End of training

README.md

@@ -1,9 +1,15 @@
 ---
 base_model: meta-llama/Llama-2-7b-hf
 tags:
+- alignment-handbook
 - trl
 - dpo
 - generated_from_trainer
+- trl
+- dpo
+- generated_from_trainer
+datasets:
+- HuggingFaceH4/ultrafeedback_binarized
 model-index:
 - name: llama2-7b-dpo-full-wo-healthsearch_qa-ep3
   results: []
@@ -14,7 +20,17 @@ should probably proofread and complete it, then remove this comment. -->
 
 # llama2-7b-dpo-full-wo-healthsearch_qa-ep3
 
-This model is a fine-tuned version of [meta-llama/Llama-2-7b-hf](https://huggingface.co/meta-llama/Llama-2-7b-hf) on an unknown dataset.
+This model is a fine-tuned version of [meta-llama/Llama-2-7b-hf](https://huggingface.co/meta-llama/Llama-2-7b-hf) on the HuggingFaceH4/ultrafeedback_binarized dataset.
+It achieves the following results on the evaluation set:
+- Loss: 0.6764
+- Rewards/chosen: 0.0307
+- Rewards/rejected: -0.0060
+- Rewards/accuracies: 0.7173
+- Rewards/margins: 0.0367
+- Logps/rejected: -867.9213
+- Logps/chosen: -604.5544
+- Logits/rejected: -0.5562
+- Logits/chosen: 0.0445
 
 ## Model description
 

all_results.json

@@ -1,5 +1,18 @@
 {
     "epoch": 1.0,
+    "eval_logits/chosen": 0.04453304782509804,
+    "eval_logits/rejected": -0.5562469959259033,
+    "eval_logps/chosen": -604.554443359375,
+    "eval_logps/rejected": -867.9213256835938,
+    "eval_loss": 0.6763553023338318,
+    "eval_rewards/accuracies": 0.7173295617103577,
+    "eval_rewards/chosen": 0.030703268945217133,
+    "eval_rewards/margins": 0.03669063001871109,
+    "eval_rewards/rejected": -0.005987361539155245,
+    "eval_runtime": 622.4746,
+    "eval_samples": 5631,
+    "eval_samples_per_second": 9.046,
+    "eval_steps_per_second": 0.283,
     "train_loss": 0.6753969060050117,
     "train_runtime": 1121.4796,
     "train_samples": 5168,

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": 32000
 }

eval_results.json

@@ -0,0 +1,16 @@
+{
+    "epoch": 1.0,
+    "eval_logits/chosen": 0.04453304782509804,
+    "eval_logits/rejected": -0.5562469959259033,
+    "eval_logps/chosen": -604.554443359375,
+    "eval_logps/rejected": -867.9213256835938,
+    "eval_loss": 0.6763553023338318,
+    "eval_rewards/accuracies": 0.7173295617103577,
+    "eval_rewards/chosen": 0.030703268945217133,
+    "eval_rewards/margins": 0.03669063001871109,
+    "eval_rewards/rejected": -0.005987361539155245,
+    "eval_runtime": 622.4746,
+    "eval_samples": 5631,
+    "eval_samples_per_second": 9.046,
+    "eval_steps_per_second": 0.283
+}