End of training

README.md

@@ -2,9 +2,10 @@
 license: gemma
 library_name: peft
 tags:
+- llama-factory
+- lora
 - trl
 - dpo
-- llama-factory
 - generated_from_trainer
 base_model: google/gemma-7b-it
 model-index:
@@ -17,7 +18,7 @@ should probably proofread and complete it, then remove this comment. -->
 
 # Gemma-7B-It-ORPO-SALT
 
-This model is a fine-tuned version of [google/gemma-7b-it](https://huggingface.co/google/gemma-7b-it) on the None dataset.
+This model is a fine-tuned version of [google/gemma-7b-it](https://huggingface.co/google/gemma-7b-it) on the dpo_mix_en and the bct_non_cot_dpo_1000 datasets.
 It achieves the following results on the evaluation set:
 - Loss: 1.2657
 - Rewards/chosen: -0.1198

all_results.json

@@ -0,0 +1,22 @@
+{
+    "epoch": 2.9969690846635686,
+    "eval_logits/chosen": 253.6037139892578,
+    "eval_logits/rejected": 253.95994567871094,
+    "eval_logps/chosen": -1.1982638835906982,
+    "eval_logps/rejected": -1.4377323389053345,
+    "eval_loss": 1.265723466873169,
+    "eval_odds_ratio_loss": 0.6745957732200623,
+    "eval_rewards/accuracies": 0.5699999928474426,
+    "eval_rewards/chosen": -0.11982638388872147,
+    "eval_rewards/margins": 0.023946860805153847,
+    "eval_rewards/rejected": -0.14377322793006897,
+    "eval_runtime": 221.2031,
+    "eval_samples_per_second": 4.973,
+    "eval_sft_loss": 1.1982638835906982,
+    "eval_steps_per_second": 2.486,
+    "total_flos": 2.1935611788745114e+18,
+    "train_loss": 1.3469306265266197,
+    "train_runtime": 24131.5713,
+    "train_samples_per_second": 1.231,
+    "train_steps_per_second": 0.077
+}

eval_results.json

@@ -0,0 +1,17 @@
+{
+    "epoch": 2.9969690846635686,
+    "eval_logits/chosen": 253.6037139892578,
+    "eval_logits/rejected": 253.95994567871094,
+    "eval_logps/chosen": -1.1982638835906982,
+    "eval_logps/rejected": -1.4377323389053345,
+    "eval_loss": 1.265723466873169,
+    "eval_odds_ratio_loss": 0.6745957732200623,
+    "eval_rewards/accuracies": 0.5699999928474426,
+    "eval_rewards/chosen": -0.11982638388872147,
+    "eval_rewards/margins": 0.023946860805153847,
+    "eval_rewards/rejected": -0.14377322793006897,
+    "eval_runtime": 221.2031,
+    "eval_samples_per_second": 4.973,
+    "eval_sft_loss": 1.1982638835906982,
+    "eval_steps_per_second": 2.486
+}

train_results.json

@@ -0,0 +1,8 @@
+{
+    "epoch": 2.9969690846635686,
+    "total_flos": 2.1935611788745114e+18,
+    "train_loss": 1.3469306265266197,
+    "train_runtime": 24131.5713,
+    "train_samples_per_second": 1.231,
+    "train_steps_per_second": 0.077
+}