microsoft

/

Florence-2-large-ft

like 293

I wrote a florence-2 multi-gpu finetuning project, especially for help understanding the process of finetuning the florence-2 model using multi-gpu training. The codes heavily references the [Hugging Face blog on fine-tuning Florence-2] (https://huggingface.co/blog/finetune-florence2) blog and their codes, but with the multi-GPU training code expanded(by chatgpt) and manual bug fixes.