README.md

metadata

license: mit

Visual Product Recognition Challenge

The trained models for the competition. The training code for the models can be found in HCA97/Product-Recognition.

How to use it?

You need to install open_clip library.

pip install open_clip

Example of loading the model:

model = open_clip.create_model_and_transforms('ViT-H-14', None)[0].visual
model.load_state_dict(th.load('path to model'))
model.half()
model.eval()