Image Classifier Server

This project contains the image classifier server

Setup

python3 -m venv venv

SigLip2

podman pull metaloom/image-classifier-server:1.0.0
podman run --rm \
    --device nvidia.com/gpu=0 --name siglip2 \
    metaloom/image-classifier-server:1.0.0 

Usage

# Single image (multipart)
curl -X POST http://localhost:8000/classify -F "file=@photo.jpg" -F "top_k=3"

# Single image (base64 JSON)
curl -X POST http://localhost:8000/classify/base64 \
  -H "Content-Type: application/json" \
  -d '{"image": "'$(base64 -w0 photo.jpg)'", "top_k": 5}'

# Batch (multipart)
curl -X POST http://localhost:8000/classify/batch \
  -F "files=@img1.jpg" -F "files=@img2.jpg" -F "top_k=3"

Endpoint	Method	Input	Description
/classify	POST	Multipart file field	Single image via file upload
/classify/base64	POST	JSON {"image": "<b64>", "top_k": 5}	Single image via base64
/classify/batch	POST	Multipart files fields	Batch via file upload
/classify/batch/base64	POST	JSON {"images": ["<b64>", ...], "top_k": 5}	Batch via base64
/health	GET	—	Readiness check
/labels	GET	—	List loaded labels