Skip to content

Conversation

chsasank
Copy link
Contributor

@chsasank chsasank commented Jun 2, 2017

Added proper docstrings for all transforms.

WIP. Document:

  • Transforms
  • Datasets
  • Mean Std to be used for models
  • Create a PR to pytorch/pytorch with rst files modified

@fmassa
Copy link
Member

fmassa commented Jun 2, 2017

Awesome, thanks @chsasank !

@chsasank
Copy link
Contributor Author

chsasank commented Jun 2, 2017

@fmassa This PR is ready to be reviewed and merged.
Once merged I'll create another PR in pytorch/pytorch.

Do all the models use this normalisation: transforms.Normalize([0.485, 0.456, 0.406], [0.229, 0.224, 0.225])

chsasank added a commit to chsasank/pytorch that referenced this pull request Jun 2, 2017
@chsasank
Copy link
Contributor Author

chsasank commented Jun 2, 2017

Latest commit fixes #180.

@fmassa
Copy link
Member

fmassa commented Jun 2, 2017

The PR looks good, thanks a lot!

Yes, the mean / std should be the same for all models (except googlenet, but there is an extra argument to the model to adapt to it).

Unfortunately the README.rst from torchvision is not automatically updated, but it's ok for the moment I think.
At some point in future, it might be good to get the docs automatically generated from the python files (like it's done in pytorch).

@fmassa fmassa merged commit 432aa00 into pytorch:master Jun 2, 2017
chsasank added a commit to chsasank/pytorch that referenced this pull request Jun 2, 2017
@chsasank chsasank deleted the documentation branch June 2, 2017 16:05
soumith pushed a commit to pytorch/pytorch that referenced this pull request Jun 5, 2017
* Modify torchvision documentation following pytorch/vision#179

* Add new datasets to docs

* Fix wording in torch.datasets

* Small clarification
rajveerb pushed a commit to rajveerb/vision that referenced this pull request Nov 30, 2023
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants