Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

[Bug] Dataset Reproducibility #286

Open
TorchDSP opened this issue Feb 21, 2025 · 1 comment
Open

[Bug] Dataset Reproducibility #286

TorchDSP opened this issue Feb 21, 2025 · 1 comment
Assignees
@TorchDSP
Labels
status: in progress Currently being worked on type: bug Something isn't working
Milestone
v1.1.0

Comments

@TorchDSP
Copy link
Owner

Version

1.0.0 (Default)

System Information

OS: Ubuntu 22.04

Description

NewDataset given the same seed does not generate the same dataset.

How to Reproduce the Bug

create two NewDataset instances, check their __getitem__ returns the same values (it does not).
@TorchDSP TorchDSP added status: in progress Currently being worked on type: bug Something isn't working labels Feb 21, 2025
@TorchDSP TorchDSP self-assigned this Feb 21, 2025
@TorchDSP TorchDSP added this to the v1.1.0 milestone Feb 21, 2025
@TorchDSP
Copy link
Owner Author

Debug print statement shows the following code:

from torchsig.signals.builders.constellation import ConstellationSignalBuilder

calls get_distribution() several times creating random distributions which are never going to be seeded; this should never happen in an import, and should be tracked down and removed/put somewhere in an init function probably

Image

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees

@TorchDSP TorchDSP

Labels
status: in progress Currently being worked on type: bug Something isn't working
Projects
None yet
Milestone
v1.1.0
Development

No branches or pull requests
1 participant
@TorchDSP