End of training

README.md

@@ -2,9 +2,15 @@
 license: apache-2.0
 base_model: Minbyul/mistral-7b-wo-medication_qa-sft
 tags:
+- alignment-handbook
 - trl
 - dpo
 - generated_from_trainer
+- trl
+- dpo
+- generated_from_trainer
+datasets:
+- HuggingFaceH4/ultrafeedback_binarized
 model-index:
 - name: mistral-7b-dpo-full-sft-wo-medication_qa
   results: []
@@ -15,17 +21,17 @@ should probably proofread and complete it, then remove this comment. -->
 
 # mistral-7b-dpo-full-sft-wo-medication_qa
 
-This model is a fine-tuned version of [Minbyul/mistral-7b-wo-medication_qa-sft](https://huggingface.co/Minbyul/mistral-7b-wo-medication_qa-sft) on an unknown dataset.
+This model is a fine-tuned version of [Minbyul/mistral-7b-wo-medication_qa-sft](https://huggingface.co/Minbyul/mistral-7b-wo-medication_qa-sft) on the HuggingFaceH4/ultrafeedback_binarized dataset.
 It achieves the following results on the evaluation set:
-- Logits/chosen: -2.5877
-- Logits/rejected: -2.3930
-- Logps/chosen: -803.1033
-- Logps/rejected: -1661.4266
-- Loss: 0.0753
+- Loss: 0.0756
+- Rewards/chosen: -3.7115
+- Rewards/rejected: -11.1989
 - Rewards/accuracies: 0.9531
-- Rewards/chosen: -3.7097
-- Rewards/margins: 7.4753
-- Rewards/rejected: -11.1850
+- Rewards/margins: 7.4875
+- Logps/rejected: -1662.8185
+- Logps/chosen: -803.2770
+- Logits/rejected: -2.3910
+- Logits/chosen: -2.5860
 
 ## Model description
 

all_results.json

@@ -1,5 +1,18 @@
 {
     "epoch": 1.0,
+    "eval_logits/chosen": -2.5860071182250977,
+    "eval_logits/rejected": -2.390984535217285,
+    "eval_logps/chosen": -803.2769775390625,
+    "eval_logps/rejected": -1662.8184814453125,
+    "eval_loss": 0.07558085024356842,
+    "eval_rewards/accuracies": 0.953125,
+    "eval_rewards/chosen": -3.7114715576171875,
+    "eval_rewards/margins": 7.48745059967041,
+    "eval_rewards/rejected": -11.198921203613281,
+    "eval_runtime": 32.9329,
+    "eval_samples": 252,
+    "eval_samples_per_second": 7.652,
+    "eval_steps_per_second": 0.243,
     "train_loss": 0.007740737387427577,
     "train_runtime": 396.513,
     "train_samples": 20740,

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": 1.0,
+    "eval_logits/chosen": -2.5860071182250977,
+    "eval_logits/rejected": -2.390984535217285,
+    "eval_logps/chosen": -803.2769775390625,
+    "eval_logps/rejected": -1662.8184814453125,
+    "eval_loss": 0.07558085024356842,
+    "eval_rewards/accuracies": 0.953125,
+    "eval_rewards/chosen": -3.7114715576171875,
+    "eval_rewards/margins": 7.48745059967041,
+    "eval_rewards/rejected": -11.198921203613281,
+    "eval_runtime": 32.9329,
+    "eval_samples": 252,
+    "eval_samples_per_second": 7.652,
+    "eval_steps_per_second": 0.243
+}