Convolutional neural network model for video classification trained on the Kinetics dataset.
Inflated i3d network with inception backbone, weights transfered from tensorflow
I3D Nonlocal ResNets in Pytorch
#计算机科学#TensorFlow code for finetuning I3D model on UCF101.
#计算机科学#keras implementation of inflated 3d from Quo Vardis paper + weights
pytorch for i3d_nonlocal
Train I3D model on ucf101 or hmdb51 by tensorflow
I3D features extractor with resnet50 backbone
Inflate DenseNet and ResNet as per I3D with ImageNet weight transfer
Extract video features from raw videos using multiple GPUs. We support RAFT flow frames as well as S3D, I3D, R(2+1)D, VGGish, CLIP, and TIMM models.