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: []
@@ -16,15 +18,15 @@ 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: 0.6820
-- Rewards/chosen: -0.2813
-- Rewards/rejected: -0.4578
-- Rewards/accuracies: 0.5743
-- Rewards/margins: 0.1765
-- Logps/rejected: -38.0889
-- Logps/chosen: -34.3862
+- Loss: 0.6833
+- Rewards/chosen: -0.2814
+- Rewards/rejected: -0.4540
+- Rewards/accuracies: 0.6009
+- Rewards/margins: 0.1726
+- Logps/rejected: -38.0842
+- Logps/chosen: -34.3863
 - Logits/rejected: -2.1265
 - Logits/chosen: -2.1312
 

all_results.json

@@ -1,18 +1,18 @@
 {
     "epoch": 4.0,
-    "eval_logits/chosen": -2.22914457321167,
-    "eval_logits/rejected": -2.224289894104004,
-    "eval_logps/chosen": -34.094017028808594,
-    "eval_logps/rejected": -37.60322189331055,
-    "eval_loss": 0.6942082643508911,
-    "eval_rewards/accuracies": 0.5099667906761169,
-    "eval_rewards/chosen": -0.047573190182447433,
-    "eval_rewards/margins": 0.021706854924559593,
-    "eval_rewards/rejected": -0.06928005069494247,
-    "eval_runtime": 145.4732,
+    "eval_logits/chosen": -2.1312131881713867,
+    "eval_logits/rejected": -2.1264569759368896,
+    "eval_logps/chosen": -34.38629150390625,
+    "eval_logps/rejected": -38.08415985107422,
+    "eval_loss": 0.683270275592804,
+    "eval_rewards/accuracies": 0.6009136438369751,
+    "eval_rewards/chosen": -0.2813924252986908,
+    "eval_rewards/margins": 0.1726410835981369,
+    "eval_rewards/rejected": -0.4540335237979889,
+    "eval_runtime": 145.7759,
     "eval_samples": 343,
-    "eval_samples_per_second": 2.358,
-    "eval_steps_per_second": 0.296,
+    "eval_samples_per_second": 2.353,
+    "eval_steps_per_second": 0.295,
     "train_loss": 0.20605388523696305,
     "train_runtime": 10804.1609,
     "train_samples": 3079,

eval_results.json

@@ -1,16 +1,16 @@
 {
-    "epoch": 1.0,
-    "eval_logits/chosen": -2.22914457321167,
-    "eval_logits/rejected": -2.224289894104004,
-    "eval_logps/chosen": -34.094017028808594,
-    "eval_logps/rejected": -37.60322189331055,
-    "eval_loss": 0.6942082643508911,
-    "eval_rewards/accuracies": 0.5099667906761169,
-    "eval_rewards/chosen": -0.047573190182447433,
-    "eval_rewards/margins": 0.021706854924559593,
-    "eval_rewards/rejected": -0.06928005069494247,
-    "eval_runtime": 145.4732,
+    "epoch": 4.0,
+    "eval_logits/chosen": -2.1312131881713867,
+    "eval_logits/rejected": -2.1264569759368896,
+    "eval_logps/chosen": -34.38629150390625,
+    "eval_logps/rejected": -38.08415985107422,
+    "eval_loss": 0.683270275592804,
+    "eval_rewards/accuracies": 0.6009136438369751,
+    "eval_rewards/chosen": -0.2813924252986908,
+    "eval_rewards/margins": 0.1726410835981369,
+    "eval_rewards/rejected": -0.4540335237979889,
+    "eval_runtime": 145.7759,
     "eval_samples": 343,
-    "eval_samples_per_second": 2.358,
-    "eval_steps_per_second": 0.296
+    "eval_samples_per_second": 2.353,
+    "eval_steps_per_second": 0.295
 }