A question about training

[QingZhuanya]

Thank you for the excellent project. May I ask why I got stuck during phase1 training?

[YTEP-ZHI]

Hi @QingZhuanya, thanks for your question. I think this issue is caused by the data loading and has something to do with the number of workers. You might need to try to set a different number of workers in your config to see how it works. For instance, set the following num_workers from 16 to 4:

Vista/configs/training/vista_phase2_stage1.yaml

Lines 257 to 258 in 6ff19b8

	num_workers: 16
	subsets:

[QingZhuanya]

Hi @QingZhuanya, thanks for your question. I think this issue is caused by the data loading and has something to do with the number of workers. You might need to try to set a different number of workers in your config to see how it works. For instance, set the following num_workers from 16 to 4:

Vista/configs/training/vista_phase2_stage1.yaml

Lines 257 to 258 in 6ff19b8

	num_workers: 16
	subsets:

Thanks for the answer, I changed it according to your method, but unfortunately it's still stuck. My environment is 8 A100. Is there any other method? Thank you.