[Errno 32] Broken pipe about lightweight-gan

[lucidrains]

@ramnoob hey thanks! i've added a --num-worker flag so you can specify the number of workers in 0.17.6

let me know if that works!

[ramnoob]

Thanks for the quick response!
The above phenomenon was occurring in the conda environment.
But then I started training in a pure Win10 CMD environment and it ran fine.
Training is expected to take a week.
After the training, I'll try ver0.17.6 in conda environment.

[ramnoob]

@lucidrains Hello!
I upgraded the version to 0.17.6 and ran it in the conda environment, but it doesn't seem to work.
Lightweight_gan --data datasets --image-size 512 --num-worker 1 I get the same [Errno 32] Broken pipe.
However, it works well in the WindowsCMD environment, so I am not troubled!

I think it would be better if a learning rate curve could be obtained at running or the end of the run, so please consider implementing it if you have enough capacity!