End of training

README.md

@@ -2,9 +2,15 @@
 license: apache-2.0
 base_model: Minbyul/mistral-7b-wo-kqa_golden-iter-sft-step1
 tags:
+- alignment-handbook
 - trl
 - dpo
 - generated_from_trainer
+- trl
+- dpo
+- generated_from_trainer
+datasets:
+- HuggingFaceH4/ultrafeedback_binarized
 model-index:
 - name: mistral-7b-wo-kqa_golden-iter-dpo-step2
   results: []
@@ -15,7 +21,17 @@ should probably proofread and complete it, then remove this comment. -->
 
 # mistral-7b-wo-kqa_golden-iter-dpo-step2
 
-This model is a fine-tuned version of [Minbyul/mistral-7b-wo-kqa_golden-iter-sft-step1](https://huggingface.co/Minbyul/mistral-7b-wo-kqa_golden-iter-sft-step1) on an unknown dataset.
+This model is a fine-tuned version of [Minbyul/mistral-7b-wo-kqa_golden-iter-sft-step1](https://huggingface.co/Minbyul/mistral-7b-wo-kqa_golden-iter-sft-step1) on the HuggingFaceH4/ultrafeedback_binarized dataset.
+It achieves the following results on the evaluation set:
+- Loss: 0.5463
+- Rewards/chosen: -0.5234
+- Rewards/rejected: -1.3980
+- Rewards/accuracies: 0.8125
+- Rewards/margins: 0.8746
+- Logps/rejected: -518.7740
+- Logps/chosen: -233.3091
+- Logits/rejected: -3.7466
+- Logits/chosen: -3.9181
 
 ## Model description
 

all_results.json

@@ -1,5 +1,18 @@
 {
     "epoch": 0.99,
+    "eval_logits/chosen": -3.918135404586792,
+    "eval_logits/rejected": -3.7465851306915283,
+    "eval_logps/chosen": -233.3091278076172,
+    "eval_logps/rejected": -518.7740478515625,
+    "eval_loss": 0.5462985038757324,
+    "eval_rewards/accuracies": 0.8125,
+    "eval_rewards/chosen": -0.5233834981918335,
+    "eval_rewards/margins": 0.8745772242546082,
+    "eval_rewards/rejected": -1.3979607820510864,
+    "eval_runtime": 37.2528,
+    "eval_samples": 290,
+    "eval_samples_per_second": 7.785,
+    "eval_steps_per_second": 0.268,
     "train_loss": 0.49557680382833375,
     "train_runtime": 1348.5444,
     "train_samples": 5827,

config.json

@@ -21,6 +21,6 @@
   "tie_word_embeddings": false,
   "torch_dtype": "bfloat16",
   "transformers_version": "4.39.0.dev0",
-  "use_cache": false,
+  "use_cache": true,
   "vocab_size": 32000
 }

eval_results.json

@@ -0,0 +1,16 @@
+{
+    "epoch": 0.99,
+    "eval_logits/chosen": -3.918135404586792,
+    "eval_logits/rejected": -3.7465851306915283,
+    "eval_logps/chosen": -233.3091278076172,
+    "eval_logps/rejected": -518.7740478515625,
+    "eval_loss": 0.5462985038757324,
+    "eval_rewards/accuracies": 0.8125,
+    "eval_rewards/chosen": -0.5233834981918335,
+    "eval_rewards/margins": 0.8745772242546082,
+    "eval_rewards/rejected": -1.3979607820510864,
+    "eval_runtime": 37.2528,
+    "eval_samples": 290,
+    "eval_samples_per_second": 7.785,
+    "eval_steps_per_second": 0.268
+}