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: 0.6938
-- Rewards/chosen: 0.0001
-- Rewards/rejected: 0.0014
-- Rewards/accuracies: 0.4809
+- Loss: 0.6939
+- Rewards/chosen: -0.0002
+- Rewards/rejected: 0.0012
+- Rewards/accuracies: 0.4751
 - Rewards/margins: -0.0013
-- Logps/rejected: -37.5030
-- Logps/chosen: -34.0337
-- Logits/rejected: -2.2388
-- Logits/chosen: -2.2437
+- Logps/rejected: -37.5047
+- Logps/chosen: -34.0361
+- Logits/rejected: -2.2390
+- Logits/chosen: -2.2439
 
 ## Model description
 

all_results.json

@@ -1,5 +1,18 @@
 {
     "epoch": 1.0,
+    "eval_logits/chosen": -2.243914842605591,
+    "eval_logits/rejected": -2.239037275314331,
+    "eval_logps/chosen": -34.036094665527344,
+    "eval_logps/rejected": -37.504669189453125,
+    "eval_loss": 0.693877100944519,
+    "eval_rewards/accuracies": 0.47508305311203003,
+    "eval_rewards/chosen": -0.00015425789752043784,
+    "eval_rewards/margins": -0.0013494607992470264,
+    "eval_rewards/rejected": 0.0011952028144150972,
+    "eval_runtime": 145.559,
+    "eval_samples": 343,
+    "eval_samples_per_second": 2.356,
+    "eval_steps_per_second": 0.295,
     "train_loss": 0.6926175067951152,
     "train_runtime": 3253.8494,
     "train_samples": 3079,

eval_results.json

@@ -1,16 +1,16 @@
 {
     "epoch": 1.0,
-    "eval_logits/chosen": -2.2436320781707764,
-    "eval_logits/rejected": -2.2387471199035645,
-    "eval_logps/chosen": -34.02302932739258,
-    "eval_logps/rejected": -37.51007080078125,
-    "eval_loss": 0.6931225061416626,
-    "eval_rewards/accuracies": 0.5328072905540466,
-    "eval_rewards/chosen": 0.00011527122114785016,
-    "eval_rewards/margins": 4.97526052640751e-05,
-    "eval_rewards/rejected": 6.551862315973267e-05,
-    "eval_runtime": 145.6938,
+    "eval_logits/chosen": -2.243914842605591,
+    "eval_logits/rejected": -2.239037275314331,
+    "eval_logps/chosen": -34.036094665527344,
+    "eval_logps/rejected": -37.504669189453125,
+    "eval_loss": 0.693877100944519,
+    "eval_rewards/accuracies": 0.47508305311203003,
+    "eval_rewards/chosen": -0.00015425789752043784,
+    "eval_rewards/margins": -0.0013494607992470264,
+    "eval_rewards/rejected": 0.0011952028144150972,
+    "eval_runtime": 145.559,
     "eval_samples": 343,
-    "eval_samples_per_second": 2.354,
+    "eval_samples_per_second": 2.356,
     "eval_steps_per_second": 0.295
 }