[executorch][serialization] Serialize PTD files.

Introduce top-level serialization file that calls:
- serialize_pte_binary for PTE file
- FlatTensor.serialize_tensors for PTD files.

Differential Revision: [D66523267](https://our.internmc.facebook.com/intern/diff/D66523267/)

[ghstack-poisoned]

Author: lucylq

exir/_serialize/TARGETS

@@ -33,6 +33,7 @@ runtime.python_library(
         "_dataclass.py",
         "_flatbuffer.py",
         "_program.py",
+        "_serialize.py",
         "utils.py",
         "data_serializer.py",
     ],

exir/_serialize/_serialize.py

@@ -0,0 +1,77 @@
+# Copyright (c) Meta Platforms, Inc. and affiliates.
+# All rights reserved.
+#
+# This source code is licensed under the BSD-style license found in the
+# LICENSE file in the root directory of this source tree.
+
+# pyre-strict
+
+
+from typing import Dict, Tuple
+
+from executorch.exir._serialize import _serialize_pte_binary
+
+from executorch.exir._serialize._cord import Cord
+from executorch.exir._serialize.data_serializer import (
+    DataSerializer,
+    SerializationInfo,
+    TensorLayout,
+)
+
+from executorch.exir.capture._config import ExecutorchBackendConfig
+from executorch.exir.emit import EmitterOutput
+from executorch.exir.schema import Tensor, TensorDataLocation
+
+
+def serialize(
+    emitter_output: EmitterOutput,
+    config: ExecutorchBackendConfig,
+    data_serializer: DataSerializer,
+) -> Tuple[Cord, Dict[str, Cord]]:
+    """Serialize the output from Emitter into ExecuTorch artifacts; PTE and PTD files."""
+    # Serialize PTE file.
+    pte: Cord = _serialize_pte_binary(
+        program=emitter_output.program,
+        mutable_data=emitter_output.mutable_data,
+        extract_delegate_segments=config.extract_delegate_segments,
+        segment_alignment=config.segment_alignment,
+        constant_tensor_alignment=config.constant_tensor_alignment,
+        delegate_alignment=config.delegate_alignment,
+    )
+
+    # Serialize PTD files.
+    ptd_files: Dict[str, Cord] = {}
+
+    # Find all external tensors and organize into {fqn: Tensor}.
+    fqn_to_tensor_layout: Dict[str, TensorLayout] = {}
+    for plan in emitter_output.program.execution_plan:
+        for evalue in plan.values:
+            if isinstance(evalue.val, Tensor):
+                tensor = evalue.val
+                if (
+                    tensor.extra_tensor_info is not None
+                    and tensor.extra_tensor_info.fully_qualified_name is not None
+                    and tensor.extra_tensor_info.location is TensorDataLocation.EXTERNAL
+                ):
+                    # pyre-ignore Undefined attribute [16]: Optional type has no attribute `fully_qualified_name`.
+                    fqn_to_tensor_layout[
+                        tensor.extra_tensor_info.fully_qualified_name
+                    ] = TensorLayout(tensor.scalar_type, tensor.sizes, tensor.dim_order)
+    if len(fqn_to_tensor_layout) > 0:
+        assert emitter_output.external_constant_map is not None
+        for (
+            file,
+            fqn_map,
+        ) in (
+            # pyre-ignore Undefined attribute [16]: Optional type has no attribute `items`.
+            emitter_output.external_constant_map.items()
+        ):
+            ptd_files[file] = data_serializer.serialize_tensors(
+                SerializationInfo(
+                    emitter_output.external_constant_buffer,
+                    fqn_map,
+                    fqn_to_tensor_layout,
+                )
+            )
+
+    return pte, ptd_files

exir/program/TARGETS

@@ -44,6 +44,7 @@ python_library(
         "//executorch/exir/passes:spec_prop_pass",
         "//executorch/exir/passes:weights_to_outputs_pass",
         "//executorch/exir/verification:verifier",
+        "//executorch/extension/flat_tensor/serialize:serialize",
     ] +  (["//executorch/exir/program/fb:logger"] if not runtime.is_oss else [])
 )
 

exir/program/_program.py

@@ -13,8 +13,9 @@
 
 import torch
 import torch._export
-from executorch.exir._serialize import _serialize_pte_binary
 from executorch.exir._serialize._cord import Cord
+from executorch.exir._serialize._serialize import serialize
+from executorch.exir._serialize.data_serializer import DataSerializer
 from executorch.exir._warnings import experimental
 from executorch.exir.backend.backend_api import to_backend
 from executorch.exir.backend.partitioner import Partitioner
@@ -56,6 +57,7 @@
     EXIREdgeDialectVerifier,
     get_aten_verifier,
 )
+from executorch.extension.flat_tensor.serialize.serialize import FlatTensorSerializer
 from torch._export.passes import ReplaceViewOpsWithViewCopyOpsPass
 from torch.export import ExportedProgram
 from torch.export._remove_auto_functionalized_pass import (
@@ -494,23 +496,23 @@ def __init__(
             )
         self.exported_program = exir_exported_program.exported_program
         self._pte_data: Optional[Cord] = None
+        self._data_files: Optional[Dict[str, Cord]] = None
         self._buffer: Optional[bytes] = None
         self._emitter_output: Optional[EmitterOutput] = None
         self._emit_stacktrace: bool = emit_stacktrace
         self._extract_delegate_segments: bool = extract_delegate_segments
         self._segment_alignment: int = segment_alignment
         self._constant_tensor_alignment: Optional[int] = constant_tensor_alignment
         self._delegate_alignment: Optional[int] = delegate_alignment
+        self._data_serializer: DataSerializer = FlatTensorSerializer()
 
     def _get_pte_data(self) -> Cord:
         if self._pte_data is None:
-            self._pte_data = _serialize_pte_binary(
-                program=self.program,
-                extract_delegate_segments=self._extract_delegate_segments,
-                segment_alignment=self._segment_alignment,
-                constant_tensor_alignment=self._constant_tensor_alignment,
-                delegate_alignment=self._delegate_alignment,
+            assert self._emitter_output is not None
+            self._pte_data, self._data_files = serialize(
+                self._emitter_output, ExecutorchBackendConfig(), self._data_serializer
             )
+        assert self._pte_data is not None
         return self._pte_data
 
     @property
@@ -1443,14 +1445,11 @@ def __init__(
             self._config_methods,
         )
 
+        self._data_serializer = FlatTensorSerializer()
+
         # Serialize emitter output, ready to be written to a file.
-        self._pte_data: Cord = _serialize_pte_binary(
-            program=self._emitter_output.program,
-            mutable_data=self._emitter_output.mutable_data,
-            extract_delegate_segments=backend_config.extract_delegate_segments,
-            segment_alignment=backend_config.segment_alignment,
-            constant_tensor_alignment=backend_config.constant_tensor_alignment,
-            delegate_alignment=backend_config.delegate_alignment,
+        self._pte_data, self._data_files = serialize(
+            self._emitter_output, ExecutorchBackendConfig(), self._data_serializer
         )
         self._buffer: Optional[bytes] = None
 
@@ -1532,3 +1531,8 @@ def write_to_file(self, open_file: io.BufferedIOBase) -> None:
         reducing the peak memory usage.
         """
         self._pte_data.write_to_file(open_file)
+
+        for filename, cord in self._data_files.items():
+            filename = filename + ".ptd"
+            with open(filename, "wb") as file:
+                cord.write_to_file(file)

extension/flat_tensor/test/test_serialize.py

@@ -25,7 +25,7 @@
 )
 
 # Test artifacts
-TEST_TENSOR_BUFFER = [b"\x11"*4, b"\x22"*32]
+TEST_TENSOR_BUFFER = [b"\x11" * 4, b"\x22" * 32]
 TEST_TENSOR_MAP = {
     "fqn1": 0,
     "fqn2": 0,