End of training

README.md

@@ -1,10 +1,13 @@
 ---
 library_name: peft
 tags:
+- alignment-handbook
 - trl
 - dpo
 - generated_from_trainer
 base_model: NbAiLab/nb-gpt-j-6B-v2
+datasets:
+- hugodk-sch/aftonposten_title_prefs
 model-index:
 - name: aftonposten-6b-align-scan
   results: []
@@ -15,17 +18,17 @@ should probably proofread and complete it, then remove this comment. -->
 
 # aftonposten-6b-align-scan
 
-This model is a fine-tuned version of [NbAiLab/nb-gpt-j-6B-v2](https://huggingface.co/NbAiLab/nb-gpt-j-6B-v2) on an unknown dataset.
+This model is a fine-tuned version of [data/ap-gpt-j-6b-sft-qlora-04-08](https://huggingface.co/data/ap-gpt-j-6b-sft-qlora-04-08) on the hugodk-sch/aftonposten_title_prefs dataset.
 It achieves the following results on the evaluation set:
-- Loss: 0.7655
-- Rewards/chosen: -0.0166
-- Rewards/rejected: -0.0295
-- Rewards/accuracies: 0.5071
-- Rewards/margins: 0.0129
-- Logps/rejected: -37.5657
-- Logps/chosen: -34.0622
-- Logits/rejected: -2.2285
-- Logits/chosen: -2.2333
+- Loss: 0.7598
+- Rewards/chosen: -0.0165
+- Rewards/rejected: -0.0300
+- Rewards/accuracies: 0.5307
+- Rewards/margins: 0.0135
+- Logps/rejected: -37.5666
+- Logps/chosen: -34.0620
+- Logits/rejected: -2.2283
+- Logits/chosen: -2.2332
 
 ## Model description
 

all_results.json

@@ -1,5 +1,18 @@
 {
     "epoch": 1.0,
+    "eval_logits/chosen": -2.2331926822662354,
+    "eval_logits/rejected": -2.2283315658569336,
+    "eval_logps/chosen": -34.061973571777344,
+    "eval_logps/rejected": -37.566593170166016,
+    "eval_loss": 0.7597638368606567,
+    "eval_rewards/accuracies": 0.5307309031486511,
+    "eval_rewards/chosen": -0.016451681032776833,
+    "eval_rewards/margins": 0.013532036915421486,
+    "eval_rewards/rejected": -0.029983719810843468,
+    "eval_runtime": 145.5395,
+    "eval_samples": 343,
+    "eval_samples_per_second": 2.357,
+    "eval_steps_per_second": 0.295,
     "train_loss": 0.5965218513042896,
     "train_runtime": 3252.2803,
     "train_samples": 3079,

eval_results.json

@@ -1,16 +1,16 @@
 {
     "epoch": 1.0,
-    "eval_logits/chosen": -2.230694055557251,
-    "eval_logits/rejected": -2.225895643234253,
-    "eval_logps/chosen": -33.75852966308594,
-    "eval_logps/rejected": -37.27712631225586,
-    "eval_loss": 0.4946732223033905,
-    "eval_rewards/accuracies": 0.5365448594093323,
-    "eval_rewards/chosen": 0.16561271250247955,
-    "eval_rewards/margins": 0.02191758342087269,
-    "eval_rewards/rejected": 0.14369513094425201,
-    "eval_runtime": 145.7007,
+    "eval_logits/chosen": -2.2331926822662354,
+    "eval_logits/rejected": -2.2283315658569336,
+    "eval_logps/chosen": -34.061973571777344,
+    "eval_logps/rejected": -37.566593170166016,
+    "eval_loss": 0.7597638368606567,
+    "eval_rewards/accuracies": 0.5307309031486511,
+    "eval_rewards/chosen": -0.016451681032776833,
+    "eval_rewards/margins": 0.013532036915421486,
+    "eval_rewards/rejected": -0.029983719810843468,
+    "eval_runtime": 145.5395,
     "eval_samples": 343,
-    "eval_samples_per_second": 2.354,
+    "eval_samples_per_second": 2.357,
     "eval_steps_per_second": 0.295
 }