Mohammed Innat

Here ([XNet-TF.Keras](https://gist.github.com/innat/2d2fd2e93b47dd5349095182eb984ef6)) is the **XNet** updated with `tf 2.x`. However, I'm using it just for casual stuff and so I only use `keras` official **EfficientNet** as an encoder. But other...

@chenmoneygithub > Currently users would need to write their own custom training loop to handle the gradient accumulation, which is not too hard, so we have not yet made this...

It would be great fit for https://keras.io/examples/keras_recipes/ cc. @fchollet

@fchollet I like to summarize the exact issue. **Background** Like, here I was using some `EfficientNet` models (same model but difrerent variants) to **train combinedly**. So, it's natual that, there...

Okay, it looks like if I rename the variablen ame as follows, it doesn't throw error. ```python for i in range(len(MODEL.weights)): MODEL.weights[i]._handle_name = str(i) + '/' + MODEL.weights[i].name 0/stem_conv/kernel:0 1/stem_bn/gamma:0...

ticket https://github.com/keras-team/keras-cv/issues/324

[This model](https://arxiv.org/abs/1611.05431) 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](https://github.com/keras-team/keras/issues/15267) too.

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

cc @taehoonlee Mentioning Lee who has once made a PR on ResNeXt HERE, https://github.com/keras-team/keras/pull/11203.

@taehoonlee Is the decision still persist? The `torchvision` provides [these](https://pytorch.org/vision/stable/models.html) models. Models like wide-resnet, shuffle-net, unfortunately are missing as well.