End of training

README.md

@@ -17,12 +17,12 @@ should probably proofread and complete it, then remove this comment. -->
 
 This model is a fine-tuned version of [deepseek-ai/deepseek-coder-1.3b-base](https://huggingface.co/deepseek-ai/deepseek-coder-1.3b-base) on the None dataset.
 It achieves the following results on the evaluation set:
-- Loss: 0.3991
-- Eval/rewards/chosen: -1.8560
-- Eval/logps/chosen: -121.7269
-- Eval/rewards/rejected: -11.4368
-- Eval/logps/rejected: -245.1979
-- Eval/rewards/margins: 9.5809
+- Loss: 0.4138
+- Eval/rewards/chosen: -2.1155
+- Eval/logps/chosen: -124.3213
+- Eval/rewards/rejected: -11.6851
+- Eval/logps/rejected: -247.6807
+- Eval/rewards/margins: 9.5696
 - Eval/kl: 0.0
 
 ## Model description
@@ -59,7 +59,7 @@ The following hyperparameters were used during training:
 
 | Training Loss | Epoch  | Step | Validation Loss |     |
 |:-------------:|:------:|:----:|:---------------:|:---:|
-| 0.1295        | 1.7058 | 100  | 0.3991          | 0.0 |
+| 0.1296        | 1.7058 | 100  | 0.4138          | 0.0 |
 
 
 ### Framework versions