George Linderman

Hi Asma, thanks for the interest in FIt-SNE! How exactly are you running it? Which of the wrappers are you using?

Thanks @jlmelville, yes, the problem is with the Seurat wrapper. As of [this](https://github.com/satijalab/seurat/pull/1127) PR, Seurat 3.0 is compatible with FIt-SNE 1.1.0. However, Seurat v2.X does not appear to be. I'm...

Good question; we discussed it a bit [here](https://github.com/KlugerLab/FIt-SNE/issues/34#issuecomment-419729622). At this time, we are not planning on supporting 3-dimensional embedding. Our interpolation scheme would work for 3 dimensions, and still be...

Thanks for posting the issue, @TonyX26. I think the problem is integer overflow. See how N*K gives a negative number here? `Going to allocate memory. N: 19113296, K: 150, N*K...

Same story. It is crashing [here](https://github.com/KlugerLab/FIt-SNE/blob/8468d5aceb9db8adbd6c7aaf31b312b28f032c71/src/tsne.cpp#L1588-L1639), and you can see the indices are integers, so they are overflowing. Try changing those to long int, particularly `n`. Although the algorithm should...

Thanks for the PR--and so sorry for the long delay. That's only for Windows though. We should have separate call-lines for Windows and non-Windows (i.e. OS X and Linux).

@rafalwoycicki Thanks for posting your fix! But yes, @dkobak is right, we need to fix the wrapper for Seurat so that it supports 1.2.1. See here https://github.com/satijalab/seurat/issues/2792. I plan to...

Thanks so much for making this happen @saketkc! Looks great. Please feel free to close this once they merge your changes.

I think it's a great idea...maybe can add it as an option? Also, in addition to the adaptive gradient descent, another reason why calling it with max_iter=1 in a loop...

I think you're referring to the R wrapper, correct? If so, then yes, it now optionally returns the KL divergence computed at every 50 iterations. I would have preferred that...