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: 1.0000
 - Rewards/chosen: 0.0001
-- Rewards/rejected: 0.0001
-- Rewards/accuracies: 0.4988
-- Rewards/margins: 0.0000
-- Logps/rejected: -37.5053
-- Logps/chosen: -34.0201
+- Rewards/rejected: 0.0002
+- Rewards/accuracies: 0.5021
+- Rewards/margins: -0.0000
+- Logps/rejected: -37.5012
+- Logps/chosen: -34.0231
 - Logits/rejected: -2.2388
-- Logits/chosen: -2.2436
+- Logits/chosen: -2.2437
 
 ## Model description
 

all_results.json

@@ -1,5 +1,18 @@
 {
     "epoch": 1.0,
+    "eval_logits/chosen": -2.2436792850494385,
+    "eval_logits/rejected": -2.23879337310791,
+    "eval_logps/chosen": -34.02314376831055,
+    "eval_logps/rejected": -37.50123596191406,
+    "eval_loss": 1.0000426769256592,
+    "eval_rewards/accuracies": 0.5020764470100403,
+    "eval_rewards/chosen": 0.0001141041939263232,
+    "eval_rewards/margins": -3.9732771256240085e-05,
+    "eval_rewards/rejected": 0.00015383693971671164,
+    "eval_runtime": 145.4546,
+    "eval_samples": 343,
+    "eval_samples_per_second": 2.358,
+    "eval_steps_per_second": 0.296,
     "train_loss": 0.9998476387618425,
     "train_runtime": 3251.6099,
     "train_samples": 3079,

eval_results.json

@@ -1,16 +1,16 @@
 {
     "epoch": 1.0,
-    "eval_logits/chosen": -2.2436861991882324,
-    "eval_logits/rejected": -2.2388129234313965,
-    "eval_logps/chosen": -34.01886749267578,
-    "eval_logps/rejected": -37.50727081298828,
-    "eval_loss": 0.34474730491638184,
-    "eval_rewards/accuracies": 0.5161960124969482,
-    "eval_rewards/chosen": 0.01411630492657423,
-    "eval_rewards/margins": 0.005696756765246391,
-    "eval_rewards/rejected": 0.008419547230005264,
-    "eval_runtime": 145.8183,
+    "eval_logits/chosen": -2.2436792850494385,
+    "eval_logits/rejected": -2.23879337310791,
+    "eval_logps/chosen": -34.02314376831055,
+    "eval_logps/rejected": -37.50123596191406,
+    "eval_loss": 1.0000426769256592,
+    "eval_rewards/accuracies": 0.5020764470100403,
+    "eval_rewards/chosen": 0.0001141041939263232,
+    "eval_rewards/margins": -3.9732771256240085e-05,
+    "eval_rewards/rejected": 0.00015383693971671164,
+    "eval_runtime": 145.4546,
     "eval_samples": 343,
-    "eval_samples_per_second": 2.352,
-    "eval_steps_per_second": 0.295
+    "eval_samples_per_second": 2.358,
+    "eval_steps_per_second": 0.296
 }