End of training

README.md

@@ -2,9 +2,15 @@
 license: apache-2.0
 base_model: Minbyul/biomistral-7b-wo-kqa_golden-iter-dpo-step2
 tags:
+- alignment-handbook
 - trl
 - dpo
 - generated_from_trainer
+- trl
+- dpo
+- generated_from_trainer
+datasets:
+- HuggingFaceH4/ultrafeedback_binarized
 model-index:
 - name: biomistral-7b-wo-kqa_golden-iter-dpo-step3
   results: []
@@ -15,7 +21,17 @@ should probably proofread and complete it, then remove this comment. -->
 
 # biomistral-7b-wo-kqa_golden-iter-dpo-step3
 
-This model is a fine-tuned version of [Minbyul/biomistral-7b-wo-kqa_golden-iter-dpo-step2](https://huggingface.co/Minbyul/biomistral-7b-wo-kqa_golden-iter-dpo-step2) on an unknown dataset.
+This model is a fine-tuned version of [Minbyul/biomistral-7b-wo-kqa_golden-iter-dpo-step2](https://huggingface.co/Minbyul/biomistral-7b-wo-kqa_golden-iter-dpo-step2) on the HuggingFaceH4/ultrafeedback_binarized dataset.
+It achieves the following results on the evaluation set:
+- Loss: 0.6914
+- Rewards/chosen: 0.0080
+- Rewards/rejected: 0.0043
+- Rewards/accuracies: 0.6964
+- Rewards/margins: 0.0037
+- Logps/rejected: -164.6167
+- Logps/chosen: -234.3960
+- Logits/rejected: -2.1831
+- Logits/chosen: -2.2946
 
 ## Model description
 

all_results.json

@@ -1,5 +1,18 @@
 {
     "epoch": 0.99,
+    "eval_logits/chosen": -2.294609308242798,
+    "eval_logits/rejected": -2.183091878890991,
+    "eval_logps/chosen": -234.39601135253906,
+    "eval_logps/rejected": -164.61671447753906,
+    "eval_loss": 0.6914091110229492,
+    "eval_rewards/accuracies": 0.6964285969734192,
+    "eval_rewards/chosen": 0.00801192969083786,
+    "eval_rewards/margins": 0.003745319088920951,
+    "eval_rewards/rejected": 0.004266609903424978,
+    "eval_runtime": 18.9978,
+    "eval_samples": 204,
+    "eval_samples_per_second": 10.738,
+    "eval_steps_per_second": 0.368,
     "train_loss": 0.6816894855905087,
     "train_runtime": 1070.9433,
     "train_samples": 6039,

config.json

@@ -21,6 +21,6 @@
   "tie_word_embeddings": false,
   "torch_dtype": "bfloat16",
   "transformers_version": "4.38.2",
-  "use_cache": false,
+  "use_cache": true,
   "vocab_size": 32000
 }

eval_results.json

@@ -1,16 +1,16 @@
 {
-    "epoch": 1.0,
-    "eval_logits/chosen": -0.8710112571716309,
-    "eval_logits/rejected": -1.522813320159912,
-    "eval_logps/chosen": -40.126407623291016,
-    "eval_logps/rejected": -60.11854934692383,
-    "eval_loss": 0.6829047799110413,
-    "eval_rewards/accuracies": 0.6299999952316284,
-    "eval_rewards/chosen": -0.050214797258377075,
-    "eval_rewards/margins": 0.02983289770781994,
-    "eval_rewards/rejected": -0.08004770427942276,
-    "eval_runtime": 65.4744,
-    "eval_samples": 785,
-    "eval_samples_per_second": 11.989,
-    "eval_steps_per_second": 0.382
+    "epoch": 0.99,
+    "eval_logits/chosen": -2.294609308242798,
+    "eval_logits/rejected": -2.183091878890991,
+    "eval_logps/chosen": -234.39601135253906,
+    "eval_logps/rejected": -164.61671447753906,
+    "eval_loss": 0.6914091110229492,
+    "eval_rewards/accuracies": 0.6964285969734192,
+    "eval_rewards/chosen": 0.00801192969083786,
+    "eval_rewards/margins": 0.003745319088920951,
+    "eval_rewards/rejected": 0.004266609903424978,
+    "eval_runtime": 18.9978,
+    "eval_samples": 204,
+    "eval_samples_per_second": 10.738,
+    "eval_steps_per_second": 0.368
 }