Add ResNeXt [50, 101] to keras.applications

[innat]

System information.

TensorFlow version (you are using): 2.5
Are you willing to contribute it (Yes/No) : Yes

Describe the feature and the current behavior/state.

ResNeXt is a well-known classification model that is weirdly missing from the keras.applications. The idea that is presented here is great and I think it's fit for the model member in the keras package.

Will this change the current api? How?
Yes.

from tensorflow.keras.applications.resnext50 import ResNeXt50
from tensorflow.keras.applications.resnext101 import ResNeXt101

Who will benefit from this feature?
ML engineers and researcher who uses tf.keras.

Others
Others implementation: https://github.com/qubvel/classification_models

[rchao]

Thanks for opening the issue. Can you provide a paragraph as a proposal for justifying adding this model? If you can provide some context of how widely this is used or some citation, that would be helpful.

[gcunhase]

Any update on this?

[innat]

This model was introduced in 2016 and now it's 2022 and still, we don't have it in keras.application even though it's cited by 6697. Same as its too.

[gcunhase]

How can that be?

[gcunhase]

@innat have you tried using ResNext from here? I'm not able to get good accuracies: qubvel/classification_models#73 .

[innat]

I did use SE-ResNeXt most often. It worked well, comparable with DenseNet121.

[innat]

cc @taehoonlee
Mentioning Lee who has once made a PR on ResNeXt HERE, keras-team/keras#11203.

[gcunhase]

Thank you @innat !

[taehoonlee]

There are so many different types of deep learning architecture these days, (as I know) so François Chollet wants to add only those with special structural changes among them.

[innat]

@taehoonlee Is the decision still persist? The torchvision provides these models. Models like wide-resnet, shuffle-net, unfortunately are missing as well.

[taehoonlee]

You can ask the maintainers. Now I am out of the keras team.