* remove dtype kwarg from .to() calls in EvoNorm as it messed up script + trace combo * BatchNormAct2d always uses custom forward (cut & paste from original) instead of super().forward. Fixes #1176 * BlurPool groups==channels, no need to use input.dim[1] |
||
---|---|---|
.. | ||
data | ||
loss | ||
models | ||
optim | ||
scheduler | ||
utils | ||
__init__.py | ||
version.py |