Toby Roseman

Several of [the versions pinned in our `test.pip`](https://github.com/apple/coremltools/blob/main/reqs/test.pip) are not going to work with Python 3.9. This is because these versions are set to work with the last versions of...

[A version of `h5py` is no longer pinned for Apple Silicon](https://github.com/apple/coremltools/blob/233286eba0a03beb3dc2e24f7133e4f161e81597/reqs/test.pip#L3).

@liam-murphy-zozo - Thanks for reporting the issue I can reproduce it. It looks like you're trying to add batching to your model. coremltools does not have built in batching support...

Thanks for the minimal example. I can reproduce this issue. This looks an issue with an optimization pass to reduce the number of transpose operations.

Here is an even simpler reproduction code: ```python import torch from torch import nn, Tensor import coremltools as ct class Net(nn.Module): def forward(self, x): d = torch.zeros(2, dtype=torch.float32) return (d,...

This is related to (possibly a duplicate of) #775. What do you mean when you say "pth converter"? Are you referring to [the unified convertor](https://apple.github.io/coremltools/source/coremltools.converters.convert.html)?

Thanks for the minimal example. I can reproduce the issue. This is not just a neural network backend issue. The following also produces a Core ML model with float output:...

Hi @olokobayusuf - I don't understand that you are saying. You have a fix for this issue? If so, please put up a PR. I can help you get it...

What is the output of `print(type(model))`? Relevant code section can be found [here](https://github.com/apple/coremltools/blob/973eae67f2f273a29e80a9b009987516a070a58b/coremltools/converters/mil/frontend/tensorflow2/load.py#L99).

Can you give us a minimal example to reproduce the problem?