RuntimeError when training on multiple GPUs

[LiuShiyu95]

I tried to train on multiple GPUs, but when reading the data, even if I set num_workers=0, I still get the error RuntimeError: unable to open shared memory object and I don't have root access, so I can't increase the openfile data. trainer = pl.Trainer(gpus=2, precision=32, callbacks=[logger]) As soon as I change gpus to 1, training works fine. Anyone have ideas?

[qingfengmingyue]

so am i

[kernelguardian]

Try increasing the shared memory size by running this on the terminal ulimit -n 16384

• check this issue and the last comment https://github.com/lllyasviel/ControlNet/issues/165#issuecomment-1464726313