End of training

README.md

@@ -2,9 +2,15 @@
 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: []
@@ -15,17 +21,17 @@ 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: 0.5066
-- Rewards/chosen: -0.9998
-- Rewards/rejected: -1.9548
+- Loss: 0.5061
+- Rewards/chosen: -1.0205
+- Rewards/rejected: -1.9944
 - Rewards/accuracies: 0.7812
-- Rewards/margins: 0.9549
-- Logps/rejected: -458.1404
-- Logps/chosen: -362.6149
-- Logits/rejected: 2.3428
-- Logits/chosen: 1.5651
+- Rewards/margins: 0.9740
+- Logps/rejected: -462.1064
+- Logps/chosen: -364.6778
+- Logits/rejected: 2.4089
+- Logits/chosen: 1.6213
 
 ## Model description
 

all_results.json

@@ -1,5 +1,18 @@
 {
     "epoch": 1.0,
+    "eval_logits/chosen": 1.6213219165802002,
+    "eval_logits/rejected": 2.4088590145111084,
+    "eval_logps/chosen": -364.67779541015625,
+    "eval_logps/rejected": -462.1063537597656,
+    "eval_loss": 0.5060754418373108,
+    "eval_rewards/accuracies": 0.78125,
+    "eval_rewards/chosen": -1.0204789638519287,
+    "eval_rewards/margins": 0.9739635586738586,
+    "eval_rewards/rejected": -1.9944424629211426,
+    "eval_runtime": 93.5321,
+    "eval_samples": 2000,
+    "eval_samples_per_second": 21.383,
+    "eval_steps_per_second": 0.342,
     "total_flos": 0.0,
     "train_loss": 0.5401068107852377,
     "train_runtime": 7660.5652,

config.json

@@ -22,6 +22,6 @@
   "tie_word_embeddings": false,
   "torch_dtype": "bfloat16",
   "transformers_version": "4.44.0",
-  "use_cache": false,
+  "use_cache": true,
   "vocab_size": 32000
 }

eval_results.json

@@ -1,16 +1,16 @@
 {
     "epoch": 1.0,
-    "eval_logits/chosen": 0.900179386138916,
-    "eval_logits/rejected": 1.8046338558197021,
-    "eval_logps/chosen": -359.44342041015625,
-    "eval_logps/rejected": -458.4696960449219,
-    "eval_loss": 0.5034610629081726,
-    "eval_rewards/accuracies": 0.78515625,
-    "eval_rewards/chosen": -0.9681348204612732,
-    "eval_rewards/margins": 0.9899410009384155,
-    "eval_rewards/rejected": -1.9580758810043335,
-    "eval_runtime": 93.6391,
+    "eval_logits/chosen": 1.6213219165802002,
+    "eval_logits/rejected": 2.4088590145111084,
+    "eval_logps/chosen": -364.67779541015625,
+    "eval_logps/rejected": -462.1063537597656,
+    "eval_loss": 0.5060754418373108,
+    "eval_rewards/accuracies": 0.78125,
+    "eval_rewards/chosen": -1.0204789638519287,
+    "eval_rewards/margins": 0.9739635586738586,
+    "eval_rewards/rejected": -1.9944424629211426,
+    "eval_runtime": 93.5321,
     "eval_samples": 2000,
-    "eval_samples_per_second": 21.359,
+    "eval_samples_per_second": 21.383,
     "eval_steps_per_second": 0.342
 }

runs/Aug19_16-52-57_simpo-8-0-0/events.out.tfevents.1724120128.simpo-8-0-0.97801.1

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ab55658549f3305129de0d5fbdc2690a8b904d4fbd79cf4ca6efa16779757894
+size 828