/online_http_pytorch

Gunicorn aiohttp PyTorch, A concurrent HTTP server in order to inference using PyTorch.

Primary LanguagePython

Gunicorn aiohttp PyTorch

A concurrent HTTP server in order to inference using PyTorch.

If running in GPU server, it will allocate one GPU per worker(process), which is authorized by Gunicorn.

You can assign the number of worker in app.sh file (-w $num_worker)

run

sh app.sh