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: 0.9279
-- Rewards/chosen: -0.0837
-- Rewards/rejected: -0.1567
-- Rewards/accuracies: 0.6125
-- Rewards/margins: 0.0730
-- Logps/rejected: -35.4817
-- Logps/chosen: -31.7010
-- Logits/rejected: -2.8247
-- Logits/chosen: -2.8267
+- Loss: 0.9294
+- Rewards/chosen: -0.0900
+- Rewards/rejected: -0.1614
+- Rewards/accuracies: 0.6009
+- Rewards/margins: 0.0715
+- Logps/rejected: -35.5053
+- Logps/chosen: -31.7323
+- Logits/rejected: -2.8259
+- Logits/chosen: -2.8279
 
 ## Model description
 

all_results.json

@@ -1,5 +1,18 @@
 {
     "epoch": 1.0,
+    "eval_logits/chosen": -2.8279476165771484,
+    "eval_logits/rejected": -2.8258962631225586,
+    "eval_logps/chosen": -31.732280731201172,
+    "eval_logps/rejected": -35.50532150268555,
+    "eval_loss": 0.9294446110725403,
+    "eval_rewards/accuracies": 0.6009136438369751,
+    "eval_rewards/chosen": -0.08996524661779404,
+    "eval_rewards/margins": 0.07146809250116348,
+    "eval_rewards/rejected": -0.16143333911895752,
+    "eval_runtime": 112.8197,
+    "eval_samples": 343,
+    "eval_samples_per_second": 3.04,
+    "eval_steps_per_second": 0.381,
     "train_loss": 0.8011291454364727,
     "train_runtime": 2726.0249,
     "train_samples": 3079,

eval_results.json

@@ -1,16 +1,16 @@
 {
     "epoch": 1.0,
-    "eval_logits/chosen": -2.8154520988464355,
-    "eval_logits/rejected": -2.8132472038269043,
-    "eval_logps/chosen": -31.453943252563477,
-    "eval_logps/rejected": -35.06647872924805,
-    "eval_loss": 0.907057523727417,
-    "eval_rewards/accuracies": 0.5714285373687744,
-    "eval_rewards/chosen": -0.08574554324150085,
-    "eval_rewards/margins": 0.0984167754650116,
-    "eval_rewards/rejected": -0.18416231870651245,
-    "eval_runtime": 112.8077,
+    "eval_logits/chosen": -2.8279476165771484,
+    "eval_logits/rejected": -2.8258962631225586,
+    "eval_logps/chosen": -31.732280731201172,
+    "eval_logps/rejected": -35.50532150268555,
+    "eval_loss": 0.9294446110725403,
+    "eval_rewards/accuracies": 0.6009136438369751,
+    "eval_rewards/chosen": -0.08996524661779404,
+    "eval_rewards/margins": 0.07146809250116348,
+    "eval_rewards/rejected": -0.16143333911895752,
+    "eval_runtime": 112.8197,
     "eval_samples": 343,
-    "eval_samples_per_second": 3.041,
+    "eval_samples_per_second": 3.04,
     "eval_steps_per_second": 0.381
 }