forked from ecker/DL-Art-School
4328c2f713
This conforms my ConvGnLelu implementation with the generally accepted negative_slope=.2. I have no idea where I got .1. This will break backwards compatibility with some older models but will likely improve their performance when freshly trained. I did some auditing to find what these models might be, and I am not actively using any of them, so probably OK. |
||
|---|---|---|
| .. | ||
| byol | ||
| fixup_resnet | ||
| flownet2@db2b7899ea | ||
| glean | ||
| srflow | ||
| stylegan | ||
| switched_conv@cb520afd4d | ||
| tecogan | ||
| transformers | ||
| __init__.py | ||
| arch_util.py | ||
| discriminator_vgg_arch.py | ||
| feature_arch.py | ||
| ProgressiveSrg_arch.py | ||
| ResGen_arch.py | ||
| RRDBNet_arch.py | ||
| spinenet_arch.py | ||
| srg2_classic.py | ||
| SwitchedResidualGenerator_arch.py | ||