End of training

README.md

@@ -3,9 +3,15 @@ library_name: transformers
 license: apache-2.0
 base_model: alignment-handbook/zephyr-7b-sft-full
 tags:
+- alignment-handbook
 - trl
 - dpo
 - generated_from_trainer
+- trl
+- dpo
+- generated_from_trainer
+datasets:
+- HuggingFaceH4/ultrafeedback_binarized
 model-index:
 - name: zephyr-7b-dpo-full
   results: []
@@ -16,17 +22,17 @@ should probably proofread and complete it, then remove this comment. -->
 
 # zephyr-7b-dpo-full
 
-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: 0.5006
-- Rewards/chosen: -1.0395
-- Rewards/rejected: -2.0108
+- Loss: 0.5004
+- Rewards/chosen: -1.0684
+- Rewards/rejected: -2.0671
 - Rewards/accuracies: 0.7852
-- Rewards/margins: 0.9713
-- Logps/rejected: -463.7658
-- Logps/chosen: -366.5239
-- Logits/rejected: 0.6358
-- Logits/chosen: -0.3905
+- Rewards/margins: 0.9987
+- Logps/rejected: -469.3939
+- Logps/chosen: -369.4198
+- Logits/rejected: 0.7735
+- Logits/chosen: -0.2945
 
 ## Model description
 

all_results.json

@@ -1,5 +1,18 @@
 {
     "epoch": 1.0,
+    "eval_logits/chosen": -0.2944760024547577,
+    "eval_logits/rejected": 0.7734581232070923,
+    "eval_logps/chosen": -369.4197998046875,
+    "eval_logps/rejected": -469.3939208984375,
+    "eval_loss": 0.5004380345344543,
+    "eval_rewards/accuracies": 0.78515625,
+    "eval_rewards/chosen": -1.0684473514556885,
+    "eval_rewards/margins": 0.998650312423706,
+    "eval_rewards/rejected": -2.0670976638793945,
+    "eval_runtime": 39.5943,
+    "eval_samples": 2000,
+    "eval_samples_per_second": 50.512,
+    "eval_steps_per_second": 0.808,
     "total_flos": 0.0,
     "train_loss": 0.5400827195355085,
     "train_runtime": 3932.0785,

config.json

@@ -22,6 +22,6 @@
   "tie_word_embeddings": false,
   "torch_dtype": "bfloat16",
   "transformers_version": "4.45.1",
-  "use_cache": false,
+  "use_cache": true,
   "vocab_size": 32000
 }

eval_results.json

@@ -0,0 +1,16 @@
+{
+    "epoch": 1.0,
+    "eval_logits/chosen": -0.2944760024547577,
+    "eval_logits/rejected": 0.7734581232070923,
+    "eval_logps/chosen": -369.4197998046875,
+    "eval_logps/rejected": -469.3939208984375,
+    "eval_loss": 0.5004380345344543,
+    "eval_rewards/accuracies": 0.78515625,
+    "eval_rewards/chosen": -1.0684473514556885,
+    "eval_rewards/margins": 0.998650312423706,
+    "eval_rewards/rejected": -2.0670976638793945,
+    "eval_runtime": 39.5943,
+    "eval_samples": 2000,
+    "eval_samples_per_second": 50.512,
+    "eval_steps_per_second": 0.808
+}

runs/Oct03_21-31-06_bohan-data-job-0-worker-0/events.out.tfevents.1727995532.bohan-data-job-0-worker-0.4780.1

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:852beaa4aa0f20f8dfd47d155f95c5c6e80468919755e90998d8e6be0c7edd4d
+size 780