Stephen Panaro

## ❓Question I see that the Linear 8-bit quantization supports per-channel scales. Is there any way to achieve something similar with palettization? I have a large weight tensor and being...

Util to expose intermediate tensors as outputs

2

## 🌱 Describe your Feature Request Sometimes when debugging a CoreML model it can be difficult to know where exactly inside of the model it is failing. I wrote a...

KVO on unmanaged Object @Persisted properties fails

5

### How frequently does the bug occur? All the time ### Description I have an Object class with many properties. In some cases, I make an unmanaged instance of that...

Same idea as [here](https://github.com/microsoft/TransformerCompression/blob/6b12cdee6ad51791d7c776b3a046bc408b9e77e9/src/slicegpt/layernorm_fusion.py#L83-L85). opt-125m is impacted by this.

Support multiple outputs in extract_submodel for >=iOS 16

3

Using the `extract_submodel` debugging_util to add additional outputs fails for deployment target >=iOS 16 when the model is converted from PyTorch. Running this script: ```python import coremltools as ct from...

Support tuple inputs in extract_submodel

2

In the `extract_submodel` debugging util, the input reachability check doesn't support ops that take list/tuple inputs e.g. concat, stack For example, this script: ```python import coremltools as ct from coremltools.converters.mil.debugging_utils...

Torch state buffer `copy_` fails to convert

1

## 🐞Describing the bug I'm trying to use the new State feature and I want to overwrite the entire state with a new value. I'm able to write a MIL...