Toby Roseman

Our support for Torch script is experimental. Conversion of your model works fine if you trace the torch model first. As a workaround, please trace your model before conversion: just...

I can reproduce this with coremltools 6.0b1. Before changing the platform tag in our wheel, we should see if we just compile with an older version (i.e. 2.5) of glibc.

I agree the error message could be more friendly. Looks like the Core ML Framework [does supports int64 as model input](https://github.com/apple/coremltools/blob/6adf95bbb48b34e40c28748f5f35c9f5fca47e29/mlmodel/format/FeatureTypes.proto#L213). Your code is not specifying `dtype` when creating `ct.TensorType`....

I can reproduce this issue. Looks like when `scripts/test.sh` is run without the `--test-package` parameter it tries to run unit tests for our dependencies. We're getting this error because [it...

We have [implemented enisum conversion for PyTorch](https://github.com/apple/coremltools/blob/5b68098dc03452741162f399e4337e67b7ede8d8/coremltools/converters/mil/frontend/torch/ops.py#L752). So I'm not sure what's going on here. @LeftAttention - Please give us complete steps to reproduce this issue. How are you using...

I'm going to close this issue, since we have not received steps to reproduce it.

@mgrachten - Thanks code to reproduce the problem. I can reproduce the problem (in coremltools 6.0b1). I'll reopen this issue.

I'm not able to reproduce this problem. I get the following TensorFlow error by just declaring your class: `AlreadyExistsError: Another metric with the same name already exists.`

I think you're using non-standard conda channels. When running step 3, I get a `PackagesNotFoundError`. If I follow your directions but install TensorFlow 2.6.0 using `pip`, I get the same...

Ok, so you're on an M1 machine and using tensorflow-metal, those are important details. > Did anyone succeed to reproduce it? Yes, using tensorflow-metal on an M1 machine, I can...