GitLab

* Add basic model testing.

Also fixes flaky test

* Fix flake8

* Add Affinetransformation

Add Affinetransformation to superseed LinearTransformation

* Add test

* Add zero mean_vector in LinearTransformation and improved docs

* update

* minor fix

* minor fix2

* fixed flake8

* fix flake8

* fixed transpose syntax

* fixed shape of mean_vector in test

* fixed test

* print est cov and mean

* fixed flake8

* debug

* reduce num_samples

* debug

* fixed num_features

* fixed rtol for cov

* fix __repr__

* Update transforms.py

* Update test_transforms.py

* Update transforms.py

* fix flake8

* Update transforms.py

* Update transforms.py

* Update transforms.py

* Update transforms.py

* Changed dim of mean_vector to 1D, doc and removed .numpy () from format_string

* Restore test_linear_transformation()

* Update test_transforms.py

Signed-off-by: Michal Zientkiewicz <michalz@nvidia.com>

* initial commit

* fixed Python2 issue

* fixed naming incorrectness and Python2 compability

* fixed preparation of train folder

* removed detection dataset

* added docstring and repr

* moved import of scipy to make the import of torchvision independent of it

* improved conversion from class string to index

* removed support for other years than 2012

* removed accidentally added file

* moved emptying of split folder to avoid accidental deletion

* removed deletion of the images

* removed error conversion for Python2

* Aligned class indices with the indices identified by ImageFolder class

In reference to #729 added comments to clarify the naming and action of the layers performing downsampling in resnets.

* Update transform doc with random offset of padding due to pad_if_needed

Updating transform docs with the random offset of padding when pad_if_needed is set to True.

* Update docs

Raise an `IndexError` for `FakeData.__getitem__()` if the index would be out of range.

* Added dimensions in the comments

The update provides the dimensions of the processed data following the style of inceptionV3 implementation.

* Changed docs and comments

Updated doc with the argument `transform_input`. Modified comments to match inceptionV3 style.

Including the `transform_input` argument in the docs of inceptionV3

* Add Caltech101 and Caltech256

* Add information about default for target_type

* Fix docs

* Add function to download from Google Drive

* Add CelebA dataset

* Only import pandas when needed

* Addressing comments

* Remove trailing whitespace

* Replace torch.LongTensor with torch.as_tensor

* Add GoogLeNet (Inception v1)

* Fix missing padding

* Add missing ReLu to aux classifier

* Add Batch normalized version of GoogLeNet

* Use ceil_mode instead of padding and initialize weights using "xavier"

* Match BVLC GoogLeNet zero initialization of classifier

* Small cleanup

* use adaptive avg pool

* adjust network to match TensorFlow

* Update url of pre-trained model and add classification results on ImageNet

* Bugfix that improves performance by 1 point

Remove explicit Python interpreter name.

* Fixes for PyTorch version of tqdm

* Flake

* Flake fix

* randomresizedmods

* lint checks

* test to randomrescrop added

* updates

* tests updated

* tests updated

* upd

* updates

* Update torchvision/transforms/transforms.py
Co-Authored-By: surgan12 <33121121+surgan12@users.noreply.github.com>

* tests changed

* trvis

* travis

* fixes syntax

* ...

* flake fixes

* flake_fixes

* flake_fixes2

* flake_fixes

* flake_fixes2

* Modifying the comments of inceptionV3 dimensions

Modifying the comments of inceptionV3 dimensions to match the pytorch convention. Relevant (https://github.com/pytorch/vision/pull/719#pullrequestreview-203194302)

* Added Batch size in comment

* Update inception.py

* Updated inceptionV3 to accept different sized images (Adaptive avg pool)

The update allows inceptionV3 to process images larger or smaller than prescribed image size (299x299) using adaptive average pooling. Will be useful while finetuning or testing on different resolution images.

* Update inception.py

The update allows VGG to process images larger or smaller than prescribed imagenet size using adaptive average pooling. Will be useful while finetuning or testing on different resolution images. Similar to https://github.com/pytorch/vision/pull/643 and https://github.com/pytorch/vision/pull/672. I did not include adaptive avg pool in features or classifier block so that these predefined blocks can be used as it is.

The update allows Alexnet to process images larger or smaller than prescribed imagenet size using adaptive average pooling. Will be useful while finetuning or testing on different resolution images. Similar to https://github.com/pytorch/vision/pull/643 and https://github.com/pytorch/vision/pull/672. I did not include adaptive avg pool in features or classifier block so that these predefined blocks can be used as it is.

The tests were previously taking 2 minutes, not they take 4 seconds

* Update URL for EMNIST

Update the dataset and download URL for EMNIST to the url from the updated arxiv version.

* Update mnist.py

* Added size information for inception v3

Since inception v3 expects in contrast to the other models a size
of 299x299x3 I added a small note in the docstring to point that out
(addresses #560)

* inception size info in note box; fixed line length

* removed trailing whitespace

Test installed version of package

To test against the installed version of the package (which is
preferably since this can catch installation file inclusion bugs) we
have to deal with quirks of coverage.py. By default the ecosystem is
set up to do development coverage tests (e.g. in src, or by using `pip
install -e .` or `python setup.py develop`).

We want to test against the version installed, to do this we have to
find the install path which we'll then pass to the `--cov` arg added by
`pytest-cov`. To do this, we have to cd out of the current folder, and
import the installed version and get it's install path (if we don't cd
out, then we end up getting the existing directory in the cwd since by
default cwd is on `sys.path`)

Once we have the install path, we pass that to the `--cov` of pytest,
however we also want to rewrite the paths for codecov to pick them up on
the website, if they don't have to code coverage with local paths, they
don't register is properly.

In order to that, we have a .coveragerc file that has the contents:

```
[paths]
source =
    torchvision
    /**/site-packages/torchvision
```

This tells codecov to treat all paths with those prefixes as the same,
so anything like
`/home/travis/miniconda/envs/test-environment/lib/python2.7/site-packages/torchvision/models/__init__.py`
would be treated the same as
`torchvision/models/__init__.py` after running the coverage combination
command. The first path seems to be special, in that all subsequent
paths are rewritten to that one.

Once we collect coverage, we then run `coverage run` to rewrite the
installation paths in the coverage file, `.coverage`, to those expected
by codecov.

Phew. And that's it.

N.B: Whilst adding test/__init__.py does solve the issue, it also results in
PWD being added to the path, and then we'll test the development version
of the package, and not the installed version (see
https://docs.pytest.org/en/latest/goodpractices.html#tests-as-part-of-application-code)

* doc change

* doc changes

* lint_checks

* Update transforms.py