8 days ago
Hi everyone,
I’m 13 years old and pretty experienced with ML and YOLO, but I’m completely new to Railway and deployment.
I want to deploy a Python + YOLO project on Railway and would appreciate some guidance from people who’ve done it before.
Things I’m trying to understand:
Best way to deploy a YOLO model on Railway
Recommended setup (FastAPI, Flask, etc.)
Managing model weights and dependencies
CPU/GPU support and limitations
Good project structure for deployment
Any example repos or tutorials
Would really appreciate any tips or advice. Thanks!
1 Replies
8 days ago
Hey!
First things first, Railway does not provides GPU so your performance is going to be impacted and you will need to use more lightweight models like the nano/small versions of YOLO.
For API setup you can choose whichever you like the most, googling a bit I see that yolo pairs better with FastAPI but it's ultimately up to you.
Weight management is a tricky one, maybe through git commits?
The full setup would be FastAPI + Dockerfile (running the ultralytics cpu bound image) + YOLO small/nano
