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: 0.9400
-- Rewards/chosen: -0.0802
-- Rewards/rejected: -0.1425
-- Rewards/accuracies: 0.5623
-- Rewards/margins: 0.0624
-- Logps/rejected: -36.6792
-- Logps/chosen: -32.8440
-- Logits/rejected: 98.1900
-- Logits/chosen: 98.2178
+- Loss: 0.9524
+- Rewards/chosen: -0.0816
+- Rewards/rejected: -0.1313
+- Rewards/accuracies: 0.5303
+- Rewards/margins: 0.0497
+- Logps/rejected: -36.6233
+- Logps/chosen: -32.8513
+- Logits/rejected: 98.1886
+- Logits/chosen: 98.2171
 
 ## Model description
 

all_results.json

@@ -1,5 +1,18 @@
 {
     "epoch": 1.0,
+    "eval_logits/chosen": 98.21707153320312,
+    "eval_logits/rejected": 98.18864440917969,
+    "eval_logps/chosen": -32.851280212402344,
+    "eval_logps/rejected": -36.62327575683594,
+    "eval_loss": 0.9524344801902771,
+    "eval_rewards/accuracies": 0.530315637588501,
+    "eval_rewards/chosen": -0.08162158727645874,
+    "eval_rewards/margins": 0.04972809553146362,
+    "eval_rewards/rejected": -0.13134966790676117,
+    "eval_runtime": 103.7326,
+    "eval_samples": 343,
+    "eval_samples_per_second": 3.307,
+    "eval_steps_per_second": 0.415,
     "train_loss": 0.8039178501475941,
     "train_runtime": 2556.1517,
     "train_samples": 3079,

eval_results.json

@@ -1,15 +1,15 @@
 {
     "epoch": 1.0,
-    "eval_logits/chosen": 98.69615173339844,
-    "eval_logits/rejected": 98.67865753173828,
-    "eval_logps/chosen": -32.57830810546875,
-    "eval_logps/rejected": -36.26156234741211,
-    "eval_loss": 0.9490477442741394,
-    "eval_rewards/accuracies": 0.5689368844032288,
-    "eval_rewards/chosen": -0.06756443530321121,
-    "eval_rewards/margins": 0.07995220273733139,
-    "eval_rewards/rejected": -0.1475166231393814,
-    "eval_runtime": 103.72,
+    "eval_logits/chosen": 98.21707153320312,
+    "eval_logits/rejected": 98.18864440917969,
+    "eval_logps/chosen": -32.851280212402344,
+    "eval_logps/rejected": -36.62327575683594,
+    "eval_loss": 0.9524344801902771,
+    "eval_rewards/accuracies": 0.530315637588501,
+    "eval_rewards/chosen": -0.08162158727645874,
+    "eval_rewards/margins": 0.04972809553146362,
+    "eval_rewards/rejected": -0.13134966790676117,
+    "eval_runtime": 103.7326,
     "eval_samples": 343,
     "eval_samples_per_second": 3.307,
     "eval_steps_per_second": 0.415