mirror of
https://github.com/pytorch/pytorch.git
synced 2025-10-21 13:44:15 +08:00
1646d6f939
This reverts commit 3279f06410032e9798e380cedf552f5b706ac6c1. Reverted https://github.com/pytorch/pytorch/pull/102593 on behalf of https://github.com/malfet due to There is nothing wrong with this PR, but it fails some internal builds that depend on outdated typing_extensions, will reland when update is done ([comment](https://github.com/pytorch/pytorch/pull/102593#issuecomment-1636062515))
46 lines
2.2 KiB
Python
46 lines
2.2 KiB
Python
from typing import List, Dict, Tuple, Optional
|
|
|
|
import torch
|
|
from torch import Tensor
|
|
from torch.autograd.grad_mode import no_grad
|
|
|
|
def _get_foreach_kernels_supported_devices() -> List[str]:
|
|
r"""
|
|
Return the device type list that supports foreach kernels.
|
|
"""
|
|
return ["cuda", torch._C._get_privateuse1_backend_name()]
|
|
|
|
def _get_fused_kernels_supported_devices() -> List[str]:
|
|
r"""
|
|
Return the device type list that supports fused kernels in optimizer.
|
|
"""
|
|
return ["cuda", "xpu", torch._C._get_privateuse1_backend_name()]
|
|
|
|
# This util function splits tensors into groups by device and dtype, which is useful before sending
|
|
# tensors off to a foreach implementation, which requires tensors to be on one device and dtype.
|
|
# If tensorlistlist contains more than one tensorlist, the following assumptions are made BUT NOT verified:
|
|
# - tensorlists CAN be None
|
|
# - all tensors in the first specified list cannot be None
|
|
# - given an index i, all specified tensorlist[i]s match in dtype and device
|
|
# with_indices (bool, optional): whether to track previous indices as the last list per dictionary entry.
|
|
# It comes in handy if there are Nones or literals in the tensorlists that are getting scattered out.
|
|
# Whereas mutating a tensor in the resulting split-up tensorlists WILL propagate changes back to the
|
|
# original input tensorlists, changing up Nones/literals WILL NOT propagate, and manual propagation
|
|
# may be necessary. Check out torch/optim/sgd.py for an example.
|
|
@no_grad()
|
|
def _group_tensors_by_device_and_dtype(
|
|
tensorlistlist: List[List[Optional[Tensor]]],
|
|
with_indices: bool = False,
|
|
) -> Dict[Tuple[torch.device, torch.dtype], Tuple[List[List[Optional[Tensor]]], List[int]]]:
|
|
return {
|
|
(device, getattr(torch, str_dtype)): value
|
|
for (device, str_dtype), value in
|
|
torch._C._group_tensors_by_device_and_dtype(tensorlistlist, with_indices).items()
|
|
}
|
|
|
|
|
|
def _has_foreach_support(tensors: List[Tensor], device: torch.device) -> bool:
|
|
if device.type not in set(_get_foreach_kernels_supported_devices() + ["cpu"]) or torch.jit.is_scripting():
|
|
return False
|
|
return all(t is None or type(t) == torch.Tensor for t in tensors)
|