insight face does not use local model even if the local path is specified with the exact same models

[hwang136]

I have already downloaded the model and included them directly inside the server app = FaceAnalysis(name="buffalo_l", root="/root", providers=['CPUExecutionProvider'])

but it still tried to access github.com/deepinsight/insightface/releases/download/v0.7/buffalo_l.zip to download that zip file

[hwang136]

@yingfeng @nttstar @nihui @mike-hearn This seems like a simple question. Does anyone know how to fix it.

[AbhinavPrabhakar050501]

Is there no solution for this yet?

[japel]

This works fine for me. It expects to provide a folder in which a folder models will contain a folder buffalo_l (e.g.) in which the onnx files (extracted from the zip) will be found.