End of training

README.md

@@ -3,10 +3,16 @@ 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
 - alignment-handbook
 - generated_from_trainer
+datasets:
+- HuggingFaceH4/ultrafeedback_binarized
 model-index:
 - name: zephyr-7b-dpo-full
   results: []
@@ -17,17 +23,17 @@ 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:
-- Logits/chosen: -1.8309
-- Logits/rejected: -1.4322
-- Logps/chosen: -726.7409
-- Logps/rejected: -1277.9103
-- Loss: 0.2493
-- Rewards/accuracies: 0.8509
-- Rewards/chosen: -4.3528
-- Rewards/margins: 5.3912
-- Rewards/rejected: -9.7441
+- Loss: 0.2473
+- Rewards/chosen: -4.6815
+- Rewards/rejected: -10.5131
+- Rewards/accuracies: 0.8525
+- Rewards/margins: 5.8316
+- Logps/rejected: -1354.8135
+- Logps/chosen: -759.6055
+- Logits/rejected: -1.2709
+- Logits/chosen: -1.7157
 
 ## Model description
 

all_results.json

@@ -9,10 +9,10 @@
     "eval_rewards/chosen": -4.681485176086426,
     "eval_rewards/margins": 5.831614017486572,
     "eval_rewards/rejected": -10.513099670410156,
-    "eval_runtime": 468.8807,
+    "eval_runtime": 465.6415,
     "eval_samples": 5126,
-    "eval_samples_per_second": 10.932,
-    "eval_steps_per_second": 0.343,
+    "eval_samples_per_second": 11.008,
+    "eval_steps_per_second": 0.346,
     "total_flos": 0.0,
     "train_loss": 0.0,
     "train_runtime": 0.0182,

config.json

@@ -22,6 +22,6 @@
   "tie_word_embeddings": false,
   "torch_dtype": "bfloat16",
   "transformers_version": "4.44.1",
-  "use_cache": false,
+  "use_cache": true,
   "vocab_size": 32000
 }

eval_results.json

@@ -9,8 +9,8 @@
     "eval_rewards/chosen": -4.681485176086426,
     "eval_rewards/margins": 5.831614017486572,
     "eval_rewards/rejected": -10.513099670410156,
-    "eval_runtime": 468.8807,
+    "eval_runtime": 465.6415,
     "eval_samples": 5126,
-    "eval_samples_per_second": 10.932,
-    "eval_steps_per_second": 0.343
+    "eval_samples_per_second": 11.008,
+    "eval_steps_per_second": 0.346
 }

runs/Aug25_00-49-50_ip-10-0-9-154.ec2.internal/events.out.tfevents.1724549094.ip-10-0-9-154.ec2.internal.57390.1

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:262ab9e0ef4b79575f759e56e6012dc0787cb705c35d0362c0f494037908c539
+size 828