Add XNN U8 op support via quantization

[GregoryComer]

Summary

Support U8 ops in the XNNPACK delegate by treating the input and output tensors as u8 asymmetric-quantized tensors with scale=1 and zero_point=0. This PR adds U8 support for upsample_bilinear2d, cat, slice, and _to_copy (when used to convert u8 to f32). More ops are possible with this method.

Conversion from u8 to f32 is done via transformation into a dequantize op. This is implemented in a new pass - ReplaceU8ConvertWithDqPass. The general U8 to quantized U8 transformation is done in define_tensor in node_visitor.py. U8 inputs are created as quantized tensors with the appropriate qparams (scale=1, zp=0).

Test plan

I've added op-level u8 tests to each of the new ops, as well as tests for the ReplaceU8ConvertWithDqPass. I've also added an end-to-end test for MobileNetV3 with a wrapper to take U8 inputs, resize and crop, and then convert to f32 and run the model.

[pytorch-bot]

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/executorch/8330

• 📄 Preview Python docs built from this PR

Note: Links to docs will display an error until the docs builds have been completed.

✅ No Failures

As of commit 1fbaf0e with merge base d99970b ():
💚 Looks good so far! There are no failures yet. 💚

This comment was automatically generated by Dr. CI and updates every 15 minutes.

[facebook-github-bot]

@GregoryComer has imported this pull request. If you are a Meta employee, you can view this diff on Phabricator.

[mcr229]

This generally looks right to me.

[digantdesai]

Ah reviewed internally first, please look at the comments there, next time I will remember.

[github-actions]

Looks like this PR hasn't been updated in a while so we're going to go ahead and mark this as Stale.
Feel free to remove the Stale label if you feel this was a mistake.
If you are unable to remove the Stale label please contact a maintainer in order to do so.
If you want the bot to never mark this PR stale again, add the no-stale label.
Stale pull requests will automatically be closed after 30 days of inactivity.

[github-actions]

Looks like this PR hasn't been updated in a while so we're going to go ahead and mark this as Stale.
Feel free to remove the Stale label if you feel this was a mistake.
If you are unable to remove the Stale label please contact a maintainer in order to do so.
If you want the bot to never mark this PR stale again, add the no-stale label.
Stale pull requests will automatically be closed after 30 days of inactivity.