How to predict with batch size ?

[Vuong02011996]

Thannks Vladimir Iglovikov! Great works, but I only predict with single image by using model.predict_jsons(image), how to use batch size in there?

[PVGorshenin]

@Vuong02011996 I assume you are using model of class Model, defined in retinaface.predict_single. predict_json is the inference just in case of one image. Take a look on retinaface.inference. It is inference for the batch form. It has data preparation and loading under the hood. You have to pass arguments to script in the command line. It seems no explanation in the READ.me. But after taking a look on the code in inference.py things start to be clean.