-
I'd like to add some semantic segmentation models.
Possible candidates:
- [ ] [Wider or Deeper: Revisiting the ResNet Model for Visual Recognition](https://arxiv.org/abs/1611.10080)
- code: h…
-
How can I modify the model to add a convolution block before the first conv layer?
-
FX Graph:
```
def forward(self, primals_1, primals_8, tangents_1): …
-
- https://arxiv.org/abs/2107.13034
- 2021
機械学習アルゴリズムの有効性は、大量のデータから有用な特徴を抽出できるかどうかで決まる。
モデルやデータセットのサイズが大きくなるにつれて、大規模なデータセットを大幅に小さくしながらも性能の高いデータセットに圧縮するデータセット蒸留法は、学習効率や有用な特徴抽出の点で価値が高くなる。
この目的のために、…
e4exp updated
2 years ago
-
Hi,
Just wondering if there is a way to use Conv1D in a heapix data. That is, it can handle, resolution(nside) and pixel orientation (nested or ring).
Many Thanks
Wasim
-
I can't find a suitable 3d version of DCN/DCNv2 on the Internet. Could you update this PyTorch version with a 3D DCN with support for large images?
For example: (N, 3, 128, 128, 128)
-
When I ran the code, I got the error :
2 root error(s) found.
(0) Unknown: Failed to get convolution algorithm. This is probably because cuDNN failed to initialize, so try looking to see if a war…
-
Thanks for your good work!
Group conv is used in a lot of ConvNets.
Will you plan to support group conv in the future version of dcn?
-
**Is your feature request related to a problem? Please describe.**
Currently, I want to apply `ndimage.convolve` to my sparse matrix. As my matrix is too sparse, only convolution results in nonzero p…
-