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-alpha_0.5_batch32
   results: []
@@ -16,17 +22,17 @@ should probably proofread and complete it, then remove this comment. -->
 
 # zephyr-7b-dpo-full-alpha_0.5_batch32
 
-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: 0.7743
-- Rewards/chosen: -1.3347
-- Rewards/rejected: -2.3725
-- Rewards/accuracies: 0.7718
-- Rewards/margins: 1.0378
-- Logps/rejected: -497.4510
-- Logps/chosen: -415.4467
-- Logits/rejected: 1.0747
-- Logits/chosen: -0.2184
+- Loss: 0.7742
+- Rewards/chosen: -1.3339
+- Rewards/rejected: -2.3724
+- Rewards/accuracies: 0.7738
+- Rewards/margins: 1.0385
+- Logps/rejected: -497.4416
+- Logps/chosen: -415.3672
+- Logits/rejected: 1.0737
+- Logits/chosen: -0.2206
 
 ## Model description
 

all_results.json

@@ -1,5 +1,18 @@
 {
     "epoch": 1.0,
+    "eval_logits/chosen": -0.2205861508846283,
+    "eval_logits/rejected": 1.0736942291259766,
+    "eval_logps/chosen": -415.36724853515625,
+    "eval_logps/rejected": -497.4416198730469,
+    "eval_loss": 0.7742469906806946,
+    "eval_rewards/accuracies": 0.773809552192688,
+    "eval_rewards/chosen": -1.33390212059021,
+    "eval_rewards/margins": 1.0384857654571533,
+    "eval_rewards/rejected": -2.3723878860473633,
+    "eval_runtime": 423.4241,
+    "eval_samples": 2000,
+    "eval_samples_per_second": 4.723,
+    "eval_steps_per_second": 0.149,
     "total_flos": 0.0,
     "train_loss": 0.45257438286507334,
     "train_runtime": 23936.1745,

config.json

@@ -22,6 +22,6 @@
   "tie_word_embeddings": false,
   "torch_dtype": "bfloat16",
   "transformers_version": "4.44.2",
-  "use_cache": false,
+  "use_cache": true,
   "vocab_size": 32000
 }

eval_results.json

@@ -0,0 +1,16 @@
+{
+    "epoch": 1.0,
+    "eval_logits/chosen": -0.2205861508846283,
+    "eval_logits/rejected": 1.0736942291259766,
+    "eval_logps/chosen": -415.36724853515625,
+    "eval_logps/rejected": -497.4416198730469,
+    "eval_loss": 0.7742469906806946,
+    "eval_rewards/accuracies": 0.773809552192688,
+    "eval_rewards/chosen": -1.33390212059021,
+    "eval_rewards/margins": 1.0384857654571533,
+    "eval_rewards/rejected": -2.3723878860473633,
+    "eval_runtime": 423.4241,
+    "eval_samples": 2000,
+    "eval_samples_per_second": 4.723,
+    "eval_steps_per_second": 0.149
+}