-
When trying to use the 'bm' command, I typed in the command wrongly as 'bmm 1' instead of 'bm 1'.
The command is wrong but the category number which was 1 is valid.
Howeverr, the error message shows …
-
Right now, it's not restored as maximized. Since I have a Mac it would be a lot easier if someone working on Windows could work on this.
See electron/main.ts where window.setBounds() is called.
ht…
-
## 🐛 Bug
The following program never terminates.
```
import torch
import torch.multiprocessing as mp
def foo():
x = torch.ones((2, 50, 10))
return torch.einsum('ijl,ikl->ijk', x…
-
Currently, `torch.einsum` is computed by performing tensor contractions from left to right and calling `torch.bmm` for each contraction, unless there are no dimensions to be summed out in which case `…
-
### Describe the bug
`Isaac-Lift-Cube-Franka-IK-Abs-v0` with robot based rotated 90 degree, step constant action in robot base frame
The Frame indicates the visualization of target pose
![frank…
-
## 📚 Documentation
In PyTorch 1.8.1, the code
```py
a = torch.rand((1,2,3))
b = torch.rand((4,5,6))
torch.bmm(a, b)
```
produces the error message
```
RuntimeError: Expected batch2_size…
-
My pytorch is 1.1.0.
_x.data.resize_(batchSize, inputSize, 1)_ in NCEAverage.py doesn't work.
_with torch.no_grad(): x.resize_(batchSize, inputSize, 1)_ still doesn't work.
I replaced it as _ x…
-
The ```generate_data``` function shuffles the labels even though the predictions are ordered since they are multiplied by the ```inv_order``` matrix.
![image](https://user-images.githubusercontent…
-
### 論文へのリンク
[[arXiv:1805.08318] Self-Attention Generative Adversarial Networks](https://arxiv.org/abs/1805.08318)
### 著者・所属機関
Han Zhang, Ian Goodfellow, Dimitris Metaxas, Augustus Odena
- …
-
图中红色框的算子与模拟器结果对比有很大异常。在模型中我们由于推理速度的原因将其中的torch.bmm()矩阵乘算子换成了F.conv2d(不是nn.conv2d),再由pth->onnx->rknn转换路线转换,rknn可视化如下:
其中MatMul下的rknn模型连板精度无异常,但是用动态卷积F.conv2d(将其中一个矩阵设为卷积权重与另一个矩阵进行运算)时候模拟器正常,但是连板精度…