NYU-DLSP20
NYU-DLSP20 copied to clipboard
[11-VAE.ipnb] TSNE fit_transform() happens a error with cuda
Following line happens a coversion error in the 16th cell. with CUDA environment. It causes to set cuda type tensor to TSNE,fit_transform().
E.append(TSNE(n_components=2).fit_transform(X[-1]))
I passed to add ,cpu() calling.
E.append(TSNE(n_components=2).fit_transform(X[-1].cpu()))
It looks good. @norihiro-ito, do you think you can send a pull request with such improvement? Otherwise @ritchieng can do it for you.
Let me double check. I'll push the fix and close the issue once I'm done.