CoreML partitioner errors out with 8-bit integer types

🐛 Describe the bug

When partitioning a model with Core ML that contains int8 or uint8 into abs, the partitioner will error out with a KeyError. The backend should ideally either correctly handle it (if supported) or not partition it.

Repro:

import torch

from executorch.backends.apple.coreml.partition import CoreMLPartitioner
from executorch.exir import to_edge_transform_and_lower
from executorch.extension.pybindings.portable_lib import _load_for_executorch_from_buffer

class Model(torch.nn.Module):
    def forward(self, x):
        return torch.abs(x)

model = Model()
inputs = (torch.randint(0, 255, (1,10)).to(torch.uint8),)
lowered = to_edge_transform_and_lower(
    torch.export.export(model, inputs),
    partitioner=[CoreMLPartitioner()],
).to_executorch()

et_model = _load_for_executorch_from_buffer(lowered.buffer)
et_model([*inputs])

Output:

File [~/miniconda3/envs/executorch/lib/python3.10/site-packages/coremltools/converters/mil/frontend/torch/exir_utils.py:60](http://localhost:8888/lab/tree/~/miniconda3/envs/executorch/lib/python3.10/site-packages/coremltools/converters/mil/frontend/torch/exir_utils.py#line=59), in _construct_ct_tensor_type_from_torch(name, tensor, symbol_name_to_ct_range_dim)
     55 def _construct_ct_tensor_type_from_torch(
     56     name: str,
     57     tensor: torch.Tensor,
     58     symbol_name_to_ct_range_dim: Dict[str, RangeDim],
     59 ) -> TensorType:
---> 60     coreml_dtype = TORCH_DTYPE_TO_MIL_DTYPE[tensor.dtype]
     61     if coreml_dtype == types.int16:
     62         coreml_dtype = types.int32

KeyError: torch.uint8

Versions

Nightly

cc @kimishpatel @YifanShenSZ @cymbalrush @metascroy

Start with CoreMLPartitioner in executorch.backends.apple.coreml.partition and run the provided uint8 abs reproduction to trace the KeyError. Done means the backend either handles supported int8 and uint8 inputs or leaves unsupported inputs unpartitioned, with coverage for the reported case.