Toby Roseman comments

Results 543 comments of


                                            Toby Roseman

Kwargs propagation from _get_model in quantize_weights to MLModel in _get_model

CI run: https://gitlab.com/coremltools1/coremltools/-/pipelines/1337526571

Avoid fp32 cast for Torch div operator

Please add a unit test to [test_torch_ops.py](https://github.com/apple/coremltools/blob/main/coremltools/converters/mil/frontend/torch/test/test_torch_ops.py) which fails without your fix but passes with your fix.

not create new symbol with immutable shape convolution. #1867

I'm not sure what you're asking. Can you please clarify the question? Also, please include all code necessary to reproduce the issue.

not create new symbol with immutable shape convolution. #1867

Can you share your complete code to reproduce this issue?

not create new symbol with immutable shape convolution. #1867

Are you not calling `ct.convert` on `conv_shape_debug`?

not create new symbol with immutable shape convolution. #1867

Can you share all the code we need to reproduce this issue? Including the `ct.convert` call and any thing else needed.

High Numerical Errors in Mish Activation with FLOAT16 Precision on Neural Engine

For security reasons, I am not able to download and run your network. Please create a minimal example to demonstrate the issue. Ideally some small amount of self contained code...

Failed to build the model execution plan using a model architecture file

@Skyline-23 that is a lot of code. Can you give us a more minimal example?

Failed to build the model execution plan using a model architecture file

@Skyline-23 - Can you give us a minimal example to reproduce this issue?

torch stft op conversion is broken for torch-2.7.0

This should be fixed in our next release.