-
Hello, I am trying to run a simplified "forward" passage of a neural network with GPU.
On CUDA/CuArray I have always the same, correct results for my output, but with oneAPI/oneArray, the first time …
-
I would like to ask that this section of the book describes the effect of Sigmoid function on data scaling, but your code uses RELU(), so can you explain it again?
-
In the current implementation ReLU is called as a function after each convolution layer.
The guided back-propagation tutorial I can find online are applying the hook function when detecting the ReLU …
-
## What
Let's fuse RELU with the proceeding operators across CONCAT.
Before
![image](https://user-images.githubusercontent.com/5449554/117227838-b0b2ef00-ae52-11eb-9d90-da7607d9ef71.png)
A…
-
I have read your codes about split-attention and I found that you use ReLU before split-attention.
https://github.com/zhanghang1989/ResNeSt/blob/76debaa9b9444742599d104609b8ee984b207332/resnest/torch…
-
Hello,
In your implementation of SpGAT,
there is this line:
edge_e = torch.exp(-self.leakyrelu(self.a.mm(edge_h).squeeze()))
However, I cannot understand why you added the minus sign in front …
-
## 🐛 Bug
I am observing a large divergence when using DeepLiftShap on a model with ReLU activations (or any type of activation) but not when using `torch.nn.Identity` instead. This is pretty puzzli…
-
### Expected behavior
expect to support qunatize leaky_relu when using relay.frontend.from_pytorch import a qat model
### Environment
tvm branch main for x86
### Steps to reproduce
…
-
### Describe the feature request
QDQ process includes symmetric quantization and asymmetric quantization by introducing the zero-offset. Many accelerators do not support zero-offset and thus symmetri…
-
I might be reading it incorrectly, but it looks like you don't apply the activation function to the final output layer? (should that be applied, in this context?)