Support for Add/Sub with int8, int16, uint8, uint16 in CPU Provider

[atodekangae]

Describe the issue

ONNX opset 14 supports Add and Sub for int8, int16, uint8, and uint16, but these do not seem to be supported by the default CPU provider in ONNX Runtime.

I have created a patch that adds support for these operations in the CPU provider. Can I just create a pull request to contribute this patch?

To reproduce

import onnx
import onnxruntime as ort

def run_model(typ):
    left = onnx.helper.make_tensor('left', typ, [1], [1])
    right = onnx.helper.make_tensor('right', typ, [1], [2])
    add_node = onnx.helper.make_node(
        'Add',
        inputs=['left', 'right'],
        outputs=['output']
    )
    output = onnx.helper.make_tensor_value_info('output', typ, [1])

    graph = onnx.helper.make_graph(
        nodes=[add_node],
        name='TestModel',
        inputs=[],
        outputs=[output],
        initializer=[left, right]
    )
    model = onnx.helper.make_model(graph, opset_imports=[onnx.helper.make_opsetid("", 14)])

    session = ort.InferenceSession(model.SerializeToString(), providers=['CPUExecutionProvider'])
    return session.run(None, {})

print(run_model(onnx.helper.TensorProto.INT32))  # good
print(run_model(onnx.helper.TensorProto.UINT8))  # [ONNXRuntimeError] : 9 : NOT_IMPLEMENTED : Could not find an implementation for Add(14) node with name ''

Urgency

No response

Platform

Linux

OS Version

Ubuntu 22.04

ONNX Runtime Installation

Released Package

ONNX Runtime Version or Commit ID

1.16.3

ONNX Runtime API

Python

Architecture

X64

Execution Provider

Default CPU

Execution Provider Library Version

No response

[carzh]

Yes, you can create PR to contribute the patch

[atodekangae]

Thank you for the reply. I created a pull request: https://github.com/microsoft/onnxruntime/pull/19244

[github-actions[bot]]

This issue has been automatically marked as stale due to inactivity and will be closed in 30 days if no further activity occurs. If further support is needed, please provide an update and/or more details.