Help with CUDA and Accessing DLLs

[Avaneesh-Ramesh-07]

Hi all,
I am getting the above error when I try to run the train.py script with the given dataset. I have followed all the steps in the README file and have reinstalled CUDA and the problem still persists. Any ideas on how I can get past this error?

[milesial]

Hi, do you see an error when running this?

import torch
m = torch.nn.Conv2d(16, 16, 3).cuda()
m(torch.randn(16, 16, 16, 16, device='cuda'))

If so, your CUDA install is broken or your machine is not supported.
I would also try to switch to Linux or docker

[Avaneesh-Ramesh-07]

Interestingly, that piece of code runs fine without error. I don't really know what the problem is in that aspect, but at the end of all the error messages, it throws the following error:

`Traceback (most recent call last):
File "D:------\Conda\envs---\lib\multiprocessing\pool.py", line 125, in worker
result = (True, func(*args, **kwds))
File "D:------\Pytorch-UNet\utils\data_loading.py", line 30, in unique_mask_values
return np.unique(mask)
File "<array_function internals>", line 5, in unique
File "D:------\Conda\envs---\lib\site-packages\numpy\lib\arraysetops.py", line 272, in unique
ret = _unique1d(ar, return_index, return_inverse, return_counts)
File "D:------\Conda\envs---\lib\site-packages\numpy\lib\arraysetops.py", line 335, in unique1d
mask = np.empty(aux.shape, dtype=np.bool)
numpy.core._exceptions._ArrayMemoryError: Unable to allocate 2.34 MiB for an array with shape (2455040,) and data type bool
"""

The above exception was the direct cause of the following exception:

Traceback (most recent call last):
File "D:------\Pytorch-UNet\train.py", line 212, in
train_model(
File "D:------\Pytorch-UNet\train.py", line 43, in train_model
dataset = CarvanaDataset(dir_img, dir_mask, img_scale)
File "D:------\Pytorch-UNet\utils\data_loading.py", line 117, in init
super().init(images_dir, mask_dir, scale, mask_suffix='_mask')
File "D:------\Pytorch-UNet\utils\data_loading.py", line 53, in init
unique = list(tqdm(
File "D:------\Conda\envs---\lib\site-packages\tqdm\std.py", line 1195, in iter
for obj in iterable:
File "D:------\Conda\envs---\lib\multiprocessing\pool.py", line 873, in next
raise value
numpy.core._exceptions._ArrayMemoryError: Unable to allocate 2.34 MiB for an array with shape (2455040,) and data type bool`

I'm really not sure how to resolve this memory error as well, do you have any tips for that? If not, then do you have an idea of how to get started running this code with linux or docker?

[milesial]

Seems like you're just running out of memory. I would try on another machine.
The README explains how to get started on linux or docker