End of training

README.md

@@ -1,10 +1,13 @@
 ---
 library_name: peft
 tags:
+- alignment-handbook
 - trl
 - dpo
 - generated_from_trainer
 base_model: NorLLM-AI/NorMistral-7B
+datasets:
+- hugodk-sch/aftonposten_title_prefs
 model-index:
 - name: norllm-ai-normistral-7b-align-scan
   results: []
@@ -15,17 +18,17 @@ should probably proofread and complete it, then remove this comment. -->
 
 # norllm-ai-normistral-7b-align-scan
 
-This model is a fine-tuned version of [NorLLM-AI/NorMistral-7B](https://huggingface.co/NorLLM-AI/NorMistral-7B) on the None dataset.
+This model is a fine-tuned version of [data/norllm-ai-normistral-7b-sft-qlora](https://huggingface.co/data/norllm-ai-normistral-7b-sft-qlora) on the hugodk-sch/aftonposten_title_prefs dataset.
 It achieves the following results on the evaluation set:
 - Loss: 0.4777
-- Rewards/chosen: 0.0978
-- Rewards/rejected: -0.0248
-- Rewards/accuracies: 0.5548
-- Rewards/margins: 0.1226
-- Logps/rejected: -34.7292
-- Logps/chosen: -31.1603
-- Logits/rejected: -2.8137
-- Logits/chosen: -2.8161
+- Rewards/chosen: 0.0944
+- Rewards/rejected: -0.0283
+- Rewards/accuracies: 0.5635
+- Rewards/margins: 0.1227
+- Logps/rejected: -34.7335
+- Logps/chosen: -31.1645
+- Logits/rejected: -2.8145
+- Logits/chosen: -2.8167
 
 ## Model description
 

all_results.json

@@ -1,5 +1,18 @@
 {
     "epoch": 1.0,
+    "eval_logits/chosen": -2.816662311553955,
+    "eval_logits/rejected": -2.814479351043701,
+    "eval_logps/chosen": -31.164457321166992,
+    "eval_logps/rejected": -34.7335090637207,
+    "eval_loss": 0.4776751399040222,
+    "eval_rewards/accuracies": 0.5635381937026978,
+    "eval_rewards/chosen": 0.09439504146575928,
+    "eval_rewards/margins": 0.12267693877220154,
+    "eval_rewards/rejected": -0.028281904757022858,
+    "eval_runtime": 112.8,
+    "eval_samples": 343,
+    "eval_samples_per_second": 3.041,
+    "eval_steps_per_second": 0.381,
     "train_loss": 0.3680442472556969,
     "train_runtime": 2722.5245,
     "train_samples": 3079,

eval_results.json

@@ -1,16 +1,16 @@
 {
     "epoch": 1.0,
-    "eval_logits/chosen": -2.817976474761963,
-    "eval_logits/rejected": -2.815365791320801,
-    "eval_logps/chosen": -31.056903839111328,
-    "eval_logps/rejected": -34.66963195800781,
-    "eval_loss": 0.49578404426574707,
-    "eval_rewards/accuracies": 0.5809800624847412,
-    "eval_rewards/chosen": 0.02255457080900669,
-    "eval_rewards/margins": 0.019702447578310966,
-    "eval_rewards/rejected": 0.0028521220665425062,
-    "eval_runtime": 112.8784,
+    "eval_logits/chosen": -2.816662311553955,
+    "eval_logits/rejected": -2.814479351043701,
+    "eval_logps/chosen": -31.164457321166992,
+    "eval_logps/rejected": -34.7335090637207,
+    "eval_loss": 0.4776751399040222,
+    "eval_rewards/accuracies": 0.5635381937026978,
+    "eval_rewards/chosen": 0.09439504146575928,
+    "eval_rewards/margins": 0.12267693877220154,
+    "eval_rewards/rejected": -0.028281904757022858,
+    "eval_runtime": 112.8,
     "eval_samples": 343,
-    "eval_samples_per_second": 3.039,
+    "eval_samples_per_second": 3.041,
     "eval_steps_per_second": 0.381
 }