End of training

README.md

@@ -2,10 +2,13 @@
 license: apache-2.0
 library_name: peft
 tags:
+- alignment-handbook
 - trl
 - dpo
 - generated_from_trainer
 base_model: alignment-handbook/zephyr-7b-sft-full
+datasets:
+- HuggingFaceH4/ultrafeedback_binarized
 model-index:
 - name: zephyr-7b-dpo-lora
   results: []
@@ -16,7 +19,7 @@ should probably proofread and complete it, then remove this comment. -->
 
 # zephyr-7b-dpo-lora
 
-This model is a fine-tuned version of [alignment-handbook/zephyr-7b-sft-full](https://huggingface.co/alignment-handbook/zephyr-7b-sft-full) on the None dataset.
+This model is a fine-tuned version of [alignment-handbook/zephyr-7b-sft-full](https://huggingface.co/alignment-handbook/zephyr-7b-sft-full) on the HuggingFaceH4/ultrafeedback_binarized dataset.
 It achieves the following results on the evaluation set:
 - Loss: 0.6776
 - Rewards/chosen: 0.0182
@@ -24,7 +27,7 @@ It achieves the following results on the evaluation set:
 - Rewards/accuracies: 0.6855
 - Rewards/margins: 0.0328
 - Logps/rejected: -262.9002
-- Logps/chosen: -280.9546
+- Logps/chosen: -280.9537
 - Logits/rejected: -2.8233
 - Logits/chosen: -2.8504
 

all_results.json

@@ -1,5 +1,18 @@
 {
     "epoch": 1.0,
+    "eval_logits/chosen": -2.850416660308838,
+    "eval_logits/rejected": -2.823277235031128,
+    "eval_logps/chosen": -280.9537353515625,
+    "eval_logps/rejected": -262.90020751953125,
+    "eval_loss": 0.6776002049446106,
+    "eval_rewards/accuracies": 0.6855000257492065,
+    "eval_rewards/chosen": 0.018197493627667427,
+    "eval_rewards/margins": 0.032815974205732346,
+    "eval_rewards/rejected": -0.014618484303355217,
+    "eval_runtime": 623.229,
+    "eval_samples": 2000,
+    "eval_samples_per_second": 3.209,
+    "eval_steps_per_second": 0.401,
     "total_flos": 0.0,
     "train_loss": 0.680465580483626,
     "train_runtime": 64957.9706,

config.json

@@ -19,7 +19,8 @@
   "rope_theta": 10000.0,
   "sliding_window": 4096,
   "tie_word_embeddings": false,
-  "transformers_version": "4.40.0",
+  "torch_dtype": "bfloat16",
+  "transformers_version": "4.40.2",
   "use_cache": true,
   "vocab_size": 32000
 }

eval_results.json

@@ -1,16 +1,16 @@
 {
     "epoch": 1.0,
-    "eval_logits/chosen": -2.7746472358703613,
-    "eval_logits/rejected": -2.752934455871582,
-    "eval_logps/chosen": -310.119873046875,
-    "eval_logps/rejected": -321.6407165527344,
-    "eval_loss": 0.5893968343734741,
-    "eval_rewards/accuracies": 0.703499972820282,
-    "eval_rewards/chosen": -0.27382245659828186,
-    "eval_rewards/margins": 0.32820531725883484,
-    "eval_rewards/rejected": -0.6020277142524719,
-    "eval_runtime": 692.2285,
+    "eval_logits/chosen": -2.850416660308838,
+    "eval_logits/rejected": -2.823277235031128,
+    "eval_logps/chosen": -280.9537353515625,
+    "eval_logps/rejected": -262.90020751953125,
+    "eval_loss": 0.6776002049446106,
+    "eval_rewards/accuracies": 0.6855000257492065,
+    "eval_rewards/chosen": 0.018197493627667427,
+    "eval_rewards/margins": 0.032815974205732346,
+    "eval_rewards/rejected": -0.014618484303355217,
+    "eval_runtime": 623.229,
     "eval_samples": 2000,
-    "eval_samples_per_second": 2.889,
-    "eval_steps_per_second": 0.361
+    "eval_samples_per_second": 3.209,
+    "eval_steps_per_second": 0.401
 }