Comment Comment by sandslides Not sure. The only inference demos are colab notebooks. The models are approx 700mb each so I imagine it will run on modest gpu parent sandslides 3 years ago