Gilad

Hi @BvB93, Thanks for the answer. Are there any thoughts about making such a type/class part of the standard (It may be helpful for statically checking conformance)?

Thanks @BvB93

If I might add a use case - some packages depend on non-python stuff. For example, cupy depends on cuda. It uses naming scheme like cupy-cuda12x. We pip compile our...

Will you be open to a PR that will apply the same fix as in unpack, to make the behavior the same?

Any plans to merge this? Anything I can do?

> In what situation would this change ever be used? The `WheelFile` class in this library is not part of any public API. Ok, then. Your choice, I just thought...

To reproduce, add the following to the file: keras/utils/torch_utils_test.py: ```python def test_save_load(self): import keras class M(keras.Model): def __init__(self,channels=10, **kwargs): super().__init__() self.sequence = torch.nn.Sequential( torch.nn.Conv2d(1, channels, kernel_size=(3, 3)), ) def call(self,...

Iti s correct that torch objects lack get_config. The comment above is the cause of the bug, not the solution to it. As stated there, Torch modules are automatically wrapped...

I also fail to run the mentioned code. Machine setup is: cuda 11.4, Tensorflow nightly build - tf_nightly-2.7.0.dev20210727, python 3.7.5, 4xRTX2080 Ti, nvidia driver 470.57.02 Running this code: ``` import...