Text Generation
Transformers
PyTorch
mpt
Composer
MosaicML
llm-foundry
conversational
custom_code
text-generation-inference

How to use all gpu?

#19
by fuliu2023 - opened

config.init_device how to use all gpu。I config config.init_device = 'cuda:0',it report error:torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 98.00 MiB (GPU 0; 15.77 GiB total capacity; 14.74 GiB already allocated; 86.88 MiB free; 14.75 GiB reserved in total by PyTorch) If reserved memory is >> allocated memory try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CON

Sign up or log in to comment