End of training

README.md

@@ -2,9 +2,15 @@
 license: apache-2.0
 base_model: mistralai/Mistral-7B-v0.1
 tags:
+- alignment-handbook
 - trl
 - dpo
 - generated_from_trainer
+- trl
+- dpo
+- generated_from_trainer
+datasets:
+- HuggingFaceH4/ultrafeedback_binarized
 model-index:
 - name: mistral-7b-dpo-full-wo-healthsearch_qa-ep3
   results: []
@@ -15,7 +21,17 @@ should probably proofread and complete it, then remove this comment. -->
 
 # mistral-7b-dpo-full-wo-healthsearch_qa-ep3
 
-This model is a fine-tuned version of [mistralai/Mistral-7B-v0.1](https://huggingface.co/mistralai/Mistral-7B-v0.1) on an unknown dataset.
+This model is a fine-tuned version of [mistralai/Mistral-7B-v0.1](https://huggingface.co/mistralai/Mistral-7B-v0.1) on the HuggingFaceH4/ultrafeedback_binarized dataset.
+It achieves the following results on the evaluation set:
+- Loss: 0.2680
+- Rewards/chosen: -0.5188
+- Rewards/rejected: -3.1722
+- Rewards/accuracies: 0.8909
+- Rewards/margins: 2.6534
+- Logps/rejected: -1410.1324
+- Logps/chosen: -651.1674
+- Logits/rejected: -3.3100
+- Logits/chosen: -3.2034
 
 ## Model description
 

all_results.json

@@ -1,5 +1,18 @@
 {
     "epoch": 0.99,
+    "eval_logits/chosen": -3.203406810760498,
+    "eval_logits/rejected": -3.3099639415740967,
+    "eval_logps/chosen": -651.1673583984375,
+    "eval_logps/rejected": -1410.1324462890625,
+    "eval_loss": 0.26801708340644836,
+    "eval_rewards/accuracies": 0.8909090757369995,
+    "eval_rewards/chosen": -0.5188338756561279,
+    "eval_rewards/margins": 2.65336275100708,
+    "eval_rewards/rejected": -3.172196626663208,
+    "eval_runtime": 727.4355,
+    "eval_samples": 7040,
+    "eval_samples_per_second": 9.678,
+    "eval_steps_per_second": 0.302,
     "train_loss": 0.44661769921752226,
     "train_runtime": 1188.3148,
     "train_samples": 5572,

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": -3.203406810760498,
+    "eval_logits/rejected": -3.3099639415740967,
+    "eval_logps/chosen": -651.1673583984375,
+    "eval_logps/rejected": -1410.1324462890625,
+    "eval_loss": 0.26801708340644836,
+    "eval_rewards/accuracies": 0.8909090757369995,
+    "eval_rewards/chosen": -0.5188338756561279,
+    "eval_rewards/margins": 2.65336275100708,
+    "eval_rewards/rejected": -3.172196626663208,
+    "eval_runtime": 727.4355,
+    "eval_samples": 7040,
+    "eval_samples_per_second": 9.678,
+    "eval_steps_per_second": 0.302
+}