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 the None 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: 0.2144
-- Rewards/chosen: -2.2032
-- Rewards/rejected: -6.3258
+- Loss: 0.2130
+- Rewards/chosen: -2.1752
+- Rewards/rejected: -6.3268
 - Rewards/accuracies: 0.8984
-- Rewards/margins: 4.1226
-- Logps/rejected: -812.5331
-- Logps/chosen: -477.7561
-- Logits/rejected: 1.2277
-- Logits/chosen: -0.1965
+- Rewards/margins: 4.1516
+- Logps/rejected: -812.6320
+- Logps/chosen: -474.9558
+- Logits/rejected: 1.0727
+- Logits/chosen: -0.3077
 
 ## Model description
 

all_results.json

@@ -1,5 +1,18 @@
 {
     "epoch": 1.0,
+    "eval_logits/chosen": -0.30772945284843445,
+    "eval_logits/rejected": 1.072676658630371,
+    "eval_logps/chosen": -474.955810546875,
+    "eval_logps/rejected": -812.6320190429688,
+    "eval_loss": 0.21300281584262848,
+    "eval_rewards/accuracies": 0.8984375,
+    "eval_rewards/chosen": -2.175225019454956,
+    "eval_rewards/margins": 4.151553153991699,
+    "eval_rewards/rejected": -6.326778411865234,
+    "eval_runtime": 97.1627,
+    "eval_samples": 2000,
+    "eval_samples_per_second": 20.584,
+    "eval_steps_per_second": 0.329,
     "train_loss": 0.3183088973476298,
     "train_runtime": 7617.3811,
     "train_samples": 61135,

config.json

@@ -21,6 +21,6 @@
   "tie_word_embeddings": false,
   "torch_dtype": "bfloat16",
   "transformers_version": "4.38.2",
-  "use_cache": false,
+  "use_cache": true,
   "vocab_size": 32000
 }

eval_results.json

@@ -1,16 +1,16 @@
 {
     "epoch": 1.0,
-    "eval_logits/chosen": -0.8698186874389648,
-    "eval_logits/rejected": -0.7152774930000305,
-    "eval_logps/chosen": -366.0130310058594,
-    "eval_logps/rejected": -477.25433349609375,
-    "eval_loss": 0.48990052938461304,
-    "eval_rewards/accuracies": 0.765625,
-    "eval_rewards/chosen": -0.8437946438789368,
-    "eval_rewards/margins": 1.0511335134506226,
-    "eval_rewards/rejected": -1.894927978515625,
-    "eval_runtime": 97.4253,
+    "eval_logits/chosen": -0.30772945284843445,
+    "eval_logits/rejected": 1.072676658630371,
+    "eval_logps/chosen": -474.955810546875,
+    "eval_logps/rejected": -812.6320190429688,
+    "eval_loss": 0.21300281584262848,
+    "eval_rewards/accuracies": 0.8984375,
+    "eval_rewards/chosen": -2.175225019454956,
+    "eval_rewards/margins": 4.151553153991699,
+    "eval_rewards/rejected": -6.326778411865234,
+    "eval_runtime": 97.1627,
     "eval_samples": 2000,
-    "eval_samples_per_second": 20.529,
-    "eval_steps_per_second": 0.328
+    "eval_samples_per_second": 20.584,
+    "eval_steps_per_second": 0.329
 }

runs/May14_18-00-17_n136-100-194/events.out.tfevents.1715689231.n136-100-194.2788262.1

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d494a654db4f89e51f698233f2a55f25765672981dc07b70a6786f66fa398583
+size 828