v2.0

Release pre-trained image-based models over the DIV2K dataset, in addition to the video-based models over the LDVv2 dataset.

• DIV2K: includes QP=27/32/37/42 and QF=20/30/40/50 models.
• LDVv2: includes QP=37 models.

v1.0

Pre-trained models.