Summary:
This is to move us along the path to removing Type from the public API.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/12355
Reviewed By: ezyang
Differential Revision: D10212616
Pulled By: gchanan
fbshipit-source-id: c9cd128d1111ab219cb0b2f3bf5b632502ab97c0
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/12288
Current implementation of Tensor takes an intrusive_ptr as an argument for storing data. But instead of requiring users to explicitly pass an intrusive_ptr we want them to pass args for intrusive ptr directly which are forwarded internally through new helper function called make_tensor
Reviewed By: ezyang
Differential Revision: D10152661
fbshipit-source-id: bfa72de161ace3fd1c4573427abcd1bfbd12e29e
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/12328
- Delete reset() from Storage, as it makes it easy to accidentally
create a null storage.
- Immediately reject a storage if it is null when passed in
Reviewed By: dzhulgakov
Differential Revision: D10200448
fbshipit-source-id: 14bfa45f8f59859cc350bd9e20e3ef8692e3991d
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/12103
This defers lookup of defaults to the site where we read
out of TensorOptions. THIS IS A BC-BREAKING BEHAVIOR CHANGE,
but we expect the bulk of uses of OptionsGuard don't allocate TensorOptions
inside the OptionsGuard region, and then use it outside of the region
(the situation where behavior could change.)
I also optimize the size of TensorOptions by rearranging fields, so that we
always fit in two 64-bit words.
Reviewed By: goldsborough
Differential Revision: D10052523
fbshipit-source-id: f454a15b4dbf8cd17bc902ab7d2016f2f689ed13
Summary:
The changes are made to clarify how the parsing between the yaml files and header files of THNN and THCUNN works. As issue #12320 shows it is not that easy to understand the existing code without a hint to the important files.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/12367
Differential Revision: D10217459
Pulled By: ezyang
fbshipit-source-id: 9b3e64dea4f156843814840e736dc3230332060c
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/12340
`long` and `int64_t` are the same type on 64-bit Android.
Reviewed By: Yangqing
Differential Revision: D10204892
fbshipit-source-id: 2d5bf707bf87b99fc597c9292b59f032e9004620
Summary:
* Replaces `prim::PythonOp` with the name of the function being called
* Delays printing values used in `prim::Return` nodes until the return
node itself if that is the only place the value is used to remove some
useless assigns
zdevito apaszke ezyang
Pull Request resolved: https://github.com/pytorch/pytorch/pull/12179
Differential Revision: D10132661
Pulled By: driazati
fbshipit-source-id: cbc4ac34137ed5872049082e25d19eb1ebc71208
Summary:
All usages of the `ndarray` construct have now been guarded with `USE_NUMPY`. This eliminates the requirement of NumPy while building PyTorch from source.
Fixes#11757
Reviewed By: Yangqing
Differential Revision: D10031862
Pulled By: SsnL
fbshipit-source-id: 32d84fd770a7714d544e2ca1895a3d7c75b3d712
Summary:
This is to fix cmake-time compilation error.
When we change script to build Caffe2 with mkldnn, we run into some cmake-time compilation support check (like in libsleef) failed due to incorrect setting of CMAKE_REQUIRED_LIBRARIES. It is a global setting which can interfere camke compilation if it is not clean up properly. FindBLAS.cmake and FindLAPACK.cmake didn't clean this flag, and causes incorrect building of libsleef.so.
yinghai gujinghui
Pull Request resolved: https://github.com/pytorch/pytorch/pull/12195
Differential Revision: D10159314
Pulled By: yinghai
fbshipit-source-id: 04908738f7d005579605b9c2a58d54f035d3baf4
Summary:
- Removed the old nccl file
- Make open-source NCCL a submodule
- CMake to make NCCL itself
NCCL2 now is in the default build.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/12312
Differential Revision: D10190845
Pulled By: teng-li
fbshipit-source-id: 08d42253b774149a66919d194f88b34628c39bae
Summary:
Tensors cannot be created globally because of static initialization order issues. So tensors for the optim_baseline test must be created lazily instead. This is fine because these functions will only be called once (in the respective test).
ezyang
Pull Request resolved: https://github.com/pytorch/pytorch/pull/12301
Differential Revision: D10201008
Pulled By: goldsborough
fbshipit-source-id: 59a041f437354e7c6600e5655b3e2d0647dbde9e
Summary:
CAFFE2_UNIQUE_LONG_TYPEMETA has been a tricky variable defined only from cmake - this is an experiment to remove it and see what exact compilers need that one set.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/12311
Reviewed By: dzhulgakov
Differential Revision: D10187777
Pulled By: Yangqing
fbshipit-source-id: 03e4ede4eafc291e947e0449382bc557cb624b34
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/11465
It happened in one of my testing workflow run that deserialization of dataset_cursor failed. The reason it fails is due to the offset vector is serialized only when it's non-empty, but deserialization always process offset_blob whenever it is called. Though I'm still checking the reason why the offset of dataset_cursor is empty, I think it's good to remove this discrepancy.
Reviewed By: aazzolini, Tianshu-Bao
Differential Revision: D9737636
fbshipit-source-id: bb111933f534b092f29469680ff29e59617655f0
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/12274
We use caffe2::int8::Int8TensorCPU for quantized tensor with uint8_t element type.
Reviewed By: llyfacebook
Differential Revision: D10156452
fbshipit-source-id: 52cf2bedc9dbb433cd5d03f0b76723f7df6a7361
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/12029
In order to remove New() function in StaticContext(to remove StaticContext) and converge to the Allocator design, we'll first change the return type of New to at::DataPtr.
Reviewed By: ezyang
Differential Revision: D9889990
fbshipit-source-id: 3257c763530b987025f428741bdd2e089d11bad4
Summary:
At long last, we will have clang-tidy enabled in CI. For a while I thought I could clean up the project enough to enable clang-tidy with all checks enabled, but I figure it's smarter to set up the minimal checks and at least have those in CI. We can fix more going forward.
ezyang apaszke
Pull Request resolved: https://github.com/pytorch/pytorch/pull/12213
Differential Revision: D10183069
Pulled By: goldsborough
fbshipit-source-id: 7ecd2d368258f46efe23a2449c0a206d10f3a769
Summary:
If a PythonOp throws an error it raises an exception to the interpreter and also releases the GIL which causes [pybind to segfault](https://github.com/potassco/clingo/issues/42)
This fix catches pybind errors while the GIL is still held and throws a `python_error` to re-capture the GIL
Fixes#12118
apaszke
Pull Request resolved: https://github.com/pytorch/pytorch/pull/12243
Differential Revision: D10182787
Pulled By: driazati
fbshipit-source-id: 719d4a7c3294af201e061cf7141bec3ca0fb1f04
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/12279
By some reason if we don't write to the wipe buffer, it doesn't really wipe out everything from caches in x86.
We also need to wipe out cache after initializing input blobs.
Reviewed By: Maratyszcza
Differential Revision: D10161211
fbshipit-source-id: c34414dd8b83947805010d7d57e4134d56de1430
Summary:
Fix things in onnxwhile op to support nested loops, correctly track loop carried deps. Nested loops should be fully supported together with https://github.com/onnx/onnx/pull/1453
Pull Request resolved: https://github.com/pytorch/pytorch/pull/12124
Differential Revision: D10108817
Pulled By: wanchaol
fbshipit-source-id: 51b948024da857c9962833213ee792f47f054e48
Summary:
If block is missing control inputs when do caffe2 net execution, this PR add them back and remove the un-SSA semantics
jamesr66a
Pull Request resolved: https://github.com/pytorch/pytorch/pull/12224
Differential Revision: D10135408
Pulled By: wanchaol
fbshipit-source-id: 746c870bde54ed4ca627167361db1b3f36cd235c
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/12266
- Put all byte-size fields together (booleans and TensorTypeId),
so they can be coalesced into a single word.
- Replace std::vector<int64_t> strides with
std::unique_ptr<int64_t[]>, saving two words.
Reviewed By: dzhulgakov
Differential Revision: D10150834
fbshipit-source-id: f54f38eec34732f3ff7e52e00b1371d7b5b210eb
Summary:
This PR adds a bool type to `IValue` and puts it into place.
* changes conds for `prim::If` and `prim::Loop` to use `bool` type
* changes operators that take `bool`s to match their native ops
* fixes ambiguous `aten` ops `aten::std` and `aten::var`
* fixes tests in `test_jit.py TestJitGenerated`
```
'test_std_dim',
'test_std_dim_1d',
'test_std_dim_1d_neg0',
'test_std_dim_neg0',
'test_var_dim',
'test_var_dim_1d',
'test_var_dim_1d_neg0',
'test_var_dim_neg0'
```
* adds `prim::BoolToTensor` and `prim::TensorToBool`
apaszke zdevito
Pull Request resolved: https://github.com/pytorch/pytorch/pull/11834
Differential Revision: D9928570
Pulled By: driazati
fbshipit-source-id: 373c53df2f1a8ffa9e33d9a517002fbeef25f3eb
Summary: Adding back import{Node,Edge} as move{Node,Edge} and adding a new function moveSubgraph
Reviewed By: duc0, yyetim
Differential Revision: D10128131
fbshipit-source-id: b0e17ec2802cb211b6455578fdb17dab2a7a425b
Summary:
Hi, I think it might be better to use https instead of http in the README.md.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/12258
Differential Revision: D10162279
Pulled By: soumith
fbshipit-source-id: 4658aa75175909b4fea6972b437765d8b49c749f
Summary:
This variable is already being used so this just serves to document that. I think it's an important variable, too, so it should definitely be documented there somewhere.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/12265
Differential Revision: D10162261
Pulled By: soumith
fbshipit-source-id: e0d01e012c2fedea63372de9967a8eaa3745fe94
Summary:
The code that reads a blob from input files are broken. Fixing them. Also, add a binary that converts input files to blobs that can be used by Caffe2 directly.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/12211
Reviewed By: llyfacebook
Differential Revision: D10121845
Pulled By: sf-wind
fbshipit-source-id: 6e48bb594680bcb3186d8d43276b602041c30d3e
