park1200656

I have a question in the source code called modeling_llama.py

5

### System Info @ArthurZucker @gante path : "src/transformers/models/llama/modeling_llama.py" Line 85 and 232 of this code contains float32 as a constant. I think, it looks like a bug. Or is there...

모델 저장 이슈

5

안녕하세요. 올려주신 코드로 모델을 학습하고 저장하던 도중, 저장 관련 문의가 있어 이슈 남깁니다. 저는 A100 40gb 8장을 사용하고, 그에 아래처럼 num_proc를 8로 셋팅을 하였습니다. tokenized_datasets = datasets.map( tokenize_function, batched=True, num_proc=8,...

How can i convert trained blip2 model to upload to huggingface

1

I trained a retrieval model using the BLIP-2 model. I would like to upload this model to Hugging Face. Is it possible to convert it so that it can be...