End of training

README.md

@@ -1,9 +1,15 @@
 ---
 base_model: dmis-lab/selfbiorag_7b
 tags:
+- alignment-handbook
 - trl
 - dpo
 - generated_from_trainer
+- trl
+- dpo
+- generated_from_trainer
+datasets:
+- HuggingFaceH4/ultrafeedback_binarized
 model-index:
 - name: selfbiorag-7b-dpo-full-wo-kqa_silver_wogold-ep3
   results: []
@@ -14,17 +20,17 @@ should probably proofread and complete it, then remove this comment. -->
 
 # selfbiorag-7b-dpo-full-wo-kqa_silver_wogold-ep3
 
-This model is a fine-tuned version of [dmis-lab/selfbiorag_7b](https://huggingface.co/dmis-lab/selfbiorag_7b) on an unknown dataset.
+This model is a fine-tuned version of [dmis-lab/selfbiorag_7b](https://huggingface.co/dmis-lab/selfbiorag_7b) on the HuggingFaceH4/ultrafeedback_binarized dataset.
 It achieves the following results on the evaluation set:
-- Loss: 0.6403
-- Rewards/chosen: 0.1251
-- Rewards/rejected: 0.0020
-- Rewards/accuracies: 0.7554
-- Rewards/margins: 0.1232
-- Logps/rejected: -171.1355
-- Logps/chosen: -150.7145
-- Logits/rejected: -1.5597
-- Logits/chosen: -1.7911
+- Loss: 0.6392
+- Rewards/chosen: 0.1232
+- Rewards/rejected: -0.0030
+- Rewards/accuracies: 0.7527
+- Rewards/margins: 0.1262
+- Logps/rejected: -171.6258
+- Logps/chosen: -150.9050
+- Logits/rejected: -1.5645
+- Logits/chosen: -1.7964
 
 ## Model description
 

all_results.json

@@ -1,5 +1,18 @@
 {
     "epoch": 1.0,
+    "eval_logits/chosen": -1.7963508367538452,
+    "eval_logits/rejected": -1.5644817352294922,
+    "eval_logps/chosen": -150.905029296875,
+    "eval_logps/rejected": -171.62582397460938,
+    "eval_loss": 0.6391857266426086,
+    "eval_rewards/accuracies": 0.7526785731315613,
+    "eval_rewards/chosen": 0.12320567667484283,
+    "eval_rewards/margins": 0.1261579990386963,
+    "eval_rewards/rejected": -0.0029523156117647886,
+    "eval_runtime": 429.2674,
+    "eval_samples": 4472,
+    "eval_samples_per_second": 10.418,
+    "eval_steps_per_second": 0.326,
     "train_loss": 0.6365382266896111,
     "train_runtime": 6736.0228,
     "train_samples": 25108,

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": -1.7963508367538452,
+    "eval_logits/rejected": -1.5644817352294922,
+    "eval_logps/chosen": -150.905029296875,
+    "eval_logps/rejected": -171.62582397460938,
+    "eval_loss": 0.6391857266426086,
+    "eval_rewards/accuracies": 0.7526785731315613,
+    "eval_rewards/chosen": 0.12320567667484283,
+    "eval_rewards/margins": 0.1261579990386963,
+    "eval_rewards/rejected": -0.0029523156117647886,
+    "eval_runtime": 429.2674,
+    "eval_samples": 4472,
+    "eval_samples_per_second": 10.418,
+    "eval_steps_per_second": 0.326
+}