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: 1.0120
-- Rewards/chosen: 0.0035
-- Rewards/rejected: 0.0155
-- Rewards/accuracies: 0.4726
-- Rewards/margins: -0.0120
-- Logps/rejected: -37.4994
-- Logps/chosen: -34.0306
-- Logits/rejected: -2.2388
-- Logits/chosen: -2.2437
+- Loss: 0.9991
+- Rewards/chosen: 0.0161
+- Rewards/rejected: 0.0150
+- Rewards/accuracies: 0.5365
+- Rewards/margins: 0.0010
+- Logps/rejected: -37.4999
+- Logps/chosen: -34.0167
+- Logits/rejected: -2.2391
+- Logits/chosen: -2.2439
 
 ## Model description
 

all_results.json

@@ -1,5 +1,18 @@
 {
     "epoch": 1.0,
+    "eval_logits/chosen": -2.2439348697662354,
+    "eval_logits/rejected": -2.2390594482421875,
+    "eval_logps/chosen": -34.016693115234375,
+    "eval_logps/rejected": -37.499900817871094,
+    "eval_loss": 0.9991074204444885,
+    "eval_rewards/accuracies": 0.5365448594093323,
+    "eval_rewards/chosen": 0.01607290655374527,
+    "eval_rewards/margins": 0.0010251295752823353,
+    "eval_rewards/rejected": 0.015047776512801647,
+    "eval_runtime": 145.7535,
+    "eval_samples": 343,
+    "eval_samples_per_second": 2.353,
+    "eval_steps_per_second": 0.295,
     "train_loss": 0.9887851405453373,
     "train_runtime": 3254.6365,
     "train_samples": 3079,

eval_results.json

@@ -1,16 +1,16 @@
 {
     "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_logits/chosen": -2.2439348697662354,
+    "eval_logits/rejected": -2.2390594482421875,
+    "eval_logps/chosen": -34.016693115234375,
+    "eval_logps/rejected": -37.499900817871094,
+    "eval_loss": 0.9991074204444885,
+    "eval_rewards/accuracies": 0.5365448594093323,
+    "eval_rewards/chosen": 0.01607290655374527,
+    "eval_rewards/margins": 0.0010251295752823353,
+    "eval_rewards/rejected": 0.015047776512801647,
+    "eval_runtime": 145.7535,
     "eval_samples": 343,
-    "eval_samples_per_second": 2.355,
+    "eval_samples_per_second": 2.353,
     "eval_steps_per_second": 0.295
 }