-
### 🚀 The feature, motivation and pitch
Currently `aten.uniform_` fallbacks to `FallbackKernel: torch.ops.aten.uniform_.default`.
```python
import torch
def foo(x):
return x.uniform_(-1, 1…
-
I was wondering if there is any way to work with tensorflow backend while tensors are in graph mode and are not EagorTensors.
-
I'm enjoying the use of tensorly, especially when sped up with a torch backend. However, I'm wondering if there's support for using multiple GPU's when using the torch backend. I could not find any in…
-
### 🚀 The feature, motivation and pitch
Currently aten.exponential_ fallbacks to `FallbackKernel: torch.ops.aten.exponential_.default`.
```python
import torch
def foo(x):
return x.expon…
-
### 🐛 Describe the bug
I tried to build the pytorch from source for a CPU version on a Ubuntu OS
Steps I followed:
1. Created a Conda environment
2. Installed the dependancies `conda install …
-
### 🐛 Describe the bug
NameError: name 'sympy' is not defined, while I already installed sympy in my venv.
```sympy 1.11.1 py310h06a4308_0 ```
```
No sympy found
Trac…
-
### 🐛 Describe the bug
Testing Pytorch 2.0 compilation for detection models from torchvision. If I run a model with no_grad for the first time, it crashes with an error. But if I launch for the first…
-
For the following input
```
Nx.tensor([
[
0.33747746329154127,
0.8448716981481403,
0.8465694977489985,
0.1926030183919778,
…
-
Hi,
I have a NN model in ONNX (converted from PyTorch), which uses HardSigmoid, which apparently is not supported by the VPU plugin.
I read a bit about the extensibility mechanism and it appears i…
-
## Description
Seg fault on declaring the derivative material property name, retrieving the name of a variable which doesnt exist because a constant value was passed instead of variable
## How to …