Richard Hakim

[FEA] TSNE and UMAP: Allow input to be precomputed distance matrix

6

**Is your feature request related to a problem? Please describe.** Yes. Currently, TSNE and UMAP do not actually allow for a precomputed distance matrix to be used. This is despite...

[BUG] TSNE: make knn_graph argument supersede X argument

11

**Describe the solution you'd like** Currently, if a user wants to use a custom distance metric for TSNE or UMAP, they need to _both_ input the original features (X), as...

CP regression method: logistic & multinomial logistic regression

10

#### Is your feature request related to a problem? Please describe. I am trying to discover low rank features in tensor data (neural time series that have been expanded into...

High memory requirement for PCA denoising

7

After upgrading suite2p, I am seeing very high memory requirements for the PCA denoising step. I'm not sure how to remedy this. Input movies are 1536x512, and there are roughly...

Feature Request: Custom distance matrix input

3

FEATURE REQUEST: In https://github.com/CannyLab/tsne-cuda/issues/8, the possibility of using a custom NN matrix is discussed and noted to be 'easy' to implement. DavidMChan: " It would be easy to add the...

Slow indexing of folder tree

1

Using the iPython OSF object, it is very slow to walk through a tree structure of folders in order to index the data repo. ``` import osfclient client = osfclient.OSF()...

As is, the roll is left-aligned, or 'real-time'. This is great for some uses, but it would be great if there was an option to have a 'centered' option such...

Feature request: parallel processing implementation

2

Hi, love the package. Thank you for making it. I noticed that there is a block of code in the 'illustrations.py' file ([here](https://github.com/marmarelis/rolling-quantiles/blob/master/python/examples/illustration.py)) demonstrating some code that hasn't been implemented...

tensor.dtype.to_complex() crashes kernel after ~100 calls in ipython kernel

10

### 🐛 Describe the bug The following block of code consistently crashes an ipython kernel without an error traceback after ~100 iterations in the loop: ```python import torch for i...