Devin Yang
Devin Yang
Same issue and this `All convolutions use padding mode “same”, and except for the last layer are followed by leaky ReLU` in original paper should mean only last conv use...
Hi @bluesky314 , you could call ```python from torchtoolbox.nn import conv dc = conv.DeformConv2d(modulation=True) ``` to use this feature.
I’ll check it out later.
Hi @pinouchon ,I read the paper again, and it didn't say that evo_norm could work with 1d conv layer, so I may not provid 1d version of this. But I...
Hi, try use this: `from torchtoolbox.transform.autoaugment import ImageNetPolicy`
I don't put `ImageNetPolicy` into `__init__` of transform. So you need to import it from `autoaugment` module. I've tried this way work for me. ```python from torchtoolbox.transform.autoaugment import ImageNetPolicy ```...
For now, I want to use Gluon-cv style to write our test module but will use pytest as test runner. CI maybe use Jenkins+Github cause we need test our datasets...
It's almost same thing. /dog
It's a good idea to compare with Pillow-SIMD.