End of training

README.md

@@ -1,11 +1,13 @@
 ---
-license: apache-2.0
 library_name: peft
 tags:
+- alignment-handbook
 - trl
 - dpo
 - generated_from_trainer
 base_model: norallm/normistral-7b-warm
+datasets:
+- hugodk-sch/aftonposten_title_prefs
 model-index:
 - name: ap-normistral-7b-align-scan
   results: []
@@ -16,17 +18,17 @@ should probably proofread and complete it, then remove this comment. -->
 
 # ap-normistral-7b-align-scan
 
-This model is a fine-tuned version of [norallm/normistral-7b-warm](https://huggingface.co/norallm/normistral-7b-warm) on the None dataset.
+This model is a fine-tuned version of [data/ap-normistral-7b-sft-qlora](https://huggingface.co/data/ap-normistral-7b-sft-qlora) on the hugodk-sch/aftonposten_title_prefs dataset.
 It achieves the following results on the evaluation set:
-- Loss: 0.9534
-- Rewards/chosen: -0.0764
-- Rewards/rejected: -0.1230
-- Rewards/accuracies: 0.5714
-- Rewards/margins: 0.0466
-- Logps/rejected: -37.1962
-- Logps/chosen: -33.2069
-- Logits/rejected: 97.6898
-- Logits/chosen: 97.7152
+- Loss: 0.9545
+- Rewards/chosen: -0.0743
+- Rewards/rejected: -0.1195
+- Rewards/accuracies: 0.5889
+- Rewards/margins: 0.0453
+- Logps/rejected: -37.1618
+- Logps/chosen: -33.1857
+- Logits/rejected: 97.6826
+- Logits/chosen: 97.7083
 
 ## Model description
 

all_results.json

@@ -1,5 +1,18 @@
 {
     "epoch": 1.0,
+    "eval_logits/chosen": 97.70832824707031,
+    "eval_logits/rejected": 97.6825942993164,
+    "eval_logps/chosen": -33.185726165771484,
+    "eval_logps/rejected": -37.16183853149414,
+    "eval_loss": 0.9545474052429199,
+    "eval_rewards/accuracies": 0.5888704061508179,
+    "eval_rewards/chosen": -0.07425507158041,
+    "eval_rewards/margins": 0.045276571065187454,
+    "eval_rewards/rejected": -0.11953166127204895,
+    "eval_runtime": 103.7845,
+    "eval_samples": 343,
+    "eval_samples_per_second": 3.305,
+    "eval_steps_per_second": 0.414,
     "train_loss": 0.88636748450143,
     "train_runtime": 2554.6221,
     "train_samples": 3079,

eval_results.json

@@ -1,15 +1,15 @@
 {
     "epoch": 1.0,
-    "eval_logits/chosen": 98.60509490966797,
-    "eval_logits/rejected": 98.58561706542969,
-    "eval_logps/chosen": -32.669063568115234,
-    "eval_logps/rejected": -36.29020309448242,
-    "eval_loss": 0.9789820313453674,
-    "eval_rewards/accuracies": 0.5215947031974792,
-    "eval_rewards/chosen": -0.09035440534353256,
-    "eval_rewards/margins": 0.039116356521844864,
-    "eval_rewards/rejected": -0.12947076559066772,
-    "eval_runtime": 103.771,
+    "eval_logits/chosen": 97.70832824707031,
+    "eval_logits/rejected": 97.6825942993164,
+    "eval_logps/chosen": -33.185726165771484,
+    "eval_logps/rejected": -37.16183853149414,
+    "eval_loss": 0.9545474052429199,
+    "eval_rewards/accuracies": 0.5888704061508179,
+    "eval_rewards/chosen": -0.07425507158041,
+    "eval_rewards/margins": 0.045276571065187454,
+    "eval_rewards/rejected": -0.11953166127204895,
+    "eval_runtime": 103.7845,
     "eval_samples": 343,
     "eval_samples_per_second": 3.305,
     "eval_steps_per_second": 0.414