nat-d / gmvae Goto Github PK

Hellos, this is great work. Thanks!

I just have a question, how can I use my own dataset with this? I have a folder of images that I would like clustered.

Thanks!

Hi, @Nat-D

I have met this error at displaying spiral grapics.

ubuntu@ubuntu:/GMVAE/plot$ python plot_latent.py
Traceback (most recent call last):
File "plot_latent.py", line 2, in
import torchfile
ImportError: No module named torchfile
ubuntu@ubuntu:/GMVAE/plot$
ubuntu@ubuntu:~/GMVAE/plot$ luarocks install torchfile

Error: No results matching query were found.

@Nat-D
I was trying to reproduce the results of this work. The code(version2, svhn) works fine in the default setting.
But, as I try to change the value of the number of Monte-Carlo samples(opt.nMC) by setting the flag in run.sh , it throws the following error:

/home/jayant/torch/install/share/lua/5.1/nn/MV.lua:35: inputs must contain the same number of minibatches
stack traceback:
[C]: in function 'assert'
/home/jayant/torch/install/share/lua/5.1/nn/MV.lua:35: in function 'func'
...e/jayant/torch/install/share/lua/5.1/nngraph/gmodule.lua:345: in function 'neteval'
...e/jayant/torch/install/share/lua/5.1/nngraph/gmodule.lua:380: in function 'forward'
main.lua:226: in function 'opfunc'
/home/jayant/torch/install/share/lua/5.1/optim/adam.lua:37: in function </home/jayant/torch/install/share/lua/5.1/optim/adam.lua:24>
main.lua:312: in main chunk
[C]: in function 'dofile'
...yant/torch/install/lib/luarocks/rocks/trepl/scm-1/bin/th:150: in main chunk
[C]: at 0x00405d50

Hi, @Nat-D

The GMVAE shows better clustering performance than other "old machine learning methods"?

"old machine learning"= knn, kmeans, cmeans, kohonen, t-sne, neuarl gas, etc ...

Thanks in advances.