Variable degree of freedom is not supported for 1D about fit-sne

[linqiaozhi]

I finally went ahead and implemented this in my fork. I tried it on a simple example, but not yet on MNIST. Can you please try it out and if it works as expected I will merge to the main repo.

[linqiaozhi]

Here's a simple test:

from fast_tsne import fast_tsne
from sklearn import datasets
import numpy as np
import matplotlib.pyplot as plt

SEED = 37

data = datasets.load_digits()
embed_df_1 = fast_tsne(data.data, df = 1, map_dims = 1, seed = SEED)
embed_df_05 = fast_tsne(data.data, df = .05, map_dims = 1, seed = SEED)

x = np.random.rand(data.data.shape[0]) 
fig, axs = plt.subplots(2)
axs[0].scatter(embed_df_1,x,  c=data.target)
axs[1].scatter(embed_df_05,x, c=data.target)
fig.show()

[dkobak]

Great! I will give it a try during next week.

[dkobak]

Sorry, was a bit overwhelmed with stuff. Hoping to try it out this week...

[dkobak]

I have finally gave it a try. I tried embedding full MNIST in 1D with various value of df. It seems to work as expected, so I guess you can go ahead and merge into master! Great that you found time to implement it.

One thing I was surprised to see, is that I did not observe any effect until I decreased df much below .5. For 2D, df=.5 was already producing a very strong effect.

And the same but rescaled horizontally:

[dkobak]

Another thing, is that digits don't split like they do in 2D, but rather the gaps between digits increase...

[dkobak]

Also, didn't in 2D the embedding typically grow in size with decreasing df? Here the size decreases when df decreases below 1...

[dkobak]

Hey, what do you think? Are you planning to merge this branch? Or do you want to investigate anything first?

[linqiaozhi]

Thanks for all the extensive testing, @dkobak. These are interesting differences between the 2D and 1D. I don't have an explanation for them, but I don't think it's due to a bug.

I went ahead and merged the fork.

[dkobak]

Cool. I will add my above tests to the example Python notebook.

[dkobak]

Added to the example notebook.