Nurul Fhakri

icon company Quest Software Technology Systems Sdn Bh icon location Aspiring AI and ML Enthusiast.

Results 1 issues of Nurul Fhakri

Issue with Merging LoRA in Qwen 2.5 (3B) GRPO

5 comment

Hi, I tested Qwen 2.5 (3B) with GRPO on Kaggle, and after merging using 16-bit, it seems like the LoRA adaptations are not applied properly. The output lacks reasoning compared...