Richard Hakim

This issue still needs to be addressed.

THIS ISSUE STILL NEEDS TO BE ADDRESSED!

@cohenjer Thanks, I understand. The constrained CP that I see progress being made on looks awesome. Do you have any tips on how to attempt trying it myself? I'm not...

I wrote one that works well. In case y'all are interested: https://github.com/RichieHakim/tensor_regression/blob/main/multinomial_tensor_regression.py

Sure, though I have never worked with Tensorly-Torch. The method is intended to be a multinomial logistic version of CPRegression in Tensorly. What would need to be done to contribute...

Ok, great. Is there somewhere you guys chat about details of development for this kind of stuff? I'm a bit green and have questions on how to integrate it properly...

I'm interested in this as well. I believe this would allow for FIR filtfilt as well, which is extremely useful.

Having numerical parity with scipy filtfilt is of significant interest to me as well. I think the underlying algorithms are pretty different (DC I vs DC II I believe).

I just want to say thank you. I'm very impressed with everyone's work here and the speed with which it was done.

This repo remains the best tool for it's intended purpose. I use it for continuous frame seeking with a buffer ([here](https://github.com/RichieHakim/basic_neural_processing_modules/blob/main/bnpm/video.py#L266)). There are 1.9K projects that use decord as a...