# Copyright 2026 Tensor Auto Inc. All rights reserved.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
"""Utilities for working with PyTorch FakeTensor.
This module provides a FakeTensorContext class and decorator for running code
with FakeTensor mode enabled, which is useful for shape inference and testing
without actual tensor computations.
"""
import functools
from torch._subclasses import FakeTensorMode
from torch.fx.experimental.symbolic_shapes import ShapeEnv
from opentau.utils.monkey_patch import (
torch_fake_tensor_beta_validate_args_patch,
torch_fake_tensor_is_inf_patch,
torch_fake_tensor_module_to_patch,
torch_fake_tensor_to_numpy_patch,
)
# Share the ShapeEnv instance across all FakeTensorContext instances
# Without this, each FakeTensor.item() call would start numbering from 0, which is wrong.
_shared_shape_env = ShapeEnv()
[docs]
class FakeTensorContext:
"""Context manager for enabling FakeTensor mode with necessary patches.
This context manager applies all necessary monkey patches for FakeTensor
compatibility and manages the FakeTensorMode lifecycle.
Args:
allow_non_fake_inputs: If True, allow non-fake tensors as inputs.
Defaults to True.
"""
[docs]
def __init__(self, allow_non_fake_inputs: bool = True):
self.mode = FakeTensorMode(
shape_env=_shared_shape_env,
allow_non_fake_inputs=allow_non_fake_inputs,
)
torch_fake_tensor_module_to_patch()
torch_fake_tensor_to_numpy_patch()
torch_fake_tensor_beta_validate_args_patch()
torch_fake_tensor_is_inf_patch()
def __enter__(self):
return self.mode.__enter__()
def __exit__(self, exc_type, exc_val, exc_tb):
return self.mode.__exit__(exc_type, exc_val, exc_tb)
[docs]
def run_with_fake_tensor(fn):
"""Decorator to run a function with FakeTensor enabled.
Args:
fn: Function to wrap.
Returns:
Wrapped function that runs with FakeTensorContext enabled.
"""
@functools.wraps(fn)
def wrapper(*args, **kwargs):
with FakeTensorContext():
return fn(*args, **kwargs)
return wrapper