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: 1.3374
-- Rewards/chosen: -14.5395
-- Rewards/rejected: -18.1313
-- Rewards/accuracies: 0.6409
-- Rewards/margins: 3.5918
-- Logps/rejected: -2073.3389
-- Logps/chosen: -1735.9152
-- Logits/rejected: -0.6740
-- Logits/chosen: -1.0018
+- Loss: 1.3377
+- Rewards/chosen: -14.5626
+- Rewards/rejected: -18.1281
+- Rewards/accuracies: 0.6389
+- Rewards/margins: 3.5654
+- Logps/rejected: -2073.0146
+- Logps/chosen: -1738.2311
+- Logits/rejected: -0.6819
+- Logits/chosen: -1.0035
 
 ## Model description
 

all_results.json

@@ -1,5 +1,18 @@
 {
     "epoch": 0.9994767137624281,
+    "eval_logits/chosen": -1.0034536123275757,
+    "eval_logits/rejected": -0.6818602085113525,
+    "eval_logps/chosen": -1738.2310791015625,
+    "eval_logps/rejected": -2073.0146484375,
+    "eval_loss": 1.3376935720443726,
+    "eval_rewards/accuracies": 0.6388888955116272,
+    "eval_rewards/chosen": -14.562629699707031,
+    "eval_rewards/margins": 3.5654470920562744,
+    "eval_rewards/rejected": -18.12807846069336,
+    "eval_runtime": 177.2147,
+    "eval_samples": 2000,
+    "eval_samples_per_second": 11.286,
+    "eval_steps_per_second": 0.356,
     "total_flos": 0.0,
     "train_loss": 2.1165736393154604,
     "train_runtime": 18133.1885,

config.json

@@ -21,6 +21,6 @@
   "tie_word_embeddings": false,
   "torch_dtype": "bfloat16",
   "transformers_version": "4.40.1",
-  "use_cache": false,
+  "use_cache": true,
   "vocab_size": 32000
 }

eval_results.json

@@ -1,16 +1,16 @@
 {
     "epoch": 0.9994767137624281,
-    "eval_logits/chosen": -0.35997089743614197,
-    "eval_logits/rejected": 0.7096247673034668,
-    "eval_logps/chosen": -440.0589599609375,
-    "eval_logps/rejected": -523.0543212890625,
-    "eval_loss": 0.4903615117073059,
-    "eval_rewards/accuracies": 0.7698412537574768,
-    "eval_rewards/chosen": -1.5809098482131958,
-    "eval_rewards/margins": 1.0475648641586304,
-    "eval_rewards/rejected": -2.6284749507904053,
-    "eval_runtime": 176.963,
+    "eval_logits/chosen": -1.0034536123275757,
+    "eval_logits/rejected": -0.6818602085113525,
+    "eval_logps/chosen": -1738.2310791015625,
+    "eval_logps/rejected": -2073.0146484375,
+    "eval_loss": 1.3376935720443726,
+    "eval_rewards/accuracies": 0.6388888955116272,
+    "eval_rewards/chosen": -14.562629699707031,
+    "eval_rewards/margins": 3.5654470920562744,
+    "eval_rewards/rejected": -18.12807846069336,
+    "eval_runtime": 177.2147,
     "eval_samples": 2000,
-    "eval_samples_per_second": 11.302,
+    "eval_samples_per_second": 11.286,
     "eval_steps_per_second": 0.356
 }

runs/Apr26_05-14-17_dsw-3393-565b8f59b6-l5p7k/events.out.tfevents.1714127938.dsw-3393-565b8f59b6-l5p7k.40537.1

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b3413bb432dd9f8df62606ad33e3d63476f2890e6d0f53b3ea6b6448fa5725c7
+size 730