Add a developer guide for exporting to executorch

[jerryzh168]

Summary:
att, the requirement for exporting a quantized model to executorch is mainly that we want to preserve soem high level ops so they can be lowered to executorch ops, examples of ops that are already preserved are quantize_affine/dequantize_affine/choose_qparams_affine which can be matched in executorch for pattern matching, this PR adds an example for how to define and preserve a quantized embedding_byte op, the main util function we use is torchao.utils._register_custom_op

The expected output is that after export we see some high level ops (such as quantized embedding op) being preserved:

GraphModule(
  (0): Module(
    (parametrizations): Module(
      (weight): Module()
    )
  )
)



def forward(self, input):
    input, = fx_pytree.tree_flatten_spec(([input], {}), self._in_spec)
    _0_parametrizations_weight_original0 = getattr(self, "0").parametrizations.weight.original0
    _0_parametrizations_weight_original1 = getattr(self, "0").parametrizations.weight.original1
    input_1 = input
    embedding_byte = torch.ops.quant.embedding_byte.default(_0_parametrizations_weight_original0, [1, 128], _0_parametrizations_weight_original1, input_1);  _0_parametrizations_weight_original0 = _0_parametrizations_weight_original1 = input_1 = None
    return pytree.tree_unflatten((embedding_byte,), self._out_spec)

Test Plan:
python tutorials/developer_api_guide/export_to_executorch.py

Reviewers:

Subscribers:

Tasks:

Tags:

[pytorch-bot]

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/ao/1219

• 📄 Preview Python docs built from this PR

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

❌ 1 New Failure

As of commit 34a4ad3 with merge base 3475aed ():

NEW FAILURE - The following job has failed:

• Run Regression Tests / test (CPU Nightly, linux.4xlarge, --pre torch --index-url https://download.pytorch.org/whl/nightl... / linux-job (gh)
  test/prototype/test_parametrization.py::TestFakeSparsity::test_jit_trace

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

[andrewor14]

Looks great! Can you also share the expected output of the tutorial in the PR description?