instructions download inswapper_128.onnx (find it yourself) run convert.py todos still todo: port rest of insightface app to be batched also, add support for batched inference