Toby Roseman
Toby Roseman
How much different are the values from the Float16 model? Are they completely wrong? I think what might be happening here is that the `neuralnetwork` model type is being used...
@marius-verdier - in order to move forward here, we need a minimal self-containded example, one which doesn't require loading a model from an untrusted 3rd party (i.e. `denoiser`).
Can you add an overview of this change to the description of the pull request? Also please add unit tests for this new functionality.
Thanks @Cyanosite for this pull request. The change looks good to me. CI: https://gitlab.com/coremltools1/coremltools/-/pipelines/1288882002
The test pipeline isn't going to work with Python 3.12. We'll need to update all of our test dependencies to version that support Python 3.12. I recommend you stick with...
Updated CI: https://gitlab.com/coremltools1/coremltools/-/pipelines/1289337174
Updated CI: https://gitlab.com/coremltools1/coremltools/-/pipelines/1290878130
I can reproduce this issue. The error is actually coming when the second saved PyTorch model is being loaded. I suspect this issue is the result of some of the...
1 - Try passing `compute_precision=coremltools.precision.FLOAT16` to `coremltools.convert`. 2 - I don't understand the issue here. Do you want your model accept images or different sizes?
That should be the right wheel for a Python 3.10 environment on an M3. Assuming you're not using rosetta (i.e. you're running natively). However that's not the latest version of...