microsoft / onnxscript

ONNX Script enables developers to naturally author ONNX functions and models using a subset of Python.
https://onnxscript.ai/
MIT License
271 stars 53 forks source link

[torchlib] Support int input for floor_divide #1854

Closed justinchuby closed 3 weeks ago

justinchuby commented 3 weeks ago

Implement floor_divide for int inputs. We implement it only for positive inputs (using integer division) because that is the usual intended case and is the most efficient.

Create op info for the aten op because the original op info does not produce the same input expected nor has the same behavior.

Mark traceable.

Fix https://github.com/pytorch/pytorch/issues/125753

codecov[bot] commented 3 weeks ago

:x: 19 Tests Failed:

Tests completed Failed Passed Skipped
12058 19 12039 2018
View the top 3 failed tests by shortest run time > > ``` > onnxscript.converter_test.TestConverter test_loops_break > ``` > >
Stack Traces | 0s run time > > > > > ``` > > No failure message available > > ``` > >
onnxscript.converter_test.TestConverter test_opset_import
Stack Traces | 0s run time > > ``` > No failure message available > ```
onnxscript.converter_test.TestConverter test_eager_op
Stack Traces | 0.001s run time > > ``` > No failure message available > ```

To view individual test run time comparison to the main branch, go to the Test Analytics Dashboard