RenYang-home/OpenDVC

question about "Training your own models"

WenBingo opened this issue · 5 comments

WenBingo commented

bpgenc -f 444 -m 9 im1.png -o im1_QP27.bpg -q 27
bpgdec im1_QP27.bpg -o im1_bpg444_QP27.png

in the " bimeo_septuplet " it has many 1m1.png , how do i to do ?

RenYang-home commented

To train your models, you have to compress im1.png in all subfolders in vimeo.

WenBingo commented

Thanks for your answear。
I just learning the direction of video compression , so I do not how to compress im1.png in all subfolders in vimeo.
Need to use the folder.npy to compress the im1.png?

RenYang-home commented

The folder.npy contains all subdirectories of every training sample (im1.png - im7.png), then every im1.png in all these subdirectories should be compressed by BPG to generate the training set.

Jiaaaaaqwe commented

how to utilize the folder.npy?

RenYang-home commented