Closed WHlTE-N0lSE closed 2 years ago
@WHlTE-N0lSE The pretraining repo of maskfeat may need to be optimized. Currently, the speed of data loader perform not well. Because the hog features extraction occupy most of time.
The speed is acceptable, but what is the shell command of maskfeat pretraining?
@WHlTE-N0lSE We have alleviated the speed problem in maskfeat. The shell command can be found here https://github.com/mx-mark/VideoTransformer-pytorch#training
Request the pretraining command for maskfeat.This doesn't seem to be provided in the code. Thank you!