frozenleaves/pytorch
1
0
Fork 0
mirror of https://github.com/pytorch/pytorch.git synced 2025-10-21 05:34:18 +08:00
Code Issues Packages Projects Releases Wiki Activity

free up dispatch key space (in C++) (#69633)

Browse Source 
Summary: Pull Request resolved: https://github.com/pytorch/pytorch/pull/69633

Test Plan: Imported from OSS

Reviewed By: albanD

Differential Revision: D33255193

Pulled By: bdhirsh

fbshipit-source-id: 79773e9c15bf4f2f27675121a49ff5ffd1375238
(cherry picked from commit eac0b1300569e035f3de28a1f0fdce03f60bd270)
This commit is contained in:
Brian Hirsh
2022-02-04 09:53:30 -08:00
committed by PyTorch MergeBot
parent 1cec719448
commit 20b8653dfa
20 changed files with 1747 additions and 515 deletions
Download Patch File Download Diff File Expand all files Collapse all files

6
torch/_python_dispatcher.py
View File

@ -15,9 +15,9 @@ keys for a single example of each use case. These use cases are listed below:
- CPU/AutogradCPU: represents in-tree backends which we usually have dedicated inference &
autograd kernel in pytorch core library.
E.g. CPU, CUDA
- QuantizedCPU/AutogradOther: represents in-tree backends which we usually have backend specific
- FPGA/AutogradOther: represents in-tree backends which we usually have backend specific
inference kernels, but they share the same autograd kernel specified in AutogradOther.
E.g. QuantizedCPU, QuantizedCUDA
E.g. FPGA, SparseCsrCPU
- XLA/AutogradXLA: represents out-of-tree backends which we don't have either inference or autograd
kernel defined in pytorch core library. Backend owner is responsible for registering both
inference & autograd kernels in their extensions(e.g. torch-xla) for the operators they support.
@ -53,7 +53,7 @@ class PythonDispatcher:
name = "foo"
runtime_keys = [
"CPU", "AutogradCPU",
"QuantizedCPU", "AutogradOther",
"FPGA", "AutogradOther",
"XLA", "AutogradXLA",
"Lazy", "AutogradLazy",
]