Closed jrajagopal closed 4 years ago
Also, FYI, regarding your comment about using the t-sne function in sklearn. If you use a large dataset, it fails with a Segmentation Error/Core Dump. This is a known documented issue in sklearn.
try: ...: from tsne.bh_sne import BH_SNE ...: except SystemError: ...: pass ...: from tsne.bh_sne import BH_SNE
this is how you can import it, but it raises an error in run
method:
73 tsne = BH_SNE()
---> 74 Y = tsne.run(X, N, X.shape[1], d, perplexity, theta, seed)
75 return Y
76
tsne/bh_sne.pyx in tsne.bh_sne.BH_SNE.run (tsne/bh_sne.cpp:1590)()
NameError: name 'np' is not defined
@areshytko I had the same error, but it was resolved when I restarted my Jupyter notebook kernel.
https://github.com/alexisbcook/tsne is a fork that works out of the box for me.
Thanks guys!
Please help me with this issue. I am currently running Python 3.4 on Ubuntu 14.04 with 60 GB RAM. I installed tsne from source with command: sudo -H pip3 install git+https://github.com/danielfrg/tsne.git
1) While that was successful, when testing the import, I got a relative import error:
2) Also, I noticed something strange however. I tried to import BH_SNE from tsne.bh_sne. If I do it right off the bat in python, it fails, with the same error:
3) However, if I first try importing tsne or bh_sne, or even repeat the above command, it succeeds: