Enable lowering for upsample_bilinear2d with scale factor

[lsy323]

This fixes #2703

Enable lowering for upsample_bilinear2d with scale factor

Added a new unit tests of upsample_bilinear2d with scale factor != 1. The test will fail with the current implementation, since it will fallback to Aten with scale factor != 1. The test passed after the lowering for upsample_bilinear2d with scale factor is supported.

[JackCaoG]

I think we need to make sure only one of (scales_h + scales_w ) and output_size is specified, otherwise I don't know which one we should believe. Ideally upsteram already does that check but we can make it more explict.

[lsy323]

After investigating using GDB, I found the output_size will always be filled by upstream (at: https://github.com/pytorch/pytorch/blob/88366a907549abdd7e2c402a961b60c2be910824/aten/src/ATen/native/UpSampleBilinear2d.cpp#L166)

With the current upstream implementation, we can rely on the output_size inferred from scales_h/w by upstream. However, I think we can keep the scale factor shape inference here to make it future-proof.

In the scale factor shape inference block, I think we can do a shape validation if output_size is not empty, to make sure the output shape and the inferred shape are the same.

[JackCaoG]

For a sanity check, @lsy323 can you check if you use python api upsample_bilinear2d with scale factor, will it fall back to CPU(you can check that by looking at coutner before after the op)?

[lsy323]

For a sanity check, @lsy323 can you check if you use python api upsample_bilinear2d with scale factor, will it fall back to CPU(you can check that by looking at coutner before after the op)?

Will do, thanks!

[lsy323]

For a sanity check, @lsy323 can you check if you use python api upsample_bilinear2d with scale factor, will it fall back to CPU(you can check that by looking at coutner before after the op)?

Tested with the following Python script, upsample_bilinear2d with scale factor will be lowered instead of falling back to CPU

import torch import torch_xla.core.xla_model as xm import torch_xla.debug.metrics as met import unittest class UpsampleBilinear2DTest(unittest.TestCase): def test_upsample_bilinear2d(self): xla_device = xm.xla_device() t1 = torch.rand(2, 2, 5, 5, device=xla_device) upsample_bilinear2d = torch.nn.UpsamplingBilinear2d( scale_factor=(8.0 / 5.0, 8.0 / 5.0)) t2 = upsample_bilinear2d(t1) xm.mark_step() assert (met.counter_value("xla::upsample_bilinear2d") == 1) if __name__ == '__main__': test = unittest.main() sys.exit(0 if test.result.wasSuccessful() else 1)