Could you provide the pre-trained encoder?

[tangyuelm]

Dear Sir/Madam,

Thanks for the code. It is really helpful for me to understand the temporal action localization. However, I want to try from the extracting features step but cannot find the I3D model checkpoint pre-trained on Kinetics-400 in https://github.com/deepmind/kinetics-i3d for Pytorch. Could you provide the RGB and flow .pth checkpoints for your encoders?

Thank you very much.