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: 0.4937
-- Rewards/chosen: 0.1639
-- Rewards/rejected: 0.1391
-- Rewards/accuracies: 0.5361
-- Rewards/margins: 0.0248
-- Logps/rejected: -37.2848
-- Logps/chosen: -33.7614
-- Logits/rejected: -2.2261
-- Logits/chosen: -2.2309
+- Loss: 0.4947
+- Rewards/chosen: 0.1656
+- Rewards/rejected: 0.1437
+- Rewards/accuracies: 0.5365
+- Rewards/margins: 0.0219
+- Logps/rejected: -37.2771
+- Logps/chosen: -33.7585
+- Logits/rejected: -2.2259
+- Logits/chosen: -2.2307
 
 ## Model description
 

all_results.json

@@ -1,5 +1,18 @@
 {
     "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_samples": 343,
+    "eval_samples_per_second": 2.354,
+    "eval_steps_per_second": 0.295,
     "train_loss": 0.46439293204963983,
     "train_runtime": 3253.2697,
     "train_samples": 3079,

eval_results.json

@@ -1,15 +1,15 @@
 {
     "epoch": 1.0,
-    "eval_logits/chosen": -2.226956844329834,
-    "eval_logits/rejected": -2.2221333980560303,
-    "eval_logps/chosen": -34.03876876831055,
-    "eval_logps/rejected": -37.55620193481445,
-    "eval_loss": 0.6902801394462585,
-    "eval_rewards/accuracies": 0.5519103407859802,
-    "eval_rewards/chosen": -0.002528026234358549,
-    "eval_rewards/margins": 0.021220751106739044,
-    "eval_rewards/rejected": -0.02374877780675888,
-    "eval_runtime": 145.6815,
+    "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_samples": 343,
     "eval_samples_per_second": 2.354,
     "eval_steps_per_second": 0.295