Four super-resolution models have been introduced: SR-Unet, RCAN, DFCAN and WDSR.
The changes that have been applied are:
engine:
init.py: introduce the DFCAN loss to the compile function in super-resolution section.
metrics.py: implement the DFCAN loss.
check_configuration.py: introduce the models' names to check they are super-resolution models.
models:
init.py: introduce the models' names and their architectures' initialization.
The architectures from all the models have been introduced: srunet.py, rcan.py, dfcan.py, wdsr.py
templates:
super_resolution:
The templates from all the models have been introduced: srunet_super-resolution.yaml, rcan_super-resolution.yaml, dfcan_super-resolution.yaml, wdsr_super-resolution.yaml
Also, due to the dependencies in wdsr.py, the following library should be introduced:
pip install tensorflow-addons==0.15.0
Four super-resolution models have been introduced: SR-Unet, RCAN, DFCAN and WDSR.
The changes that have been applied are:
Also, due to the dependencies in wdsr.py, the following library should be introduced:
pip install tensorflow-addons==0.15.0