End of training

README.md

@@ -1,9 +1,15 @@
 ---
 base_model: Minbyul/selfbiorag-7b-wo-kqa_golden-iter-sft-step1
 tags:
+- alignment-handbook
 - trl
 - dpo
 - generated_from_trainer
+- trl
+- dpo
+- generated_from_trainer
+datasets:
+- HuggingFaceH4/ultrafeedback_binarized
 model-index:
 - name: selfbiorag-7b-wo-kqa_golden-iter-dpo-step2
   results: []
@@ -14,17 +20,17 @@ should probably proofread and complete it, then remove this comment. -->
 
 # selfbiorag-7b-wo-kqa_golden-iter-dpo-step2
 
-This model is a fine-tuned version of [Minbyul/selfbiorag-7b-wo-kqa_golden-iter-sft-step1](https://huggingface.co/Minbyul/selfbiorag-7b-wo-kqa_golden-iter-sft-step1) on an unknown dataset.
+This model is a fine-tuned version of [Minbyul/selfbiorag-7b-wo-kqa_golden-iter-sft-step1](https://huggingface.co/Minbyul/selfbiorag-7b-wo-kqa_golden-iter-sft-step1) on the HuggingFaceH4/ultrafeedback_binarized dataset.
 It achieves the following results on the evaluation set:
-- Loss: 0.3923
-- Rewards/chosen: -1.1034
-- Rewards/rejected: -3.4359
-- Rewards/accuracies: 0.7422
-- Rewards/margins: 2.3325
-- Logps/rejected: -785.2989
-- Logps/chosen: -345.1770
-- Logits/rejected: -0.8936
-- Logits/chosen: -1.1380
+- Loss: 0.4028
+- Rewards/chosen: -1.3389
+- Rewards/rejected: -4.1640
+- Rewards/accuracies: 0.7344
+- Rewards/margins: 2.8251
+- Logps/rejected: -858.1113
+- Logps/chosen: -368.7221
+- Logits/rejected: -0.8026
+- Logits/chosen: -1.0600
 
 ## Model description
 

all_results.json

@@ -1,5 +1,18 @@
 {
     "epoch": 1.0,
+    "eval_logits/chosen": -1.0600295066833496,
+    "eval_logits/rejected": -0.8026158809661865,
+    "eval_logps/chosen": -368.7221374511719,
+    "eval_logps/rejected": -858.111328125,
+    "eval_loss": 0.4027799963951111,
+    "eval_rewards/accuracies": 0.734375,
+    "eval_rewards/chosen": -1.3388758897781372,
+    "eval_rewards/margins": 2.8251354694366455,
+    "eval_rewards/rejected": -4.164011001586914,
+    "eval_runtime": 60.6412,
+    "eval_samples": 505,
+    "eval_samples_per_second": 8.328,
+    "eval_steps_per_second": 0.264,
     "train_loss": 0.26888106432225967,
     "train_runtime": 2670.405,
     "train_samples": 10364,

config.json

@@ -23,6 +23,6 @@
   "tie_word_embeddings": false,
   "torch_dtype": "bfloat16",
   "transformers_version": "4.39.0.dev0",
-  "use_cache": false,
+  "use_cache": true,
   "vocab_size": 32016
 }

eval_results.json

@@ -0,0 +1,16 @@
+{
+    "epoch": 1.0,
+    "eval_logits/chosen": -1.0600295066833496,
+    "eval_logits/rejected": -0.8026158809661865,
+    "eval_logps/chosen": -368.7221374511719,
+    "eval_logps/rejected": -858.111328125,
+    "eval_loss": 0.4027799963951111,
+    "eval_rewards/accuracies": 0.734375,
+    "eval_rewards/chosen": -1.3388758897781372,
+    "eval_rewards/margins": 2.8251354694366455,
+    "eval_rewards/rejected": -4.164011001586914,
+    "eval_runtime": 60.6412,
+    "eval_samples": 505,
+    "eval_samples_per_second": 8.328,
+    "eval_steps_per_second": 0.264
+}