C3D model for EgoGesture video recognition
Python3 \ torch == 1.2 \ torchvision == 0.4.0 \ scipy \ sklearn \ tqdm \ torchsummary \ pillow==6.2.1 \ matplotlib \ opencv-python-headless \ pandas \ scikit-image
First use create_annotation.py to create annotation files. \ Then run C3D_fine_tune.py or R3D_fine_tune_all.py to train C3D or R3D model.
The C3D pretrained model can be download on this link. \ The R3D pretrained model is here. The originial repos is on this link.