Image dimension for ploting

[dcharua]

Hello, first let me congratulate for an excellent code. It's well documented and well written. All the other plots that are tried were not good enough to understand the data, this one is very clear.
Anyway to the point. I tried it with my own dataset, I'm able to plot in 3d and 2d, as well as to do the embedding, but I cannt overlay the images, I get the following error

Traceback (most recent call last):
File "main.py", line 134, in
main(config)
File "main.py", line 115, in main
im = OffsetImage(X_sample[i], zoom=0.1, cmap='gray')
File "/usr/local/lib/python3.5/dist-packages/matplotlib/offsetbox.py", line 1300, in init
self.set_data(arr)
File "/usr/local/lib/python3.5/dist-packages/matplotlib/offsetbox.py", line 1304, in set_data
self.image.set_data(self._data)
File "/usr/local/lib/python3.5/dist-packages/matplotlib/image.py", line 653, in set_data
raise TypeError("Invalid dimensions for image data")
TypeError: Invalid dimensions for image data

My X_sample is of shape (5000, 3, 32, 32)
where 5000 is the data size
X_sample[i].shape is (3, 32, 32)

What is the correct dimension?

Thank you for your time and code

[simplystart-from-begining]

I'm lucky to be able to focus on your work,l'm trying this code with my own datasets,but i'm not able to plot in 2d and 3d,i get error ''KeyError: 'train_imgs''
what's the correct code and command?thanks in advance

[dcharua]

Not sure what the question is, but I worked on this project a few months ago, you can see my code adaptation in this repo

[simplystart-from-begining]

thanks a lot to your reply.i have reviewed your code adaptation,what about your 'data4torch.h5' file?I hope get your help,thanks in advance.
[email protected]