End of training

README.md

@@ -2,10 +2,16 @@
 license: mit
 library_name: peft
 tags:
+- alignment-handbook
+- trl
+- dpo
+- generated_from_trainer
 - trl
 - dpo
 - generated_from_trainer
 base_model: HuggingFaceH4/mistral-7b-sft-beta
+datasets:
+- HuggingFaceH4/ultrafeedback_binarized
 model-index:
 - name: mistral-sft-7b-dpo-qlora
   results: []
@@ -16,7 +22,17 @@ should probably proofread and complete it, then remove this comment. -->
 
 # mistral-sft-7b-dpo-qlora
 
-This model is a fine-tuned version of [HuggingFaceH4/mistral-7b-sft-beta](https://huggingface.co/HuggingFaceH4/mistral-7b-sft-beta) on the None dataset.
+This model is a fine-tuned version of [HuggingFaceH4/mistral-7b-sft-beta](https://huggingface.co/HuggingFaceH4/mistral-7b-sft-beta) on the HuggingFaceH4/ultrafeedback_binarized dataset.
+It achieves the following results on the evaluation set:
+- Loss: 0.6932
+- Rewards/chosen: 0.0023
+- Rewards/rejected: 0.0013
+- Rewards/accuracies: 0.5625
+- Rewards/margins: 0.0009
+- Logps/rejected: -122.8616
+- Logps/chosen: -86.2666
+- Logits/rejected: -3.0452
+- Logits/chosen: -2.9819
 
 ## Model description
 

all_results.json

@@ -1,5 +1,18 @@
 {
     "epoch": 5.0,
+    "eval_logits/chosen": -2.9818739891052246,
+    "eval_logits/rejected": -3.045213222503662,
+    "eval_logps/chosen": -86.26663208007812,
+    "eval_logps/rejected": -122.861572265625,
+    "eval_loss": 0.6932176947593689,
+    "eval_rewards/accuracies": 0.5625,
+    "eval_rewards/chosen": 0.0022567124105989933,
+    "eval_rewards/margins": 0.000948949484154582,
+    "eval_rewards/rejected": 0.0013077628100290895,
+    "eval_runtime": 11.9466,
+    "eval_samples": 20,
+    "eval_samples_per_second": 1.674,
+    "eval_steps_per_second": 0.167,
     "train_loss": 0.34364445209503175,
     "train_runtime": 77.9014,
     "train_samples": 10,

eval_results.json

@@ -1,16 +1,16 @@
 {
-    "epoch": 3.0,
-    "eval_logits/chosen": -2.9840264320373535,
-    "eval_logits/rejected": -3.045552968978882,
-    "eval_logps/chosen": -86.40597534179688,
-    "eval_logps/rejected": -122.90658569335938,
-    "eval_loss": 0.6938564777374268,
+    "epoch": 5.0,
+    "eval_logits/chosen": -2.9818739891052246,
+    "eval_logits/rejected": -3.045213222503662,
+    "eval_logps/chosen": -86.26663208007812,
+    "eval_logps/rejected": -122.861572265625,
+    "eval_loss": 0.6932176947593689,
     "eval_rewards/accuracies": 0.5625,
-    "eval_rewards/chosen": 0.0008632946410216391,
-    "eval_rewards/margins": 5.57894236408174e-06,
-    "eval_rewards/rejected": 0.000857715611346066,
-    "eval_runtime": 11.9269,
+    "eval_rewards/chosen": 0.0022567124105989933,
+    "eval_rewards/margins": 0.000948949484154582,
+    "eval_rewards/rejected": 0.0013077628100290895,
+    "eval_runtime": 11.9466,
     "eval_samples": 20,
-    "eval_samples_per_second": 1.677,
-    "eval_steps_per_second": 0.168
+    "eval_samples_per_second": 1.674,
+    "eval_steps_per_second": 0.167
 }