": -0.8786855936050415,
+    "eval_logps/rejected": -0.9948004484176636,
+    "eval_loss": 0.9497246742248535,
+    "eval_odds_ratio_loss": 0.7103896737098694,
+    "eval_rewards/accuracies": 0.5163636207580566,
+    "eval_rewards/chosen": -0.0878685712814331,
+    "eval_rewards/margins": 0.011611479334533215,
+    "eval_rewards/rejected": -0.0994800478219986,
+    "eval_runtime": 192.2901,
+    "eval_samples_per_second": 5.721,
+    "eval_sft_loss": 0.8786855936050415,
+    "eval_steps_per_second": 2.86,
+    "total_flos": 1.9948570754930442e+18,
+    "train_loss": 0.9750770799807618,
+    "train_runtime": 17949.5667,
+    "train_samples_per_second": 1.654,
+    "train_steps_per_second": 0.103
+}

eval_results.json

@@ -0,0 +1,17 @@
+{
+    "epoch": 2.9969690846635686,
+    "eval_logits/chosen": -0.3775150775909424,
+    "eval_logits/rejected": -0.3581116795539856,
+    "eval_logps/chosen": -0.8786855936050415,
+    "eval_logps/rejected": -0.9948004484176636,
+    "eval_loss": 0.9497246742248535,
+    "eval_odds_ratio_loss": 0.7103896737098694,
+    "eval_rewards/accuracies": 0.5163636207580566,
+    "eval_rewards/chosen": -0.0878685712814331,
+    "eval_rewards/margins": 0.011611479334533215,
+    "eval_rewards/rejected": -0.0994800478219986,
+    "eval_runtime": 192.2901,
+    "eval_samples_per_second": 5.721,
+    "eval_sft_loss": 0.8786855936050415,
+    "eval_steps_per_second": 2.86
+}

train_results.json

@@ -0,0 +1,8 @@
+{
+    "epoch": 2.9969690846635686,
+    "total_flos": 1.9948570754930442e+18,
+    "train_loss": 0.9750770799807618,
+    "train_runtime": 17949.5667,
+    "train_samples_per_second": 1.654,
+    "train_steps_per_second": 0.103
+}