End of training

README.md

@@ -2,10 +2,16 @@
 license: apache-2.0
 base_model: alignment-handbook/zephyr-7b-sft-full
 tags:
+- alignment-handbook
+- trl
+- dpo
+- generated_from_trainer
 - trl
 - dpo
 - alignment-handbook
 - generated_from_trainer
+datasets:
+- HuggingFaceH4/ultrafeedback_binarized
 model-index:
 - name: zephyr-7b-dpo-full-ultrabin-high-margin
   results: []
@@ -16,17 +22,17 @@ should probably proofread and complete it, then remove this comment. -->
 
 # zephyr-7b-dpo-full-ultrabin-high-margin
 
-This model is a fine-tuned version of [alignment-handbook/zephyr-7b-sft-full](https://huggingface.co/alignment-handbook/zephyr-7b-sft-full) on an unknown 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: 0.5598
-- Rewards/chosen: -0.6746
-- Rewards/rejected: -1.5654
-- Rewards/accuracies: 0.75
-- Rewards/margins: 0.8907
-- Logps/rejected: -419.1961
-- Logps/chosen: -330.0835
-- Logits/rejected: 0.2134
-- Logits/chosen: -0.2417
+- Loss: 0.5523
+- Rewards/chosen: -0.7174
+- Rewards/rejected: -1.6454
+- Rewards/accuracies: 0.7461
+- Rewards/margins: 0.9280
+- Logps/rejected: -427.1986
+- Logps/chosen: -334.3585
+- Logits/rejected: 0.5161
+- Logits/chosen: 0.0387
 
 ## Model description
 

all_results.json

@@ -1,17 +1,17 @@
 {
     "epoch": 0.9965156794425087,
-    "eval_logits/chosen": -0.16245996952056885,
-    "eval_logits/rejected": 0.3127332031726837,
-    "eval_logps/chosen": -328.00042724609375,
-    "eval_logps/rejected": -417.0811462402344,
-    "eval_loss": 0.5565306544303894,
-    "eval_rewards/accuracies": 0.7578125,
-    "eval_rewards/chosen": -0.653814435005188,
-    "eval_rewards/margins": 0.8903896808624268,
-    "eval_rewards/rejected": -1.5442042350769043,
-    "eval_runtime": 102.1382,
+    "eval_logits/chosen": 0.03869672119617462,
+    "eval_logits/rejected": 0.5161129236221313,
+    "eval_logps/chosen": -334.3584899902344,
+    "eval_logps/rejected": -427.1986389160156,
+    "eval_loss": 0.5522767305374146,
+    "eval_rewards/accuracies": 0.74609375,
+    "eval_rewards/chosen": -0.7173951864242554,
+    "eval_rewards/margins": 0.9279837608337402,
+    "eval_rewards/rejected": -1.6453789472579956,
+    "eval_runtime": 102.3923,
     "eval_samples": 2000,
-    "eval_samples_per_second": 19.581,
+    "eval_samples_per_second": 19.533,
     "eval_steps_per_second": 0.313,
     "total_flos": 0.0,
     "train_loss": 0.42797933008287337,

config.json

@@ -21,6 +21,6 @@
   "tie_word_embeddings": false,
   "torch_dtype": "bfloat16",
   "transformers_version": "4.41.2",
-  "use_cache": false,
+  "use_cache": true,
   "vocab_size": 32000
 }

eval_results.json

@@ -1,16 +1,16 @@
 {
     "epoch": 0.9965156794425087,
-    "eval_logits/chosen": -0.16245996952056885,
-    "eval_logits/rejected": 0.3127332031726837,
-    "eval_logps/chosen": -328.00042724609375,
-    "eval_logps/rejected": -417.0811462402344,
-    "eval_loss": 0.5565306544303894,
-    "eval_rewards/accuracies": 0.7578125,
-    "eval_rewards/chosen": -0.653814435005188,
-    "eval_rewards/margins": 0.8903896808624268,
-    "eval_rewards/rejected": -1.5442042350769043,
-    "eval_runtime": 102.1382,
+    "eval_logits/chosen": 0.03869672119617462,
+    "eval_logits/rejected": 0.5161129236221313,
+    "eval_logps/chosen": -334.3584899902344,
+    "eval_logps/rejected": -427.1986389160156,
+    "eval_loss": 0.5522767305374146,
+    "eval_rewards/accuracies": 0.74609375,
+    "eval_rewards/chosen": -0.7173951864242554,
+    "eval_rewards/margins": 0.9279837608337402,
+    "eval_rewards/rejected": -1.6453789472579956,
+    "eval_runtime": 102.3923,
     "eval_samples": 2000,
-    "eval_samples_per_second": 19.581,
+    "eval_samples_per_second": 19.533,
     "eval_steps_per_second": 0.313
 }