qym7 / sparsediff Goto Github PK

When I run the command ‘python3 main.py general.name=test’, the console displays an error message:

Error executing job with overrides: ['general.name=test']
Traceback (most recent call last):
File "/Users/wzn/code/SparseDiff/sparse_diffusion/main.py", line 121, in main
dataset_infos.compute_input_dims(
File "/Users/wzn/code/SparseDiff/sparse_diffusion/datasets/abstract_dataset.py", line 204, in compute_input_dims
ex_extra_feat = extra_features(example_data)
File "/Users/wzn/code/SparseDiff/sparse_diffusion/diffusion/extra_features.py", line 52, in call
noisy_data = utils.densify_noisy_data(sparse_noisy_data)
AttributeError: module 'sparse_diffusion.utils' has no attribute 'densify_noisy_data'

And I didn't find that method in utils.py.
How should I handle it to ensure that the code can run properly?

Hi,

Congrats for this nice work and thanks for sharing your code.

I'm a bit confused with something you do in the computation of the of the Laplacian :

L = self.compute_laplacian(A, normalize=False)
mask_diag = 2 * L.shape[-1] * torch.eye(A.shape[-1]).type_as(L).unsqueeze(0)
mask_diag = mask_diag * (~mask.unsqueeze(1)) * (~mask.unsqueeze(2))
L = L * mask.unsqueeze(1) * mask.unsqueeze(2) + mask_diag

Could you explain me why you add this mask to the diagonal of the Laplacian ?

Best,
Antoine

Dear Author,

I hope this message finds you well. I am currently working with your SparseDiff project and have encountered a challenge that I believe requires your expertise. I am experiencing an HTTP 403 Forbidden error when attempting to download the QM9 dataset.

The error trace is as follows:

Error executing job with overrides: ['dataset=qm9']
Traceback (most recent call last):
...
urllib.error.HTTPError: HTTP Error 403: Forbidden
I have verified that the file paths are correct and have followed the code implementation you provided. Specifically, the error occurs when executing the following code snippet:

file_path = download_url(self.raw_url, self.raw_dir)
extract_zip(file_path, self.raw_dir)
os.unlink(file_path)
_ = download_url(self.raw_url2, self.raw_dir)
os.rename(
osp.join(self.raw_dir, "3195404"),
osp.join(self.raw_dir, "uncharacterized.txt"),
)