End of training

README.md

@@ -2,10 +2,16 @@
 license: llama2
 base_model: Minbyul/meditron-7b-wo-medication_qa-sft
 tags:
+- alignment-handbook
+- trl
+- dpo
+- generated_from_trainer
 - trl
 - dpo
 - alignment-handbook
 - generated_from_trainer
+datasets:
+- HuggingFaceH4/ultrafeedback_binarized
 model-index:
 - name: meditron-7b-dpo-full-sft-wo-medication_qa
   results: []
@@ -16,7 +22,17 @@ should probably proofread and complete it, then remove this comment. -->
 
 # meditron-7b-dpo-full-sft-wo-medication_qa
 
-This model is a fine-tuned version of [Minbyul/meditron-7b-wo-medication_qa-sft](https://huggingface.co/Minbyul/meditron-7b-wo-medication_qa-sft) on an unknown dataset.
+This model is a fine-tuned version of [Minbyul/meditron-7b-wo-medication_qa-sft](https://huggingface.co/Minbyul/meditron-7b-wo-medication_qa-sft) on the HuggingFaceH4/ultrafeedback_binarized dataset.
+It achieves the following results on the evaluation set:
+- Loss: 0.6347
+- Rewards/chosen: -0.0755
+- Rewards/rejected: -0.3042
+- Rewards/accuracies: 0.7812
+- Rewards/margins: 0.2287
+- Logps/rejected: -613.1549
+- Logps/chosen: -395.2336
+- Logits/rejected: -1.0640
+- Logits/chosen: -1.1745
 
 ## Model description
 

all_results.json

@@ -1,18 +1,18 @@
 {
     "epoch": 0.99,
-    "eval_logits/chosen": -1.1746665239334106,
-    "eval_logits/rejected": -1.0633888244628906,
-    "eval_logps/chosen": -394.94110107421875,
-    "eval_logps/rejected": -613.10498046875,
-    "eval_loss": 0.6342011094093323,
+    "eval_logits/chosen": -1.1745316982269287,
+    "eval_logits/rejected": -1.0640313625335693,
+    "eval_logps/chosen": -395.233642578125,
+    "eval_logps/rejected": -613.1549072265625,
+    "eval_loss": 0.6347063779830933,
     "eval_rewards/accuracies": 0.78125,
-    "eval_rewards/chosen": -0.07253288477659225,
-    "eval_rewards/margins": 0.23117530345916748,
-    "eval_rewards/rejected": -0.30370816588401794,
-    "eval_runtime": 30.5473,
+    "eval_rewards/chosen": -0.0754583477973938,
+    "eval_rewards/margins": 0.22874917089939117,
+    "eval_rewards/rejected": -0.3042075037956238,
+    "eval_runtime": 30.2202,
     "eval_samples": 240,
-    "eval_samples_per_second": 7.857,
-    "eval_steps_per_second": 0.262,
+    "eval_samples_per_second": 7.942,
+    "eval_steps_per_second": 0.265,
     "train_loss": 0.5081947011462713,
     "train_runtime": 896.7003,
     "train_samples": 3790,

config.json

@@ -23,6 +23,6 @@
   "tie_word_embeddings": false,
   "torch_dtype": "bfloat16",
   "transformers_version": "4.39.0.dev0",
-  "use_cache": false,
+  "use_cache": true,
   "vocab_size": 32017
 }

eval_results.json

@@ -1,16 +1,16 @@
 {
     "epoch": 0.99,
-    "eval_logits/chosen": -1.1746665239334106,
-    "eval_logits/rejected": -1.0633888244628906,
-    "eval_logps/chosen": -394.94110107421875,
-    "eval_logps/rejected": -613.10498046875,
-    "eval_loss": 0.6342011094093323,
+    "eval_logits/chosen": -1.1745316982269287,
+    "eval_logits/rejected": -1.0640313625335693,
+    "eval_logps/chosen": -395.233642578125,
+    "eval_logps/rejected": -613.1549072265625,
+    "eval_loss": 0.6347063779830933,
     "eval_rewards/accuracies": 0.78125,
-    "eval_rewards/chosen": -0.07253288477659225,
-    "eval_rewards/margins": 0.23117530345916748,
-    "eval_rewards/rejected": -0.30370816588401794,
-    "eval_runtime": 30.5473,
+    "eval_rewards/chosen": -0.0754583477973938,
+    "eval_rewards/margins": 0.22874917089939117,
+    "eval_rewards/rejected": -0.3042075037956238,
+    "eval_runtime": 30.2202,
     "eval_samples": 240,
-    "eval_samples_per_second": 7.857,
-    "eval_steps_per_second": 0.262
+    "eval_samples_per_second": 7.942,
+    "eval_steps_per_second": 0.265
 }