Broken repo

[TBS-CL]

Thank you for the inspiration, but this code doesn't work. It's either not finished or tested.
There are typos in the handler execution file name, and even with minor fixes, every possible Runpod GPU type gives CUDA out of memory error, while the pytorch tries to occupy 100% of the GPU memory. Stackideas tweaks didn't help.

For these seeking a working upscaler on runpod workers, I recommend this repo: https://github.com/ashleykleynhans/runpod-worker-real-esrgan - works out of the box.