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: 0.4785
-- Rewards/chosen: -0.2460
-- Rewards/rejected: -0.4418
-- Rewards/accuracies: 0.5631
-- Rewards/margins: 0.1958
-- Logps/rejected: -35.4345
-- Logps/chosen: -31.6925
-- Logits/rejected: -2.8303
-- Logits/chosen: -2.8337
+- Loss: 0.4787
+- Rewards/chosen: -0.2528
+- Rewards/rejected: -0.4471
+- Rewards/accuracies: 0.5602
+- Rewards/margins: 0.1943
+- Logps/rejected: -35.4433
+- Logps/chosen: -31.7037
+- Logits/rejected: -2.8306
+- Logits/chosen: -2.8338
 
 ## Model description
 

all_results.json

@@ -1,5 +1,18 @@
 {
     "epoch": 4.0,
+    "eval_logits/chosen": -2.8337740898132324,
+    "eval_logits/rejected": -2.830592393875122,
+    "eval_logps/chosen": -31.703718185424805,
+    "eval_logps/rejected": -35.44328689575195,
+    "eval_loss": 0.4786534011363983,
+    "eval_rewards/accuracies": 0.560215950012207,
+    "eval_rewards/chosen": -0.2527596652507782,
+    "eval_rewards/margins": 0.1943179965019226,
+    "eval_rewards/rejected": -0.4470776319503784,
+    "eval_runtime": 112.8079,
+    "eval_samples": 343,
+    "eval_samples_per_second": 3.041,
+    "eval_steps_per_second": 0.381,
     "train_loss": 0.15808662716057392,
     "train_runtime": 11226.5517,
     "train_samples": 3079,

eval_results.json

@@ -1,16 +1,16 @@
 {
     "epoch": 4.0,
-    "eval_logits/chosen": -2.7188398838043213,
-    "eval_logits/rejected": -2.7167177200317383,
-    "eval_logps/chosen": -99.52666473388672,
-    "eval_logps/rejected": -108.2430191040039,
-    "eval_loss": 0.690004289150238,
-    "eval_rewards/accuracies": 0.5485880374908447,
-    "eval_rewards/chosen": -0.6824421286582947,
-    "eval_rewards/margins": 0.05300649255514145,
-    "eval_rewards/rejected": -0.7354486584663391,
-    "eval_runtime": 112.821,
+    "eval_logits/chosen": -2.8337740898132324,
+    "eval_logits/rejected": -2.830592393875122,
+    "eval_logps/chosen": -31.703718185424805,
+    "eval_logps/rejected": -35.44328689575195,
+    "eval_loss": 0.4786534011363983,
+    "eval_rewards/accuracies": 0.560215950012207,
+    "eval_rewards/chosen": -0.2527596652507782,
+    "eval_rewards/margins": 0.1943179965019226,
+    "eval_rewards/rejected": -0.4470776319503784,
+    "eval_runtime": 112.8079,
     "eval_samples": 343,
-    "eval_samples_per_second": 3.04,
+    "eval_samples_per_second": 3.041,
     "eval_steps_per_second": 0.381
 }