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: 5.9715
-- Rewards/chosen: 0.0004
-- Rewards/rejected: -0.0246
-- Rewards/accuracies: 0.5748
-- Rewards/margins: 0.0250
-- Logps/rejected: -34.8210
-- Logps/chosen: -31.2804
-- Logits/rejected: -2.8025
-- Logits/chosen: -2.8048
+- Loss: 6.0745
+- Rewards/chosen: -0.0067
+- Rewards/rejected: -0.0282
+- Rewards/accuracies: 0.5660
+- Rewards/margins: 0.0215
+- Logps/rejected: -34.8390
+- Logps/chosen: -31.3159
+- Logits/rejected: -2.8030
+- Logits/chosen: -2.8055
 
 ## Model description
 

all_results.json

@@ -1,5 +1,18 @@
 {
     "epoch": 1.0,
+    "eval_logits/chosen": -2.805530071258545,
+    "eval_logits/rejected": -2.802974224090576,
+    "eval_logps/chosen": -31.315858840942383,
+    "eval_logps/rejected": -34.838993072509766,
+    "eval_loss": 6.074533939361572,
+    "eval_rewards/accuracies": 0.5660299062728882,
+    "eval_rewards/chosen": -0.006680913269519806,
+    "eval_rewards/margins": 0.021486198529601097,
+    "eval_rewards/rejected": -0.0281671155244112,
+    "eval_runtime": 112.8101,
+    "eval_samples": 343,
+    "eval_samples_per_second": 3.041,
+    "eval_steps_per_second": 0.381,
     "train_loss": 4.931960544338474,
     "train_runtime": 2724.0851,
     "train_samples": 3079,

eval_results.json

@@ -1,16 +1,16 @@
 {
     "epoch": 1.0,
-    "eval_logits/chosen": -2.80399751663208,
-    "eval_logits/rejected": -2.801586151123047,
-    "eval_logps/chosen": -31.23209571838379,
-    "eval_logps/rejected": -34.70168685913086,
-    "eval_loss": 1.0830976963043213,
-    "eval_rewards/accuracies": 0.5776578187942505,
-    "eval_rewards/chosen": 0.025178169831633568,
-    "eval_rewards/margins": 0.026945210993289948,
-    "eval_rewards/rejected": -0.0017670434899628162,
-    "eval_runtime": 112.7053,
+    "eval_logits/chosen": -2.805530071258545,
+    "eval_logits/rejected": -2.802974224090576,
+    "eval_logps/chosen": -31.315858840942383,
+    "eval_logps/rejected": -34.838993072509766,
+    "eval_loss": 6.074533939361572,
+    "eval_rewards/accuracies": 0.5660299062728882,
+    "eval_rewards/chosen": -0.006680913269519806,
+    "eval_rewards/margins": 0.021486198529601097,
+    "eval_rewards/rejected": -0.0281671155244112,
+    "eval_runtime": 112.8101,
     "eval_samples": 343,
-    "eval_samples_per_second": 3.043,
-    "eval_steps_per_second": 0.382
+    "eval_samples_per_second": 3.041,
+    "eval_steps_per_second": 0.381
 }