frozenleaves/pytorch
1
0
Fork 0
mirror of https://github.com/pytorch/pytorch.git synced 2025-10-20 21:14:14 +08:00
Code Issues Packages Projects Releases Wiki Activity
Files
501d0729cb7cfd25aebb2c0e82cef035bf4e7e78
pytorch/torch
History
Kshiteej K 849b08f14b [reland][chalf] where(cpu and cuda), pow(cuda) (#78665) 
Reland: https://github.com/pytorch/pytorch/pull/77640
Ref: #74537
Pull Request resolved: https://github.com/pytorch/pytorch/pull/78665
Approved by: https://github.com/ngimel
2022-06-02 18:04:06 +00:00
..
_C
enable with semantics
2022-06-01 21:14:45 +00:00
_C_flatbuffer
[4/5]Testing jit module in flatbuffer in Python. (#74387)
2022-03-24 23:29:47 +00:00
_decomp
Norm decomposition (#78582)
2022-06-02 00:25:43 +00:00
_lazy
Revert "Revert "[LT] Codegen ReuseNode for supported ops""
2022-05-16 20:14:42 +00:00
_masked
Revert "masked logsumexp/logaddexp"
2022-05-24 16:12:35 +00:00
_prims
Cleanup impl_nvfuser for unary ops (#78670)
2022-06-02 16:17:47 +00:00
_refs
Reference implementations for rot90, roll, atleast_1d,2d,3d (#78080)
2022-06-02 09:05:11 +00:00
_subclasses
Modernize FakeTensorMode, throw on non-fake inputs
2022-06-01 21:43:59 +00:00
amp
Update amp document with CPU Training/Inference Examples (#77244)
2022-05-11 15:42:45 +00:00
ao
Add nonuniform observer class and tests
2022-06-02 16:29:21 +00:00
autograd
[forward ad] forbid non-float non-complex tangent and primal
2022-05-31 20:58:19 +00:00
backends
[coreml] Introducing Quantization (#78108)
2022-06-01 17:10:17 +00:00
contrib
Revert D30279364: [codemod][lint][fbcode/c*] Enable BLACK by default
2021-08-12 11:45:01 -07:00
cpu
add autocast cpu doc
2022-03-22 02:02:43 +00:00
csrc
Reenable assert after test update
2022-06-02 16:40:06 +00:00
cuda
Resolve TODO after Python 2 for custom_fwd (#78592)
2022-06-01 05:17:41 +00:00
distributed
Update distributed/CONTRIBUTING.md to remove ProcessGroupAgent references and add test instructions
2022-06-01 21:31:12 +00:00
distributions
Add mode property to distributions. (#76690)
2022-05-11 18:26:56 +00:00
fft
[complex32] fft support (cuda only) (#74857)
2022-05-12 04:28:55 +00:00
futures
#71946 Remove Python 3.6 references (#72211)
2022-02-08 03:46:20 +00:00
fx
Fake Tensor Part 1
2022-05-31 16:20:35 +00:00
jit
Adding SSA support for convolution_backward
2022-05-20 18:39:47 +00:00
legacy
…
lib
[NOOP][clangformat][codemod] Enable CLANGFORMAT for some folders in caffe2/* (#67746)
2021-11-03 12:23:14 -07:00
linalg
[CUDA][Linalg] Add a driver= kwarg to torch.linalg.svd and svdvals; add cusolver gesvdaStridedBatched driver to svd (#74521)
2022-05-31 16:11:53 +00:00
monitor
torch/monitor: TensorboardEventHandler (#71658)
2022-01-27 08:33:55 +00:00
multiprocessing
Restore old names for private funcs in legacy storages (#77861)
2022-05-20 02:03:34 +00:00
nested
[PyTorch] Delete NestedTensor Python wrapper (#74691)
2022-03-29 19:13:40 +00:00
nn
Revert "Fix embedding jvp support by making embedding_renorm ignore forward mode AD (#78560)"
2022-06-02 17:40:34 +00:00
onnx
[ONNX][WIP] Refactor verification.py
2022-05-31 18:49:39 +00:00
optim
Adding maximize to Adamax (#77409)
2022-05-16 17:34:44 +00:00
package
torch: Fix black linter
2022-05-20 01:14:08 +00:00
profiler
Add __all__ definition in torch.profiler to fix Pylance type check er… (#78553)
2022-06-02 16:48:36 +00:00
quantization
[quant][fx][improvement] Renamed default_affine_fixed_qparams_observer and default_symmetric_fixed_qparams_observer (#76637)
2022-05-04 02:39:20 +00:00
sparse
Compressed sparse layout conversion stubs (#77489)
2022-05-16 18:37:42 +00:00
special
Bessel functions (#78451)
2022-06-02 14:06:20 +00:00
testing
[reland][chalf] where(cpu and cuda), pow(cuda) (#78665)
2022-06-02 18:04:06 +00:00
utils
[DataLoader] DataLoader now automatically apply sharding to DataPipes
2022-06-02 17:40:29 +00:00
__config__.py
Revert D30279364: [codemod][lint][fbcode/c*] Enable BLACK by default
2021-08-12 11:45:01 -07:00
__future__.py
Revert D30279364: [codemod][lint][fbcode/c*] Enable BLACK by default
2021-08-12 11:45:01 -07:00
__init__.py
Register the extension device module as a native module under torch namespace (#78329)
2022-05-31 17:39:18 +00:00
_appdirs.py
Revert D30279364: [codemod][lint][fbcode/c*] Enable BLACK by default
2021-08-12 11:45:01 -07:00
_classes.py
Revert D30279364: [codemod][lint][fbcode/c*] Enable BLACK by default
2021-08-12 11:45:01 -07:00
_deploy.py
[lint] upgrade mypy to latest version
2022-05-03 20:51:34 +00:00
_jit_internal.py
Revert "Reland "[pytorch][PR] Support dataclasses in TorchScript" take 2 (#74353)"
2022-03-31 04:17:33 -07:00
_linalg_utils.py
Remove deprecated torch.solve (#70986)
2022-05-10 13:44:07 +00:00
_lobpcg.py
remove inverse from LOBPCG
2022-04-20 19:03:00 +00:00
_lowrank.py
Prefer mT and mH over transpose(-2, -1) and transpose(-2, -1).conj() (#64181)
2021-10-18 13:02:25 -07:00
_meta_registrations.py
addr ref
2022-05-25 01:40:11 +00:00
_namedtensor_internals.py
Revert D30279364: [codemod][lint][fbcode/c*] Enable BLACK by default
2021-08-12 11:45:01 -07:00
_ops.py
[S271837] alias torch._utils_internal (#78597)
2022-06-01 05:21:52 +00:00
_python_dispatcher.py
Lint fix
2022-05-05 05:52:40 +00:00
_six.py
Update min python version to 3.7 in setup.py and mypy configs (#71494)
2022-01-20 00:03:57 +00:00
_sources.py
Revert "Reland "[pytorch][PR] Support dataclasses in TorchScript" take 2 (#74353)"
2022-03-31 04:17:33 -07:00
_storage_docs.py
Merge torch.cuda._UntypedStorage into torch._UntypedStorage (#75459)
2022-05-19 13:54:39 +00:00
_tensor_docs.py
DOC Corrects default value for storage_offset in as_strided (#78202)
2022-05-31 19:28:36 +00:00
_tensor_str.py
Support str for Sparse Compressed tensors
2022-05-18 12:58:54 +00:00
_tensor.py
Support saving Bfloat16 tensors for XLA/HPU (#77534)
2022-06-01 14:19:09 +00:00
_torch_docs.py
Fix docs for torch.real (#78644)
2022-06-02 04:17:03 +00:00
_utils_internal.py
Revert D30279364: [codemod][lint][fbcode/c*] Enable BLACK by default
2021-08-12 11:45:01 -07:00
_utils.py
[DOCS] Add docstring to _get_async_or_non_blocking in _utils.py (#78036)
2022-06-01 16:19:43 +00:00
_VF.py
Revert D30279364: [codemod][lint][fbcode/c*] Enable BLACK by default
2021-08-12 11:45:01 -07:00
_vmap_internals.py
More aggressively market functorch.vmap when torch.vmap gets called (#67347)
2021-11-12 16:10:16 -08:00
abi-check.cpp
…
CMakeLists.txt
Revert "[cuDNN V8 API] (reopen) Allow the number of kernels profiled under torch.backends.cudnn.benchmark = True to be limitedCudnnv8 benchmark limit (#77002)"
2022-05-24 21:52:35 +00:00
custom_class_detail.h
[NOOP][clangformat][codemod] Enable CLANGFORMAT for some folders in caffe2/* (#67746)
2021-11-03 12:23:14 -07:00
custom_class.h
Fix some typos.
2022-04-11 21:55:59 +00:00
deploy.h
…
extension.h
…
functional.py
Revert "stft: remove non-center overload and python functional wrapper"
2022-05-09 19:59:46 +00:00
hub.py
Minor torchhub docs
2022-05-10 11:01:02 +00:00
library.h
Back out Dispatcher change that makes Messenger Desktop crash on M1 devices (#77414)
2022-05-13 17:33:53 +00:00
library.py
Add a check to ensure input func to Library.impl is callable
2022-06-02 16:55:39 +00:00
overrides.py
Bessel functions (#78451)
2022-06-02 14:06:20 +00:00
py.typed
…
quasirandom.py
Revert D30279364: [codemod][lint][fbcode/c*] Enable BLACK by default
2021-08-12 11:45:01 -07:00
random.py
Adds return type annotation for fork_rng function (#63724)
2021-08-27 09:03:40 -07:00
README.txt
…
return_types.py
Register torch.return_types.* as pytree nodes
2022-04-19 13:46:20 +00:00
script.h
[NOOP][clangformat][codemod] Enable CLANGFORMAT for some folders in caffe2/* (#67746)
2021-11-03 12:23:14 -07:00
serialization.py
Merge torch.cuda._UntypedStorage into torch._UntypedStorage (#75459)
2022-05-19 13:54:39 +00:00
storage.py
Fix _free_weak_ref error (#78575)
2022-06-01 00:07:48 +00:00
torch_version.py
Import packaging.version in torch_version, if available (#71902)
2022-02-22 21:30:14 +00:00
types.py
Introducing SymInt to Pytorch (for tracing size arithmetic) (master rebase) (#74861)
2022-03-31 21:59:59 +00:00

README.txt 

Note [TH abstraction violation]
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

TH/THC provide some hpp headers, which are proper C++ headers rather than
C headers.  These headers serve double duty as *internal implementation
detail* headers, whose contents should largely not be used by external
clients.

Ideally, we would not install these headers at all; instead, you should
use public functions (in headers like `THTensor.h`, NOT `THTensor.hpp`)
to manipulate these structs.  However, there are a few places
in torch/csrc where we violate this abstraction.  They are marked with
a pointer to this note.  Each of those sites will have to be refactored
when we refactor the guts of THTensor and related structures.