End of training

README.md

@@ -1,8 +1,11 @@
 ---
 base_model: alignment-handbook/zephyr-7b-sft-full
+datasets:
+- HuggingFaceH4/ultrafeedback_binarized
 library_name: peft
 license: apache-2.0
 tags:
+- alignment-handbook
 - trl
 - dpo
 - generated_from_trainer
@@ -16,17 +19,17 @@ should probably proofread and complete it, then remove this comment. -->
 
 # zephyr-7b-dpo-lora-r16-20k
 
-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: 0.5301
-- Rewards/chosen: -0.7870
-- Rewards/rejected: -1.4645
-- Rewards/accuracies: 0.7202
-- Rewards/margins: 0.6775
-- Logps/rejected: -394.4780
-- Logps/chosen: -361.9388
-- Logits/rejected: -2.5045
-- Logits/chosen: -2.5477
+- Loss: 0.5302
+- Rewards/chosen: -0.7891
+- Rewards/rejected: -1.4667
+- Rewards/accuracies: 0.7183
+- Rewards/margins: 0.6776
+- Logps/rejected: -394.6997
+- Logps/chosen: -362.1445
+- Logits/rejected: -2.5080
+- Logits/chosen: -2.5508
 
 ## Model description
 

all_results.json

@@ -1,5 +1,18 @@
 {
     "epoch": 1.0,
+    "eval_logits/chosen": -2.550788640975952,
+    "eval_logits/rejected": -2.5080249309539795,
+    "eval_logps/chosen": -362.1445007324219,
+    "eval_logps/rejected": -394.6997375488281,
+    "eval_loss": 0.5302429795265198,
+    "eval_rewards/accuracies": 0.7182539701461792,
+    "eval_rewards/chosen": -0.7890641093254089,
+    "eval_rewards/margins": 0.6776372790336609,
+    "eval_rewards/rejected": -1.4667012691497803,
+    "eval_runtime": 166.4681,
+    "eval_samples": 500,
+    "eval_samples_per_second": 3.004,
+    "eval_steps_per_second": 0.378,
     "total_flos": 0.0,
     "train_loss": 0.5873338260650635,
     "train_runtime": 15803.1996,

eval_results.json

@@ -1,16 +1,16 @@
 {
     "epoch": 1.0,
-    "eval_logits/chosen": -2.553934335708618,
-    "eval_logits/rejected": -2.510233163833618,
-    "eval_logps/chosen": -362.36248779296875,
-    "eval_logps/rejected": -395.8988952636719,
-    "eval_loss": 0.5367478132247925,
-    "eval_rewards/accuracies": 0.7103174328804016,
-    "eval_rewards/chosen": -0.7912437319755554,
-    "eval_rewards/margins": 0.6874489188194275,
-    "eval_rewards/rejected": -1.4786925315856934,
-    "eval_runtime": 165.9175,
+    "eval_logits/chosen": -2.550788640975952,
+    "eval_logits/rejected": -2.5080249309539795,
+    "eval_logps/chosen": -362.1445007324219,
+    "eval_logps/rejected": -394.6997375488281,
+    "eval_loss": 0.5302429795265198,
+    "eval_rewards/accuracies": 0.7182539701461792,
+    "eval_rewards/chosen": -0.7890641093254089,
+    "eval_rewards/margins": 0.6776372790336609,
+    "eval_rewards/rejected": -1.4667012691497803,
+    "eval_runtime": 166.4681,
     "eval_samples": 500,
-    "eval_samples_per_second": 3.014,
-    "eval_steps_per_second": 0.38
+    "eval_samples_per_second": 3.004,
+    "eval_steps_per_second": 0.378
 }