-
This might be a stupid question but I couldn't find a solution anywhere.
When I use gpu to run non-negative decompositions for a random tensor, it is much slower than using a cpu (for various size…
-
### 🐛 Describe the bug
I have gotten this error on several occasions, most recently using NanDetect in the training loop of fourier neural operator. I was trying to track down NaNs while trainin…
-
Currently, the `unfolding_dot_khatri_rao` function is implemented by extracting the component vectors and using the `multi_mode_dot` function in `tenalg`. However, this is very slow (for all backends)…
-
#### Is your feature request related to a problem? Please describe.
Dear Tensorly team,
I am using your package for my research, I am applying non-negative PARAFAC to my data. And in my dataset …
-
Several backends have now implemented the [NumPy random generator interface](https://numpy.org/doc/stable/reference/random/generator.html). With this, we can rethink how random state is passed through…
-
### 🐛 Describe the bug
I have a fresh virtual environment with pytorch installed from `pip install -r requirements.txt` which contain
```
tensorly
numpy
torch
torchvision
torchtyping
scipy
…
-
Decomposing large tensors can be time-consuming, and it would therefore be useful to have an easy-to-use interface for storing these decompositions to disc. I am happy to work on this once we decide t…
-
If nobody does a PR on this I'll get around to it in a few days.
-
#### Describe the solution you'd like
Implementations of traditional tensor completion algorithms like rank minimization.
#### Describe alternatives you've considered
Maybe https://github.com/isk…
-
#### Describe the bug
When feeding one dimensional factors in `tl.cp_tensor.cp_to_tensor`, a shape mismatch error message is prompted even though there is no shape mismatch.
#### Steps or Code to …