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: 7.6485
-- Rewards/chosen: 0.0015
-- Rewards/rejected: -0.0116
-- Rewards/accuracies: 0.5390
-- Rewards/margins: 0.0131
-- Logps/rejected: -36.0246
-- Logps/chosen: -32.4356
-- Logits/rejected: 98.9867
-- Logits/chosen: 99.0023
+- Loss: 8.0178
+- Rewards/chosen: 0.0011
+- Rewards/rejected: 0.0010
+- Rewards/accuracies: 0.5220
+- Rewards/margins: 0.0001
+- Logps/rejected: -35.9615
+- Logps/chosen: -32.4377
+- Logits/rejected: 98.9752
+- Logits/chosen: 98.9888
 
 ## Model description
 

all_results.json

@@ -1,5 +1,18 @@
 {
     "epoch": 1.0,
+    "eval_logits/chosen": 98.98882293701172,
+    "eval_logits/rejected": 98.9751968383789,
+    "eval_logps/chosen": -32.437740325927734,
+    "eval_logps/rejected": -35.96149826049805,
+    "eval_loss": 8.017751693725586,
+    "eval_rewards/accuracies": 0.5220099687576294,
+    "eval_rewards/chosen": 0.0010875342413783073,
+    "eval_rewards/margins": 8.220935706049204e-05,
+    "eval_rewards/rejected": 0.0010053252335637808,
+    "eval_runtime": 103.7392,
+    "eval_samples": 343,
+    "eval_samples_per_second": 3.306,
+    "eval_steps_per_second": 0.415,
     "train_loss": 5.985443219271573,
     "train_runtime": 2557.1257,
     "train_samples": 3079,

eval_results.json

@@ -1,16 +1,16 @@
 {
     "epoch": 1.0,
-    "eval_logits/chosen": 98.9942626953125,
-    "eval_logits/rejected": 98.98635864257812,
-    "eval_logps/chosen": -32.3518180847168,
-    "eval_logps/rejected": -35.857688903808594,
-    "eval_loss": 2.3100497722625732,
-    "eval_rewards/accuracies": 0.5215947031974792,
-    "eval_rewards/chosen": 0.045678745955228806,
-    "eval_rewards/margins": -0.00873997900635004,
-    "eval_rewards/rejected": 0.05441872030496597,
-    "eval_runtime": 103.7492,
+    "eval_logits/chosen": 98.98882293701172,
+    "eval_logits/rejected": 98.9751968383789,
+    "eval_logps/chosen": -32.437740325927734,
+    "eval_logps/rejected": -35.96149826049805,
+    "eval_loss": 8.017751693725586,
+    "eval_rewards/accuracies": 0.5220099687576294,
+    "eval_rewards/chosen": 0.0010875342413783073,
+    "eval_rewards/margins": 8.220935706049204e-05,
+    "eval_rewards/rejected": 0.0010053252335637808,
+    "eval_runtime": 103.7392,
     "eval_samples": 343,
     "eval_samples_per_second": 3.306,
-    "eval_steps_per_second": 0.414
+    "eval_steps_per_second": 0.415
 }