Question reproducing rombodawg/Rombos-LLM-V2.5-Qwen-32b #2
Description
I'm trying to reproduce rombodawg/Rombos-LLM-V2.5-Qwen-32b
and have the following questions:
- Dataset: was all the training examples in rombodawg/Everything_Instruct used? or subset, if this is the wrong dataset could you point to the correct one please
- Training: Was the LoRA script Train_model_Lora_Tune used then mergelora.py
Any insight that would be reproduce this great model would be appreciated