Stardist error when create dataset obeject

[mdoaskndfweon]

input image size is 500500, patch size = 128 or 256 or 64, batch size is 1, but when create dataset object, it cause gpu crash for unknown reason, I tried backbone = UNet and grid size increase (8,8) but still failed, does some way can solved this, my image is too big? I had already divided 10001000 into 5*5, that's my parameter below and bug A clear and concise description of what the bug is.

To Reproduce Steps to reproduce the behavior:

1. Go to '...'
2. Click on '....' <img width="1623" alt="Screenshot 2023-03-31 at 9 42 43 PM" src="https://user

   -images.githubusercontent.com/88061206/229260122-720faa15-8715-45bf-b19a-355af5a22aea.png">

   

   always reach out GPU memory

3. Scroll down to '....'
4. See error

Expected behavior A clear and concise description of what you expected to happen.

Screenshots If applicable, add screenshots to help explain your problem.

Desktop (please complete the following information):

• OS: [e.g. iOS]
• Browser [e.g. chrome, safari]
• Version [e.g. 22]

Smartphone (please complete the following information):

• Device: [e.g. iPhone6]
• OS: [e.g. iOS8.1]
• Browser [e.g. stock browser, safari]
• Version [e.g. 22]

Additional context Add any other context about the problem here.

[mdoaskndfweon]

@esgomezm

[mdoaskndfweon]

image is 32-bit

[mdoaskndfweon]

hello, sorry for that confusing , I had already solve above those problem, but I have question relate my result for the nuclei segmentation , suppose that input image size is 1000,

this is the test result ,

for example , right hand side of the grapth is segmentation result, but when I try to predict new image without label,

the result are showed below

you see, totally differnt from the test data, but i want the format look like the same for the test data format.

my input block size x and y is 756, min overlap is256, When I decrease of block size or decrease of lin overlap, it show error and can't output the grapth

[mdoaskndfweon]

@mdoaskndfweon