lowerquality / gentle

gentle forced aligner
https://lowerquality.com/gentle/
MIT License
1.45k stars 295 forks source link

How to get Gentle to run on Render? #335

Open ncioffi1 opened 7 months ago

ncioffi1 commented 7 months ago

Hello,

I am building an application that uses Gentle. I got the service running locally on Docker on my own machine, and it does what I need.

However, when I try to deploy gentle via Docker on a web server, specifically Render.com, it doesn't work.

If I deploy Gentle via the public image (docker.io/lowerquality/gentle), it loads the front end, but crashes out any time I try to run a transcription.

If I deploy Gentle via the included Dockerfile, it fails both via Render, and my local machine. For your reference, when I deploy via the included Dockerfile, it tends to seem to error out at the package Nvidia-cuda-dev...

Big question: What exactly are the requirements to running gentle? Is it dependent on NVIDIA / a GPU? Is it possible to get this working on a CPU / Can you run it on a CPU?

Any help / insight would be appreciated - gentle works great for what I need but can't seem to figure this out. Thanks for reading!