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

PyTorch Lightning tests shouldn't hit network #551

Open
bfirsh opened this issue Mar 8, 2021 · 4 comments
Open

PyTorch Lightning tests shouldn't hit network #551

bfirsh opened this issue Mar 8, 2021 · 4 comments
Labels
good first issue Good for newcomers help wanted Extra attention is needed type/chore Something not user-facing

Comments

@bfirsh
Copy link
Member

bfirsh commented Mar 8, 2021

Currently the PyTorch Lightning tests download mnist and train a real model. We shouldn't do that to just test that the callback works -- they should run a fake training process of some kind.
@bfirsh bfirsh added good first issue Good for newcomers help wanted Extra attention is needed type/chore Something not user-facing labels Mar 8, 2021
@bfirsh bfirsh mentioned this issue Mar 8, 2021
Closed
bfirsh referenced this issue in bfirsh/keepsake Mar 8, 2021
@bfirsh
Fix broken PyTorch Lightning test
afd04e2 
Closes #547
Ref #551

Signed-off-by: Ben Firshman <[email protected]>
@bfirsh bfirsh mentioned this issue Mar 8, 2021
Merged
bfirsh referenced this issue in bfirsh/keepsake Mar 8, 2021
@bfirsh
Fix broken PyTorch Lightning test
00dacb5 
Closes #547
Ref #551

Signed-off-by: Ben Firshman <[email protected]>
bfirsh added a commit that referenced this issue Mar 8, 2021
@bfirsh
Fix broken PyTorch Lightning test
95fd2b2 
Closes #547
Ref #551

Signed-off-by: Ben Firshman <[email protected]>
@gabrielmbmb
Copy link
Contributor

Maybe, some kind of fake data could be generated with NumPy or PyTorch instead of using a "real" dataset?

@bfirsh bfirsh mentioned this issue Mar 11, 2021
Merged
@bfirsh
Copy link
Member Author

bfirsh commented Mar 11, 2021

Yeah this doesn't need to do anything real.

bfirsh referenced this issue in bfirsh/keepsake Mar 11, 2021
@bfirsh
Upgrade torchvision
86fb7ab 
Stopgap fix for #551

Signed-off-by: Ben Firshman <[email protected]>
@bfirsh
Copy link
Member Author

bfirsh commented Mar 11, 2021

These files are pretty small. Could just include them in the repo. Ideally we'd have a test that didn't actually do anything on real data though. https://github.com/golbin/TensorFlow-MNIST/tree/master/mnist/data

@gabrielmbmb
Copy link
Contributor

A pytext fixture with dummy generated train data could be nice, as it could be used in future test cases easily.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees
No one assigned
Labels
good first issue Good for newcomers help wanted Extra attention is needed type/chore Something not user-facing
Projects
None yet
Milestone
No milestone
Development

No branches or pull requests
2 participants
@bfirsh @gabrielmbmb