End of training

README.md

@@ -1,11 +1,13 @@
 ---
 library_name: peft
 tags:
+- alignment-handbook
 - trl
 - dpo
-- alignment-handbook
 - 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: []
@@ -16,17 +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: 0.4892
-- Rewards/chosen: 0.0584
-- Rewards/rejected: 0.0040
-- Rewards/accuracies: 0.6092
-- Rewards/margins: 0.0545
-- Logps/rejected: -37.4968
-- Logps/chosen: -33.7423
-- Logits/rejected: -2.1019
-- Logits/chosen: -2.1066
+- Loss: 0.4886
+- Rewards/chosen: 0.0591
+- Rewards/rejected: 0.0021
+- Rewards/accuracies: 0.6184
+- Rewards/margins: 0.0569
+- Logps/rejected: -37.5059
+- Logps/chosen: -33.7391
+- Logits/rejected: -2.1016
+- Logits/chosen: -2.1063
 
 ## Model description
 

all_results.json

@@ -1,17 +1,17 @@
 {
     "epoch": 4.0,
-    "eval_logits/chosen": -2.2319114208221436,
-    "eval_logits/rejected": -2.2271130084991455,
-    "eval_logps/chosen": -33.75971984863281,
-    "eval_logps/rejected": -37.29259490966797,
-    "eval_loss": 0.49754369258880615,
-    "eval_rewards/accuracies": 0.5394518375396729,
-    "eval_rewards/chosen": 0.05496572330594063,
-    "eval_rewards/margins": 0.010160263627767563,
-    "eval_rewards/rejected": 0.044805459678173065,
-    "eval_runtime": 145.6941,
+    "eval_logits/chosen": -2.1062839031219482,
+    "eval_logits/rejected": -2.101640224456787,
+    "eval_logps/chosen": -33.739131927490234,
+    "eval_logps/rejected": -37.50592041015625,
+    "eval_loss": 0.48858171701431274,
+    "eval_rewards/accuracies": 0.6183555126190186,
+    "eval_rewards/chosen": 0.05908490717411041,
+    "eval_rewards/margins": 0.05694425478577614,
+    "eval_rewards/rejected": 0.002140657976269722,
+    "eval_runtime": 145.7832,
     "eval_samples": 343,
-    "eval_samples_per_second": 2.354,
+    "eval_samples_per_second": 2.353,
     "eval_steps_per_second": 0.295,
     "train_loss": 0.30868637050901143,
     "train_runtime": 10803.6,

eval_results.json

@@ -1,16 +1,16 @@
 {
-    "epoch": 1.0,
-    "eval_logits/chosen": -2.2319114208221436,
-    "eval_logits/rejected": -2.2271130084991455,
-    "eval_logps/chosen": -33.75971984863281,
-    "eval_logps/rejected": -37.29259490966797,
-    "eval_loss": 0.49754369258880615,
-    "eval_rewards/accuracies": 0.5394518375396729,
-    "eval_rewards/chosen": 0.05496572330594063,
-    "eval_rewards/margins": 0.010160263627767563,
-    "eval_rewards/rejected": 0.044805459678173065,
-    "eval_runtime": 145.6941,
+    "epoch": 4.0,
+    "eval_logits/chosen": -2.1062839031219482,
+    "eval_logits/rejected": -2.101640224456787,
+    "eval_logps/chosen": -33.739131927490234,
+    "eval_logps/rejected": -37.50592041015625,
+    "eval_loss": 0.48858171701431274,
+    "eval_rewards/accuracies": 0.6183555126190186,
+    "eval_rewards/chosen": 0.05908490717411041,
+    "eval_rewards/margins": 0.05694425478577614,
+    "eval_rewards/rejected": 0.002140657976269722,
+    "eval_runtime": 145.7832,
     "eval_samples": 343,
-    "eval_samples_per_second": 2.354,
+    "eval_samples_per_second": 2.353,
     "eval_steps_per_second": 0.295
 }