How to run bark on two GPU cards through configuration?

[BillLucky]

I have two 1070Ti GPU cards, each with 8G memory, but bark can't assign tasks to these two cards.

Tried configuration: os.environ["CUDA_VISIBLE_DEVICES"] = "0,1" doesn't work either, is there any way to support it, thank you.

The error message is as follows: OutOfMemoryError: CUDA out of memory. Tried to allocate 20.00 MiB (GPU 0; 7.93 GiB total capacity; 7.07 GiB already allocated; 21.44 MiB free; 7.11 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_CONF

[gkucsko]

right now you would have to run them in separate processes. like you could write a little tiny script and launch it twice once with each env var. for the oom error have a look at the low memory tutorial in the notebook tutorials sections of the repo, you have a few choices