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: 0.9998
-- Rewards/chosen: 0.0135
-- Rewards/rejected: 0.0132
-- Rewards/accuracies: 0.4904
-- Rewards/margins: 0.0004
-- Logps/rejected: -37.5002
-- Logps/chosen: -34.0177
-- Logits/rejected: -2.2386
-- Logits/chosen: -2.2435
+- Loss: 0.9960
+- Rewards/chosen: 0.0097
+- Rewards/rejected: 0.0053
+- Rewards/accuracies: 0.5141
+- Rewards/margins: 0.0044
+- Logps/rejected: -37.5100
+- Logps/chosen: -34.0224
+- Logits/rejected: -2.2385
+- Logits/chosen: -2.2434
 
 ## Model description
 

all_results.json

@@ -1,5 +1,18 @@
 {
     "epoch": 1.0,
+    "eval_logits/chosen": -2.2434139251708984,
+    "eval_logits/rejected": -2.2385354042053223,
+    "eval_logps/chosen": -34.0224494934082,
+    "eval_logps/rejected": -37.51002883911133,
+    "eval_loss": 0.9960440397262573,
+    "eval_rewards/accuracies": 0.5141196250915527,
+    "eval_rewards/chosen": 0.0096820630133152,
+    "eval_rewards/margins": 0.004409347660839558,
+    "eval_rewards/rejected": 0.005272718146443367,
+    "eval_runtime": 145.6562,
+    "eval_samples": 343,
+    "eval_samples_per_second": 2.355,
+    "eval_steps_per_second": 0.295,
     "train_loss": 0.9929689927534624,
     "train_runtime": 3255.1859,
     "train_samples": 3079,

eval_results.json

@@ -1,15 +1,15 @@
 {
     "epoch": 1.0,
-    "eval_logits/chosen": -2.243932008743286,
-    "eval_logits/rejected": -2.239053249359131,
-    "eval_logps/chosen": -34.03658676147461,
-    "eval_logps/rejected": -37.49867630004883,
-    "eval_loss": 1.0139278173446655,
-    "eval_rewards/accuracies": 0.4663621187210083,
-    "eval_rewards/chosen": -0.0014253630070015788,
-    "eval_rewards/margins": -0.013985722325742245,
-    "eval_rewards/rejected": 0.01256035827100277,
-    "eval_runtime": 145.6648,
+    "eval_logits/chosen": -2.2434139251708984,
+    "eval_logits/rejected": -2.2385354042053223,
+    "eval_logps/chosen": -34.0224494934082,
+    "eval_logps/rejected": -37.51002883911133,
+    "eval_loss": 0.9960440397262573,
+    "eval_rewards/accuracies": 0.5141196250915527,
+    "eval_rewards/chosen": 0.0096820630133152,
+    "eval_rewards/margins": 0.004409347660839558,
+    "eval_rewards/rejected": 0.005272718146443367,
+    "eval_runtime": 145.6562,
     "eval_samples": 343,
     "eval_samples_per_second": 2.355,
     "eval_steps_per_second": 0.295