search

ryanxingql / powerqe

Empower the quality enhancement approaches for compressed images/videos.
49 stars 3 forks source link

Asking for pretrained model #3

Closed ZhangYuef closed 3 years ago

ZhangYuef commented 3 years ago

Hi, thanks for sharing! I am wondering if you could offer pretrained models for reproducing the results. If so, when could them be available?

ryanxingql commented 3 years ago

Hi Yes, I will release the pre-trained models soon. I'm still conducting training and IQA, since the DIV2K images are too big and thus the training and IQA are somehow time-consuming.

v1 provides the pre-trained models trained at down-sampled DIV2K images. You can obtain those models at Release. See also v1 Instruction.

ZhangYuef commented 3 years ago

Hi Yes, I will release the pre-trained models soon. I'm still conducting training and IQA, since the DIV2K images are too big and thus the training and IQA are somehow time-consuming.

v1 provides the pre-trained models trained at down-sampled DIV2K images. You can obtain those models at Release. See also v1 Instruction.

Thanks for instructions! Looking forward to the V2 pretrained models : )