Distributed code cannot run about paella HOT 3 OPEN

and cannot run the run.sh. I am new in slurm command, not know how to print error message if u use srun to start the process.

from paella.

The problem seems with the dist_file here https://github.com/dome272/Paella/blob/main/src_distributed/utils.py#L92

Try changing the path to some directory that exists.

In more recent runs I always do:
init_method=f"file://{os.getcwd()}/dist_file"

from paella.

Thank you so much. I can run now but have one more question. How much GPU memory does Paella need? I tried to run Paella on LSUNChurches dataset, and only use byt5 model(not using clip model), but I can only run batchsize 1 on V100.

from paella.