TerryTsao [Tue, 11 Dec 2018 18:41:37 +0000 (10:41 -0800)]
Fix CMakeLists.txt for Int8 python bindings (#15047)
Summary:
Currently in caffe2, one cannot properly fetch the content of Int8 blobs.
Upon digging the source code, it turns out that the relevant source code is not being compiled. Adding the source to CMakeLists.txt fixes this issue.
First time ever doing a pull request. Please let me know if there's any rule I should follow. Thanks.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/15047
Differential Revision:
D13417583
Pulled By: bddppq
fbshipit-source-id:
dd39575971a3012635edbf97a045d80e4b62a8eb
Orion Reblitz-Richardson [Tue, 11 Dec 2018 17:59:28 +0000 (09:59 -0800)]
Install cpp tests when built (#15000)
Summary:
This is broken out of https://github.com/pytorch/pytorch/pull/13733/
We want to install cpp tests so they can ultimately be runnable from that location for Caffe2 tests run from PyTorch builds.
cc pjh5 yf225 anderspapitto
Pull Request resolved: https://github.com/pytorch/pytorch/pull/15000
Reviewed By: pjh5
Differential Revision:
D13416253
Pulled By: orionr
fbshipit-source-id:
51280be0a22557a742f90c9f303c58c35cbd4a38
Michael Carilli [Tue, 11 Dec 2018 17:46:25 +0000 (09:46 -0800)]
Stashing checkpointing RNG states based on devices of arg tensors (#14518)
Summary:
This PR intends to address apaszke's concerns in https://github.com/pytorch/pytorch/pull/14253#issuecomment-
441740016. Preserving the rng state is now controlled by a kwarg rather than a global state, hopefully in a python 2.7-compatible way.
Additionally, the checkpointing function stashes and restores the RNG states of
1. devices associated with all input tensor args to run_fn as well as
2. the current device.
I could easily change this to only save and restore the RNG states associated 1. alone. This would simplify the logic to create a [deduplicated, ordered](https://github.com/pytorch/pytorch/compare/master...mcarilli:checkpointing_rng_touchup?expand=1#diff-58da227fc9b1d56752b7dfad90428fe0R37) list of devices considered active.
I'm wondering if the [get_device_states](https://github.com/pytorch/pytorch/compare/master...mcarilli:checkpointing_rng_touchup?expand=1#diff-58da227fc9b1d56752b7dfad90428fe0R32) and [set_device_states](https://github.com/pytorch/pytorch/compare/master...mcarilli:checkpointing_rng_touchup?expand=1#diff-58da227fc9b1d56752b7dfad90428fe0R47) functions are general enough to reside elsewhere (presumably torch/random.py). I'm also wondering if the check on [torch.cuda._initialized](https://github.com/pytorch/pytorch/compare/master...mcarilli:checkpointing_rng_touchup?expand=1#diff-58da227fc9b1d56752b7dfad90428fe0R47) would be better placed within `get_device_states`.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/14518
Differential Revision:
D13356210
Pulled By: ezyang
fbshipit-source-id:
afa4cc21ce7862142d5cb1dec3750018df222039
svcscm [Tue, 11 Dec 2018 15:38:23 +0000 (07:38 -0800)]
Updating submodules
Reviewed By: cdelahousse
fbshipit-source-id:
d39b31f12ab2ab570548f3e8a65949332a64a0ff
Marat Dukhan [Tue, 11 Dec 2018 08:46:55 +0000 (00:46 -0800)]
Switch Int8Softmax, Int8Relu, and Int8LeakyRelu to QNNPACK (#14933)
Summary:
Int8Softmax: 4x-5x speedup compared to previous implementation
Pull Request resolved: https://github.com/pytorch/pytorch/pull/14933
Differential Revision:
D13406820
Pulled By: Maratyszcza
fbshipit-source-id:
ea8cbe1b861ddb7ff1b851d06d52c6fd6d04ed01
Lingyi Liu [Tue, 11 Dec 2018 06:49:47 +0000 (22:49 -0800)]
Adjust the API call to deserilize the tensorproto (#14132)
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/14132
as title
Reviewed By: jerryzh168
Differential Revision:
D13110697
fbshipit-source-id:
822c9079de11951f90aec3d26f0e4108847e7dac
Natalia Gimelshein [Tue, 11 Dec 2018 06:48:16 +0000 (22:48 -0800)]
use datatype dependent tolerance in data parallel tests
Summary: Pull Request resolved: https://github.com/pytorch/pytorch/pull/14856
Differential Revision:
D13413560
Pulled By: soumith
fbshipit-source-id:
b3a0cfe93477ed332e6eaa2e39ef5f4cc8b36481
paland3 [Tue, 11 Dec 2018 06:34:17 +0000 (22:34 -0800)]
Update pooling.py (#14998)
Summary:
Strange line in the documentation.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/14998
Differential Revision:
D13413235
Pulled By: soumith
fbshipit-source-id:
80d05ec1185719b785f0aac914bc2369c1174f2f
Zachary DeVito [Tue, 11 Dec 2018 06:10:11 +0000 (22:10 -0800)]
Clean up casting ops (#14947)
Summary:
This removes FloatToInt style names replacing it with just the destination
name (e.g. FloatToInt -> Float). This makes it more consistent with the
syntax and makes it easier to add type conversions (just add a new
prim::Int op, for instance).
None of these ops get serialized so this should not effect loading of
old models.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/14947
Differential Revision:
D13408409
Pulled By: zdevito
fbshipit-source-id:
d773fe863f14d9de893f686832769f8cc8903a8e
Jongsoo Park [Tue, 11 Dec 2018 06:08:04 +0000 (22:08 -0800)]
share code between adagrad and rowwise adagrad tests (#14692)
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/14692
Remove some code duplication
Reviewed By: chocjy
Differential Revision:
D13296731
fbshipit-source-id:
5924e037ca64fc4b89234be922bc5ca47fb8bd32
Ilia Cherniavskii [Tue, 11 Dec 2018 05:30:53 +0000 (21:30 -0800)]
TBB task graph (#15041)
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/15041
Adding an alternative implementation of a task graph based on TBB
Reviewed By: dmudiger
Differential Revision:
D13412517
fbshipit-source-id:
f5efedd680bbe0072bf38d504e5682ab51dd630f
bddppq [Tue, 11 Dec 2018 05:25:45 +0000 (21:25 -0800)]
Enable more caffe2 fp16 rocm tests (#15040)
Summary:
cc rohithkrn petrex
Pull Request resolved: https://github.com/pytorch/pytorch/pull/15040
Reviewed By: houseroad
Differential Revision:
D13413068
Pulled By: bddppq
fbshipit-source-id:
b2967f16f8da0b9e80083138fb8632c14e9e9b63
Lu Fang [Tue, 11 Dec 2018 05:22:44 +0000 (21:22 -0800)]
Enable the build of tests in ATen/core (#15032)
Summary:
Otherwise they won't build
Pull Request resolved: https://github.com/pytorch/pytorch/pull/15032
Reviewed By: yinghai
Differential Revision:
D13409801
Pulled By: houseroad
fbshipit-source-id:
95464aa8f3604835997ba1bb7f3c3e51485d1686
Gregory Chanan [Tue, 11 Dec 2018 03:51:27 +0000 (19:51 -0800)]
More scaffolding for LegacyTHDispatch. (#14852)
Summary:
1) at::functions are now also exposed in the at::legacy::th namespace and we move relevant calls over to use them (to avoid merge conflicts)
2) LegacyTHDispatch now handles device-type initialization
3) We generate derived LegacyTHDispatchers, e.g. THLegacyCPULongDispatcher, although they are currently empty.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/14852
Reviewed By: ezyang
Differential Revision:
D13360852
Pulled By: gchanan
fbshipit-source-id:
af6705aeba3593ea5dba9bfc62890e5257bc81f8
Ilia Cherniavskii [Tue, 11 Dec 2018 03:18:06 +0000 (19:18 -0800)]
Back out "Revert
D13043261: [caffe2] Task graph and task future abstractions in executor"
Summary: Pull Request resolved: https://github.com/pytorch/pytorch/pull/15030
Reviewed By: bddppq
Differential Revision:
D13408998
fbshipit-source-id:
9eb675e09fbc4829eab34df7aa660a0590816feb
Jerry Zhang [Tue, 11 Dec 2018 03:16:18 +0000 (19:16 -0800)]
Tensor construction codemod - 2/3 (#14836)
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/14836
Codemod generated with clangr shard mode, 25 files per diff,
motivation: https://github.com/pytorch/pytorch/pull/12407
Reviewed By: bddppq
Differential Revision:
D13335176
fbshipit-source-id:
8d89510670e2cf70559d2f75e68f7181feb0b6d9
Jesse Hellemn [Tue, 11 Dec 2018 02:16:42 +0000 (18:16 -0800)]
Fixing reading of FBGEMM from env variables
Summary: Pull Request resolved: https://github.com/pytorch/pytorch/pull/15023
Reviewed By: orionr
Differential Revision:
D13406778
Pulled By: pjh5
fbshipit-source-id:
2265f01170fb7969cbdf4e44ca6ef183f5d8017d
Syed Tousif Ahmed [Tue, 11 Dec 2018 01:53:33 +0000 (17:53 -0800)]
Alignas Array struct (#14920)
Summary:
This PR aligns the Array struct such that cuda vector performance improvements can be utilized.
I tested this by using it on our Philox header. Note how the vector store instruction gets used for cuda vector types and when using alignas on Array, vs when not using alignas on Array.
With cuda vector type (uint4, uint2, float4): https://godbolt.org/z/UaWOmR
With alignas: https://godbolt.org/z/Eeh0t5
Without alignas: https://godbolt.org/z/QT63gq
Pull Request resolved: https://github.com/pytorch/pytorch/pull/14920
Differential Revision:
D13406751
Pulled By: soumith
fbshipit-source-id:
685b1010ef1f576dde30c278b1e9b642f87c843d
rohithkrn [Tue, 11 Dec 2018 01:25:46 +0000 (17:25 -0800)]
Integrate rocBLAS fp16 api into Caffe2 (#14882)
Summary:
This PR integrates rocBLAS half and mixed precision APIs in to Caffe2.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/14882
Differential Revision:
D13407840
Pulled By: bddppq
fbshipit-source-id:
75cb0d74da066776fa66575f1d255e879d36121e
Junjie Bai [Tue, 11 Dec 2018 01:19:15 +0000 (17:19 -0800)]
Fix old tensor CopyFrom usage in boolean mask operator
Summary: Pull Request resolved: https://github.com/pytorch/pytorch/pull/15025
Differential Revision:
D13407323
Pulled By: bddppq
fbshipit-source-id:
1bc1d28ad0c6c71d25d788549be18917e393ee50
Jongsoo Park [Tue, 11 Dec 2018 01:16:32 +0000 (17:16 -0800)]
unit test with multiple omp threads (#14958)
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/14958
Test with multiple threads
Reviewed By: jianyuh
Differential Revision:
D13394791
fbshipit-source-id:
931a6c3bda15ebc816807e537dd0841c383e7a6f
Jerry Zhang [Tue, 11 Dec 2018 01:13:51 +0000 (17:13 -0800)]
Remove partially initialized Tensor in Deserialization (#14197)
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/14197
Pull Request resolved: https://github.com/pytorch/pytorch/pull/13642
Previously we pass in a patially initialized Tensor to Deserialize and it will fill
it with the result of deserialization of a tensor proto. Now we want it to return
a Tensor directly since it's just a shared pointer to TensorImpl.
Reviewed By: dzhulgakov
Differential Revision:
D12874357
fbshipit-source-id:
12b80a763375da23cfa64a74d6bc186d8d03b94f
Junjie Bai [Mon, 10 Dec 2018 23:56:37 +0000 (15:56 -0800)]
Revert
D13043261: [caffe2] Task graph and task future abstractions in executor
Differential Revision:
D13043261
Original commit changeset:
d89424354aea
fbshipit-source-id:
b307e3281c4d83b60ba2bfadcbcf69afb7a41412
James Reed [Mon, 10 Dec 2018 23:35:11 +0000 (15:35 -0800)]
apply() for ScriptModules (#14655)
Summary:
This can be use to initialize state that is not necessarily eligible for serialization/is implementation-specific. Concretely, I'm going to use this to pack the weight matrices for quantized Linear modules according to the FBGEMM APIs
Pull Request resolved: https://github.com/pytorch/pytorch/pull/14655
Differential Revision:
D13404438
Pulled By: jamesr66a
fbshipit-source-id:
2d327cef5520fdd716b5b1b29effd60a049e8a4a
Edward Yang [Mon, 10 Dec 2018 23:16:27 +0000 (15:16 -0800)]
Simplify THPPointer implementation for Storage. (#14897)
Summary:
We've virtualized the destructor for storage, so we
no longer have to forward to a particular backend.
Signed-off-by: Edward Z. Yang <ezyang@fb.com>
Pull Request resolved: https://github.com/pytorch/pytorch/pull/14897
Differential Revision:
D13399216
Pulled By: ezyang
fbshipit-source-id:
531d29c3f278477cfa8759f30ab4f304d695b659
Edward Yang [Mon, 10 Dec 2018 23:10:23 +0000 (15:10 -0800)]
Disable getNumGPUs rewrite (#14993)
Summary:
cc iotamudelta
Signed-off-by: Edward Z. Yang <ezyang@fb.com>
Pull Request resolved: https://github.com/pytorch/pytorch/pull/14993
Differential Revision:
D13405804
Pulled By: ezyang
fbshipit-source-id:
c4aa9ed29ee2a4f3abf76c1e0fa8babfd738db35
Sebastian Messmer [Mon, 10 Dec 2018 23:06:30 +0000 (15:06 -0800)]
Fix include path for WrapDimMinimal.h
Summary: Pull Request resolved: https://github.com/pytorch/pytorch/pull/14794
Reviewed By: dzhulgakov
Differential Revision:
D13336842
fbshipit-source-id:
ca49a9fd1d409d8a75e43eeb9b9b02c305ebb79a
Sebastian Messmer [Mon, 10 Dec 2018 23:06:30 +0000 (15:06 -0800)]
Move WrapDimMinimal to c10
Summary: Pull Request resolved: https://github.com/pytorch/pytorch/pull/14793
Reviewed By: ezyang
Differential Revision:
D13336841
fbshipit-source-id:
4365a799e1856cc68dd94a273e97663fee5f51db
Edward Yang [Mon, 10 Dec 2018 22:52:16 +0000 (14:52 -0800)]
Stop disabling maybeOverlappingIndices (#14999)
Summary:
Signed-off-by: Edward Z. Yang <ezyang@fb.com>
cc iotamudelta
Pull Request resolved: https://github.com/pytorch/pytorch/pull/14999
Differential Revision:
D13405754
Pulled By: ezyang
fbshipit-source-id:
98459496494390ad1115b4f1f6738d53c14f0745
Jane Wang [Mon, 10 Dec 2018 22:29:41 +0000 (14:29 -0800)]
add gloo allgather support on GPU (#14576)
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/14576
as titled
Reviewed By: pietern
Differential Revision:
D13266063
fbshipit-source-id:
e262f77d63724a7504a7112907bbfba49612fe75
Ilia Cherniavskii [Mon, 10 Dec 2018 22:21:24 +0000 (14:21 -0800)]
Task graph and task future abstractions in executor
Summary: Pull Request resolved: https://github.com/pytorch/pytorch/pull/14116
Reviewed By: dmudiger
Differential Revision:
D13043261
fbshipit-source-id:
d89424354aea14d1d14eb8320fb3aa34908a4e81
Jerry Zhang [Mon, 10 Dec 2018 22:17:43 +0000 (14:17 -0800)]
caffe2/caffe2/contrib/script (#15007)
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/15007
Pull Request resolved: https://github.com/pytorch/pytorch/pull/14979
att
Reviewed By: dzhulgakov
Differential Revision:
D13286191
fbshipit-source-id:
b8a6bc7aea44487aea4dcf7f44c858fd30c6293c
Michael Suo [Mon, 10 Dec 2018 21:43:11 +0000 (13:43 -0800)]
s/Torch Script/TorchScript/g (#15011)
Summary:
pls
Pull Request resolved: https://github.com/pytorch/pytorch/pull/15011
Differential Revision:
D13404158
Pulled By: suo
fbshipit-source-id:
e906281463d65c86e4e9073eb0c0a26f4f29e307
Yuxin Wu [Mon, 10 Dec 2018 20:48:11 +0000 (12:48 -0800)]
Improve the docs of interpolate(align_corners=) (#14806)
Summary:
ailzhang
Pull Request resolved: https://github.com/pytorch/pytorch/pull/14806
Reviewed By: ailzhang
Differential Revision:
D13366332
Pulled By: ppwwyyxx
fbshipit-source-id:
08fcea95d5c86b11cdfe464fdd9daa50050871f1
Giuseppe Ottaviano [Mon, 10 Dec 2018 19:54:45 +0000 (11:54 -0800)]
Improve build time of register_symbols.cpp without compiler hacks (#14911)
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/14911
In optimized modes the compiler tries to inline all the
`unordered_map::operator[]` calls, creating a massive amount of code
which takes several minutes to optimize. Instead, create a table of
PODs and populate the maps using a simple loop.
Reviewed By: soumith, luciang
Differential Revision:
D13382948
fbshipit-source-id:
b6752921e0f7213595d26b39e4397f6a3897960b
Edward Yang [Mon, 10 Dec 2018 18:44:13 +0000 (10:44 -0800)]
Delete defunct THP_API.h header. (#14899)
Summary:
Signed-off-by: Edward Z. Yang <ezyang@fb.com>
Pull Request resolved: https://github.com/pytorch/pytorch/pull/14899
Differential Revision:
D13383687
Pulled By: ezyang
fbshipit-source-id:
f2a08a769cc3775ba55f9c58d622a83df622d816
Edward Yang [Mon, 10 Dec 2018 18:40:25 +0000 (10:40 -0800)]
Disable test_leaf_variable_sharing on ASAN runs
Summary: Pull Request resolved: https://github.com/pytorch/pytorch/pull/15001
Reviewed By: orionr
Differential Revision:
D13399119
fbshipit-source-id:
6b1d098e55a67b1f5bc6d08a8ee3c1be8234a654
Edward Yang [Mon, 10 Dec 2018 18:29:43 +0000 (10:29 -0800)]
Revert
D13306052: [pytorch][PR] Allow converting CharTensor to np arrays
Differential Revision:
D13306052
Original commit changeset:
202d038f139c
fbshipit-source-id:
11f6bdd687f8ea5ce2e5f28f48d19449a5c403eb
Edward Yang [Mon, 10 Dec 2018 17:32:55 +0000 (09:32 -0800)]
Non-INTERFACE AT_LINK_STYLE is dead code (#14822)
Summary:
Signed-off-by: Edward Z. Yang <ezyang@fb.com>
Pull Request resolved: https://github.com/pytorch/pytorch/pull/14822
Differential Revision:
D13355574
Pulled By: ezyang
fbshipit-source-id:
a7173084f8735424619b2e393df2715a05918b44
SsnL [Mon, 10 Dec 2018 16:05:06 +0000 (08:05 -0800)]
Support torch.load with encoding (#14743)
Summary:
Addresses a common compatibility issue when loading Py2 checkpoints in Py3 regarding to bytes.
E.g.,
[1] https://github.com/pytorch/pytorch/issues/5994,
[2] https://github.com/CSAILVision/places365/issues/25,
[3] https://discuss.pytorch.org/t/how-to-load-a-saved-model-trained-on-pytorch-0-3-1-python-2-7-on-pyorch-1-0-python-3-7/31212
Pull Request resolved: https://github.com/pytorch/pytorch/pull/14743
Reviewed By: weiyangfb
Differential Revision:
D13350888
Pulled By: soumith
fbshipit-source-id:
2df4e828a8b70509118a355307ca3ebe51e108f6
SsnL [Mon, 10 Dec 2018 15:36:06 +0000 (07:36 -0800)]
Convert int8 numpy array to CharTensor (#14700)
Summary:
When rewriting `default_collate`, I noticed that `from_numpy` and `as_tensor` and `tensor` all do not work on `np.int8` arrays.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/14700
Reviewed By: weiyangfb
Differential Revision:
D13305297
Pulled By: soumith
fbshipit-source-id:
2937110f65ed714ee830d50098db292238e9b2a9
SsnL [Mon, 10 Dec 2018 15:33:26 +0000 (07:33 -0800)]
Allow converting CharTensor to np arrays (#14710)
Summary:
The other direction of #14700
cc soumith
Pull Request resolved: https://github.com/pytorch/pytorch/pull/14710
Reviewed By: weiyangfb
Differential Revision:
D13306052
Pulled By: soumith
fbshipit-source-id:
202d038f139cf05e01069ff8d05268c66354c983
Jongsoo Park [Mon, 10 Dec 2018 09:06:17 +0000 (01:06 -0800)]
pre-pack operation of dnnlowp conv with 16-bit accumulation (#14881)
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/14881
This diff allows us to pre-quantize and pre-pack weight matrix used in DNNLOWP_ACC16 .
The intended use pattern is run Int8ConvPackWeight in init_net that generates a packed weight and Int8Conv with DNNLOWP_ACC16 engine uses the the packed weight.
Reviewed By: csummersea
Differential Revision:
D13374662
fbshipit-source-id:
dd02b9a4eb7af1fe208aa857fcd0b445e6e395af
Zachary DeVito [Mon, 10 Dec 2018 06:45:18 +0000 (22:45 -0800)]
Respect -q of setup.py (#14972)
Summary:
1. Changes the prints along the 'rebuild' pathway to respect the '-q' flag of setup.py
A clean rebuild now only prints:
[zdevito@devgpu172.prn2 /data/users/zdevito/pytorch] python setup.py -q rebuild develop
[0/1] Install the project...
-- Install configuration: "RelWithDebInfo"
ninja: no work to do.
ninja: no work to do.
ninja: no work to do.
ninja: no work to do.
ninja: no work to do.
ninja: no work to do.
2. Deletes apparently dead calls to `generate_code`. Now that CMake builds these files,
it appears that it is getting called twice and the second version is never used.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/14972
Reviewed By: soumith
Differential Revision:
D13396330
Pulled By: zdevito
fbshipit-source-id:
83c45143bbc6a6d2c1cfee929291ec059f2b5dc3
SsnL [Mon, 10 Dec 2018 05:10:39 +0000 (21:10 -0800)]
_get_device_index supports parsing device strings
Summary: Pull Request resolved: https://github.com/pytorch/pytorch/pull/14929
Reviewed By: weiyangfb
Differential Revision:
D13394498
Pulled By: soumith
fbshipit-source-id:
948c6118abdf6c1e1a8a17709333954cafb2345e
Soumith Chintala [Mon, 10 Dec 2018 04:41:44 +0000 (20:41 -0800)]
remove mingfeima mkldnn reference from README, as no longer necessary (#14975)
Summary: we now get mkldnn automatically from third_party/ideep
Differential Revision:
D13396480
Pulled By: soumith
fbshipit-source-id:
20f819ba4b78cbe9c7d0baeab1c575669cbf6c20
Zachary DeVito [Mon, 10 Dec 2018 00:29:38 +0000 (16:29 -0800)]
fixing some rebuild issues (#14969)
Summary:
This fixes rebuild issues with the ninja part of the build. With this patch all ninja files will now report `nothing to do` if nothing has changed assuming `BUILD_CAFFE2_OPS=0`.
1. This only does the python file processing for caffe2 when BUILD_CAFFE2_OPS=1, this part of the build file is written in such a way that it is always required to rerun and can take substantial time to move files around in the no-op build. In the future this part should be rewritten to use a faster method of copying the files or should treat copying the files as part of the build rules and only run when the files are out of date.
2. This points `sleef` to a patched version that fixes a dead build output that is causing everything to relink all the time. See https://github.com/shibatch/sleef/pull/231#partial-pull-merging for the upstream change.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/14969
Reviewed By: soumith
Differential Revision:
D13395998
Pulled By: zdevito
fbshipit-source-id:
ca85b7be9e99c5c578103c144ef0f2c3b927e724
vishwakftw [Sun, 9 Dec 2018 23:53:34 +0000 (15:53 -0800)]
Remove deprecated info argument in btrifact (#14935)
Summary:
As specified in title.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/14935
Differential Revision:
D13394449
Pulled By: soumith
fbshipit-source-id:
569d59414f3a1a43ea641bded4b5433eb53e3490
Soumith Chintala [Sun, 9 Dec 2018 23:52:25 +0000 (15:52 -0800)]
add fix for CUDA 10 (#14971)
Summary:
Linux binaries-only fix for CUDA10
Pull Request resolved: https://github.com/pytorch/pytorch/pull/14971
Differential Revision:
D13395932
Pulled By: soumith
fbshipit-source-id:
a72d6ab6b98c6c936e6391d55d2e4e45b9f1e6dd
Your Name [Sun, 9 Dec 2018 16:55:26 +0000 (08:55 -0800)]
Fix mismatched test_{full,ones,zeros}_like onnx expect files (#14956)
Summary:
master broken #14903
Pull Request resolved: https://github.com/pytorch/pytorch/pull/14956
Differential Revision:
D13395363
Pulled By: bddppq
fbshipit-source-id:
31f0913843292e557807fd5a976f8907fa6cae4b
Yiming Wu [Sun, 9 Dec 2018 16:23:36 +0000 (08:23 -0800)]
fix auto grad summing for IfOp where intermediate output needs renaming (#14772)
Summary:
fix auto grad summing for IfOp where intermediate output needs renaming.
Bug before this diff:
- we only renames the output of IfOp without changing the subnet ops output
- this results in blob not found error
the unittest provides an example
this diff fix that for IfOp
Pull Request resolved: https://github.com/pytorch/pytorch/pull/14772
Differential Revision:
D13327090
Pulled By: harouwu
fbshipit-source-id:
ec40ee88526ace3619c54551e223dd71158a02f8
Spandan Tiwari [Sun, 9 Dec 2018 06:46:03 +0000 (22:46 -0800)]
Export ones_like, zeros_like and full_like using ONNX ConstantLike op. (#14903)
Summary:
This PR does the following:
1) Updates the ONNX export for `torch.zeros_like` and `torch.full_like` ops to use ONNX op `ConstantLike`. This reduces the export of experimental op `ConstantFill`, which may possibly be removed in future, see https://github.com/onnx/onnx/pull/1434).
2) It also adds export support for `torch.ones_like`.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/14903
Differential Revision:
D13383700
Pulled By: houseroad
fbshipit-source-id:
566d00a943e9497172fcd5a034b638a650ab13a2
Edward Yang [Sun, 9 Dec 2018 03:32:01 +0000 (19:32 -0800)]
Canonicalize all includes in PyTorch. (#14849)
Summary:
Anywhere we used #include "foo.h", we now say #include <foo.h>
Paths are adjusted to be rooted out of aten/src, torch/lib, or
the root level directory.
I modified CMakeLists.txt by hand to remove TH and THC from
the include paths.
I used the following script to do the canonicalization:
```
import subprocess
import re
import os.path
files = subprocess.check_output(['git', 'ls-files']).decode('utf-8').rstrip().split('\n')
for fn in files:
if not any(fn.endswith(suff) for suff in ['.cu', '.cpp', '.in', '.h', '.hpp', '.cu', '.cuh', '.cc']):
continue
if not any(fn.startswith(pref) for pref in ["aten/", "torch/"]):
continue
with open(fn, 'r') as f:
c = f.read()
def fmt(p):
return "#include <{}>".format(p)
def repl(m):
p = m.group(1)
if p in ["dlfcn.h", "unistd.h", "nvrtc.h", "cuda.h", "cuda_runtime.h", "cstdint", "cudnn.h", "Python.h", "cusparse.h", "cuda_runtime_api.h", "cuda_fp16.h", "cublas_v2.h", "stdint.h", "curand_kernel.h"]:
return fmt(p)
if any(p.startswith(pref) for pref in ["torch/csrc", "c10/", "ATen/", "caffe2/", "TH/", "THC/", "Eigen/", "gtest/", "zdl/", "gloo/", "onnx/", "miopen/"]):
return fmt(p)
for root in ["aten/src", "torch/lib", ""]:
for bad_root in [os.path.dirname(fn), "aten/src/TH", "aten/src/THC", "torch/csrc"]:
new_p = os.path.relpath(os.path.join(bad_root, p), root)
if not new_p.startswith("../") and (os.path.exists(os.path.join(root, new_p)) or os.path.exists(os.path.join(root, new_p + ".in"))):
return fmt(new_p)
print("ERROR: ", fn, p)
return m.group(0)
new_c = re.sub(r'#include "([^"]+)"', repl, c)
if new_c != c:
print(fn)
with open(fn, 'w') as f:
f.write(new_c)
```
Signed-off-by: Edward Z. Yang <ezyang@fb.com>
Pull Request resolved: https://github.com/pytorch/pytorch/pull/14849
Reviewed By: dzhulgakov
Differential Revision:
D13363445
Pulled By: ezyang
fbshipit-source-id:
52361f878a672785f9306c9e9ab2513128092b68
Jongsoo Park [Sun, 9 Dec 2018 02:15:00 +0000 (18:15 -0800)]
race condition fix of calling mutable_data inside a openmp region (#14921)
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/14921
Fix race condition introduced in
D13188595 .
Let's reminder ourselves "never call mutable_data from an OpenMP region!!!"
Reviewed By: jianyuh
Differential Revision:
D13387692
fbshipit-source-id:
6a3aeedeeda55a9ede660de8f1f44d4eee76ae2b
Fei Sun [Sat, 8 Dec 2018 19:12:40 +0000 (11:12 -0800)]
Add crop argument, can crop rec as well, first resize and then crop
Summary: Pull Request resolved: https://github.com/pytorch/pytorch/pull/14894
Reviewed By: llyfacebook
Differential Revision:
D13377604
Pulled By: sf-wind
fbshipit-source-id:
333d0d864e6c2dc85f405baa25ed58029d62750f
Marat Dukhan [Sat, 8 Dec 2018 10:45:41 +0000 (02:45 -0800)]
Switch Int8Sigmoid to QNNPACK (#14883)
Summary:
50x-100x speedup compared to current version.
Also, fixes a bug in the current version when batch size exceeds 1 (current version processes only the first image in this case).
Pull Request resolved: https://github.com/pytorch/pytorch/pull/14883
Differential Revision:
D13390655
Pulled By: Maratyszcza
fbshipit-source-id:
1b33a97bf2d0866d38faa2b42e64fd2859017898
Your Name [Sat, 8 Dec 2018 09:04:02 +0000 (01:04 -0800)]
ONNX changes to use int32_t (instead of enum) to store data type
Summary: Pull Request resolved: https://github.com/pytorch/pytorch/pull/14926
Reviewed By: houseroad
Differential Revision:
D13390642
Pulled By: bddppq
fbshipit-source-id:
c2314b24d9384f188fda2b9a5cc16465ad39581e
Sebastian Messmer [Sat, 8 Dec 2018 08:26:14 +0000 (00:26 -0800)]
Remove at references from c10
Summary: Pull Request resolved: https://github.com/pytorch/pytorch/pull/14432
Reviewed By: dzhulgakov
Differential Revision:
D13223904
fbshipit-source-id:
43b06e33e088e7789ccea6d92267936fe30d8571
Brennan Vincent [Sat, 8 Dec 2018 04:13:31 +0000 (20:13 -0800)]
Implement `std` for multiple dimensions on CPU devices. (#14535)
Summary:
Tested on a tensor with 1 billion elements and 3 dimensions on a powerful, highly
multi-core Linux machine.
parallelized: All operations (e.g., `t.std(1)`) that could be done in the old code are now several times faster. All
new operations (e.g., `t.std((0,2))` are significantly faster than the NumPy equivalents.
`t.std((0, 1, 2))`, a new operation, is logically equivalent to the
old `t.std()`, but faster.
serial: The above comment about old operationos now being faster still
holds, but `t.std((t1, ..., tn))` is now a few
times slower than `t.std()`. If this turns out to be important, we can
special-case that to use the old algorithm.
The approach is to create a new method, `TensorIterator::foreach_reduced_elt`,
valid for `TensorIterator`s that represent a dimension reduction. This
method calls a supplied function for each element in the output,
supplying it with the input elements that correspond to that output.
Given that primitive, we can implement reductions like the following pseudocode:
If there is more than one output element:
```
PARALLEL FOR EACH element IN output:
accumulator = identity
SERIAL FOR EACH data_point IN element.corresponding_input:
accumulator.update(data_point)
element = accumulator.to_output()
```
If there is only one output element, we still want to parallelize, so we
do so along the *input* instead:
```
accumulators[n_threads]
PARALLEL FOR EACH input_chunk IN input.chunks():
accumulators[thread_num()] = identity
SERIAL FOR EACH data_point IN input_chunk:
accumulators[thread_num()].update_with_data(data_point)
accumulator = identity
SERIAL FOR EACH acc in accumulators:
accumulator.update_with_other_accumulator(acc)
output_element = accumulator.to_output()
```
Note that accumulators and data points do not have to be the same type
in general, since it might be necessary to track arbitrary amounts of
data at intermediate stages.
For example, for `std`, we use a parallel version of Welford's
algorithm, which requies us to track the mean, second moment, and number
of elements, so the accumulator type for `std` contains three pieces of
data.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/14535
Differential Revision:
D13283887
Pulled By: umanwizard
fbshipit-source-id:
8586b7bf00bf9f663c55d6f8323301e257f5ec3f
Orion Reblitz-Richardson [Sat, 8 Dec 2018 03:48:38 +0000 (19:48 -0800)]
Add CAFFE2_API to video processing functions (#14900)
Summary:
Extracted from https://github.com/pytorch/pytorch/pull/13733
Some tests were failing because these methods didn't have an export.
cc pjh5 yf225
Pull Request resolved: https://github.com/pytorch/pytorch/pull/14900
Reviewed By: pjh5
Differential Revision:
D13381130
Pulled By: orionr
fbshipit-source-id:
030536f8fb09765c09a7b0bd45400161053f2e18
Johannes M Dieterich [Sat, 8 Dec 2018 02:55:21 +0000 (18:55 -0800)]
Enable unit tests known to work on ROCm (#14011)
Summary:
* Enable unit tests known to work on ROCm.
* Disable a few that are known to be flaky for the time being.
* Use std::abs for Half
* No more special casing for ROCm in TensorMathReduce
* Document an important detail for a hardcoded block size w.r.t. ROCm in TensorMathReduce
ezyang bddppq for awareness
Pull Request resolved: https://github.com/pytorch/pytorch/pull/14011
Differential Revision:
D13387679
Pulled By: bddppq
fbshipit-source-id:
4177f2a57b09d866ccbb82a24318f273e3292f71
Lu Fang [Sat, 8 Dec 2018 01:24:01 +0000 (17:24 -0800)]
update of fbcode/onnx to
aca8473a40cf43f01958c81b648efcee7f3a755a (#14865)
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/14865
Previous import was
42804705bdbf179d1a98394008417e1392013547
Included changes:
- **[aca8473](https://github.com/onnx/onnx/commit/aca8473)**: Add Erf operator for computing error function (#1675) <bddppq>
- **[3fc82ca](https://github.com/onnx/onnx/commit/3fc82ca)**: Add IsNaN operator. (#1656) <Pranav Sharma>
- **[0685f01](https://github.com/onnx/onnx/commit/0685f01)**: Add Sign Op (#1658) <Rui Zhu>
- **[2a8fae8](https://github.com/onnx/onnx/commit/2a8fae8)**: Fix unused var warning (#1669) <Yinghai Lu>
- **[e212833](https://github.com/onnx/onnx/commit/e212833)**: Update scan (#1653) <G. Ramalingam>
Reviewed By: zrphercule
Differential Revision:
D13370727
fbshipit-source-id:
13a93d5acc8d4758f682278ea162ec9124ced22d
rohithkrn [Sat, 8 Dec 2018 01:23:49 +0000 (17:23 -0800)]
Enable fp16 for MIOPEN operators in Caffe2 (#14905)
Summary:
This PR enables fp16 MIOPEN operators in Caffe2.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/14905
Differential Revision:
D13383439
Pulled By: bddppq
fbshipit-source-id:
840afa8d08bef2952ca0039dee2423f1542bb330
Gu, Jinghui [Sat, 8 Dec 2018 00:42:39 +0000 (16:42 -0800)]
Upgrade MKL-DNN to version 0.17 (#14308)
Summary:
upgrade MKL-DNN to version 0.17
update mkldnn bridge to latest.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/14308
Differential Revision:
D13383102
Pulled By: yinghai
fbshipit-source-id:
c434f0e0ddff2ee2c86db2d6c44a37298fd005a3
Daniel Bermond [Sat, 8 Dec 2018 00:37:30 +0000 (16:37 -0800)]
Fix build with OpenCV 4.0 (#14356)
Summary:
Fixes #14355
Pull Request resolved: https://github.com/pytorch/pytorch/pull/14356
Differential Revision:
D13356237
Pulled By: bddppq
fbshipit-source-id:
2bf6ee21995c2c7b617c4e78ea7341f975f1b937
Sebastian Messmer [Sat, 8 Dec 2018 00:18:20 +0000 (16:18 -0800)]
Remove unused TensorImpl dependencies
Summary: Pull Request resolved: https://github.com/pytorch/pytorch/pull/14792
Reviewed By: ezyang
Differential Revision:
D13336843
fbshipit-source-id:
12f84799a70c2e90a8b934dd8dc031c09a6782f0
Sebastian Messmer [Sat, 8 Dec 2018 00:18:20 +0000 (16:18 -0800)]
Remove TensorImpl -> context_base dependency (#14658)
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/14658
Remove this dependency by moving at::CopyBytes to c10.
The implementations for at::CopyBytes will have to live in aten/caffe2 for now because they're not unified for CUDA yet.
They'll be moved into c10/backend/xxx later.
Reviewed By: dzhulgakov
Differential Revision:
D13288655
fbshipit-source-id:
1c92379345308b3cd39a402779d7b7999613fc0d
Sebastian Messmer [Sat, 8 Dec 2018 00:18:19 +0000 (16:18 -0800)]
Fix include paths for TensorOptions
Summary: Pull Request resolved: https://github.com/pytorch/pytorch/pull/14747
Reviewed By: ezyang
Differential Revision:
D13318645
fbshipit-source-id:
f5ba77a93f6019fbf5faffb47a2837c95fad474d
James Reed [Fri, 7 Dec 2018 23:06:48 +0000 (15:06 -0800)]
Update graph printouts in JIT docs (#14914)
Summary:
Tracing records variable names and we have new types and stuff in the IR, so this updates the graph printouts in the docs
Pull Request resolved: https://github.com/pytorch/pytorch/pull/14914
Differential Revision:
D13385101
Pulled By: jamesr66a
fbshipit-source-id:
6477e4861f1ac916329853763c83ea157be77f23
Ailing Zhang [Fri, 7 Dec 2018 22:56:56 +0000 (14:56 -0800)]
Improve hub documentation (#14862)
Summary:
Added a few examples and explains to how publish/load models.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/14862
Differential Revision:
D13384790
Pulled By: ailzhang
fbshipit-source-id:
008166e84e59dcb62c0be38a87982579524fb20e
James Reed [Fri, 7 Dec 2018 22:14:25 +0000 (14:14 -0800)]
USE_FBGEMM=True by default
Summary: Pull Request resolved: https://github.com/pytorch/pytorch/pull/14868
Differential Revision:
D13383390
Pulled By: jamesr66a
fbshipit-source-id:
1880c07dfd239e19153bd4fde2ab2c8d0604f956
Sergei Nikolaev [Fri, 7 Dec 2018 21:52:56 +0000 (13:52 -0800)]
USE_TENSORRT support and TensorRT 5 compatibility
Summary: Pull Request resolved: https://github.com/pytorch/pytorch/pull/13945
Differential Revision:
D13317525
Pulled By: yinghai
fbshipit-source-id:
8630dfec1bbc5aac19539e344e7c38a7fd8b051d
Orion Reblitz-Richardson [Fri, 7 Dec 2018 21:35:58 +0000 (13:35 -0800)]
Add __init__.py so files get picked up on install (#14898)
Summary:
This will let us install tests and other Caffe2 python code as a part of running Caffe2 tests in PyTorch.
Broken out of https://github.com/pytorch/pytorch/pull/13733/
cc pjh5 yf225
Pull Request resolved: https://github.com/pytorch/pytorch/pull/14898
Reviewed By: pjh5
Differential Revision:
D13381123
Pulled By: orionr
fbshipit-source-id:
0ec96629b0570f6cc2abb1d1d6fce084e7464dbe
Gregory Chanan [Fri, 7 Dec 2018 20:37:03 +0000 (12:37 -0800)]
Replace calls of Type::_th_tensor. (#14877)
Summary:
_th_tensor is moving off Type, so these calls need to be replaced.
Unfortunately, replacing these with a full-fledged solution [e.g. from_storage(..., TensorOptions)] is a bit complicated because the storage itself fully defines the Type (modulo variable). It's simpler to just wait for the Variable/Tensor merge rather than to solve this now, so instead I changed the call sites to: at::empty({0}, type.options()).set_(storage...).
This isn't great because we are also trying to get rid of Type::options, but this seems to be the lesser-of-two-evils.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/14877
Differential Revision:
D13374310
Pulled By: gchanan
fbshipit-source-id:
eb953ed041507e6190d6f32e383912e5a08311cd
Peter Goldsborough [Fri, 7 Dec 2018 20:22:49 +0000 (12:22 -0800)]
Large scale fix of python-related files in torch/csrc/
Summary: Pull Request resolved: https://github.com/pytorch/pytorch/pull/14515
Differential Revision:
D13247966
Pulled By: goldsborough
fbshipit-source-id:
7a127c508fc576a7a92626dd6b729f660162d628
PenghuiCheng [Fri, 7 Dec 2018 20:01:44 +0000 (12:01 -0800)]
Implementation of WeightedSum op for mkl-dnn and fix FC op output shape issue.
Summary: Pull Request resolved: https://github.com/pytorch/pytorch/pull/14407
Reviewed By: yinghai
Differential Revision:
D13364364
Pulled By: wesolwsk
fbshipit-source-id:
e69bcd1bc52e35b2f0e45e5dc40184f1bd66605d
Yudong Guang [Fri, 7 Dec 2018 17:58:54 +0000 (09:58 -0800)]
Revert
D13205604: Move numa.{h, cc} to c10/util
Differential Revision:
D13205604
Original commit changeset:
54166492d318
fbshipit-source-id:
89b6833518c0b554668c88ae38d97fbc47e2de17
vishwakftw [Fri, 7 Dec 2018 15:25:55 +0000 (07:25 -0800)]
Expose torch.roll function and method (#14880)
Summary: Fixes #14859 .
Differential Revision:
D13376915
Pulled By: zou3519
fbshipit-source-id:
f1fc0e8492a159431a3fc0a19a41aa10429ecc80
Junjie Bai [Fri, 7 Dec 2018 08:07:05 +0000 (00:07 -0800)]
Make autograd engine compatible with hip
Summary: Pull Request resolved: https://github.com/pytorch/pytorch/pull/14873
Differential Revision:
D13375053
Pulled By: bddppq
fbshipit-source-id:
f3051640386667bbf0566856ed433eb83276c39e
Jon Crall [Fri, 7 Dec 2018 07:55:34 +0000 (23:55 -0800)]
Fixed ConvT docstring (#14876)
Summary:
Fixes #14099
I attempted to be as consistent as possible with the formatting, hence why my equation reads d*(k - 1) instead of (k - 1)*d.
Also there is an unused variable on line 46: `n = self.in_channels`. I could fix that here too if that's not too out of scope.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/14876
Differential Revision:
D13374317
Pulled By: soumith
fbshipit-source-id:
a9f110acafa58cdb4206956dbe3ab4738d48292d
svcscm [Fri, 7 Dec 2018 06:50:09 +0000 (22:50 -0800)]
Updating submodules
Reviewed By: yns88
fbshipit-source-id:
7da015701f18f8a0b5a8092aae02a42ede7bfd44
David Riazati [Fri, 7 Dec 2018 05:50:35 +0000 (21:50 -0800)]
Remove weak module test expect files (#14871)
Summary:
This PR removes some expect files that aren't really testing anything
Pull Request resolved: https://github.com/pytorch/pytorch/pull/14871
Differential Revision:
D13373762
Pulled By: driazati
fbshipit-source-id:
e3537ee83df23b3b3b854f9b1253fd0cc8e9dd33
Wei Yang [Fri, 7 Dec 2018 01:58:16 +0000 (17:58 -0800)]
gradcheck (#14596)
Summary:
- allow gradcheck to take sparse tensor as input
- sparse output is not allowed yet at gradcheck
- add backward for `to_dense()` to get around sparse output
- calling gradcheck at test_sparse, so that we can use `_gen_sparse()` and also easily cover coalesced / uncoalesced test cases
Pull Request resolved: https://github.com/pytorch/pytorch/pull/14596
Differential Revision:
D13271904
Pulled By: weiyangfb
fbshipit-source-id:
5317484104404fd38058884c86e987546011dd86
Teng Li [Fri, 7 Dec 2018 01:22:04 +0000 (17:22 -0800)]
Skipping two c10d tests only if there are multi-GPUs (#14860)
Summary:
Otherwise, these tests will fail, even though there are never meant to run on single GPU machines.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/14860
Differential Revision:
D13369060
Pulled By: teng-li
fbshipit-source-id:
8a637a6d57335491ba8602cd09927700b2bbf8a0
Sebastian Messmer [Thu, 6 Dec 2018 23:52:15 +0000 (15:52 -0800)]
Move TensorOptions, DefaultTensorOptions to c10
Summary: Pull Request resolved: https://github.com/pytorch/pytorch/pull/14746
Reviewed By: ezyang
Differential Revision:
D13318644
fbshipit-source-id:
b703d7dc67e75d9e9571c80d62a100c5fc4e84df
Marat Dukhan [Thu, 6 Dec 2018 23:12:35 +0000 (15:12 -0800)]
Switch Int8MaxPool operator to QNNPACK (#14832)
Summary:
1.6-2.4X speedup on ARM when compiled with gcc
Pull Request resolved: https://github.com/pytorch/pytorch/pull/14832
Differential Revision:
D13358160
Pulled By: Maratyszcza
fbshipit-source-id:
39e9791886fac62650bb53a9df341889f0bb5d49
Richard Zou [Thu, 6 Dec 2018 22:55:55 +0000 (14:55 -0800)]
collect_env.py: get conda magma and mkl information (#14854)
Summary:
Fixes #12371
Pull Request resolved: https://github.com/pytorch/pytorch/pull/14854
Differential Revision:
D13363635
Pulled By: zou3519
fbshipit-source-id:
f8b5d05038bf5ce451399dfeed558ae298178128
zrphercule [Thu, 6 Dec 2018 22:04:44 +0000 (14:04 -0800)]
Add LogSigmoid support in ONNX symbolic (#14830)
Summary:
Add LogSigmoid:
torch.LogSigmoid(x) = onnx.Log(onnx.Sigmoid(x))
Pull Request resolved: https://github.com/pytorch/pytorch/pull/14830
Differential Revision:
D13353891
Pulled By: zrphercule
fbshipit-source-id:
bf456170b9e6c4edad07b3333cd5797f8e0fa97f
Ashwin Bharambe [Thu, 6 Dec 2018 21:44:33 +0000 (13:44 -0800)]
Kill GPU memory logs in normal runs (#14838)
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/14838
The GPU memory tracking logs are incredibly annoying and merely serve
to pollute output. I `VLOG(1)`ed them. Hopefully, this is non-controversial.
Reviewed By: kuttas
Differential Revision:
D13343290
fbshipit-source-id:
b3cae99346c97b66e97ea660061e15dc5c99b9fc
Junjie Bai [Thu, 6 Dec 2018 21:17:28 +0000 (13:17 -0800)]
Stop inserting static casts in Hipify (#14853)
Summary:
Latest hcc can now properly cast to correct type internally, so there is no need to insert static_cast in hipify scripts anymore.
However the hcc included in the latest ROCm release (1.9.2) doesn't have this fix, so leaving a flag to continue doing static_cast for those using the official ROCm releases.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/14853
Differential Revision:
D13363171
Pulled By: bddppq
fbshipit-source-id:
a36476a8511222ff3c933d31788e8a0ffb04f5ca
Jerry Zhang [Thu, 6 Dec 2018 19:16:07 +0000 (11:16 -0800)]
Tensor construction codemod - 3/3 (#14835)
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/14835
Codemod generated with clangr shard mode, 25 files per diff,
motivation: https://github.com/pytorch/pytorch/pull/12407
Reviewed By: bddppq
Differential Revision:
D13335184
fbshipit-source-id:
26d8247e16b30bdff045530034af9b72c76d066f
Jerry Zhang [Thu, 6 Dec 2018 19:14:48 +0000 (11:14 -0800)]
Tensor construction codemod - 1/3 (#14828)
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/14828
Codemod generated with clangr shard mode, 25 files per diff,
motivation: https://github.com/pytorch/pytorch/pull/12407
Reviewed By: bddppq
Differential Revision:
D13335160
fbshipit-source-id:
a3ae4c5a86bfbdaf2d5aa14e0eef57255e829fd4
Jerry Zhang [Thu, 6 Dec 2018 18:56:14 +0000 (10:56 -0800)]
Move numa.{h, cc} to c10/util (#14393)
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/14393
att
Reviewed By: ezyang
Differential Revision:
D13205604
fbshipit-source-id:
54166492d31827b0343ed070cc36a825dd86e2ed
Johannes M Dieterich [Thu, 6 Dec 2018 18:04:37 +0000 (10:04 -0800)]
Upgrade CI to ROCm 1.9.2 (#14216)
Summary:
Drop custom hcc/hip as the 1.9.2 release should contain the relevant patches therein.
Most notable feature in 1.9.2 is mixed precision support in rocBLAS and MIOpen. These features will be enabled by subsequent PRs.
bddppq ezyang
Pull Request resolved: https://github.com/pytorch/pytorch/pull/14216
Differential Revision:
D13354294
Pulled By: bddppq
fbshipit-source-id:
2541d4a196af21c9432c1aff7f6e65b572628028
Jan Schlüter [Thu, 6 Dec 2018 17:29:01 +0000 (09:29 -0800)]
Allow linspace and logspace with steps=1 and start != end like numpy (#14748)
Summary:
`torch.linspace(0, 1, 1)` fails with `RuntimeError: invalid argument 3: invalid number of points at ../aten/src/TH/generic/THTensorMoreMath.cpp:2119`, while `np.linspace(0, 1, 1)` works fine.
Looking at the code, there is even a comment by gchanan asking: "NumPy allows you to pass different points even if n <= 1 -- should we?"
I would say "yes". Currently, I would need to handle the case of `steps == 1` or `steps == 0` separately, making sure to change the `end` when calling `torch.linspace`. This is impractical. If we support `start != end`, there are two possibilities for the result: Either we ensure the first value in the resulting sequence always equals `start`, or we ensure the last value in the resulting sequence always equals `end`. Numpy chose the former, which also allows it to support a boolean `endpoint` flag. I'd say we should follow numpy.
This PR adapts `linspace` and `logspace` to mimic the behavior of numpy, adapts the tests accordingly, and extends the docstrings to make clear what happens when passing `steps=1`.
If you decide against this PR, the error message should become explicit about what I did wrong, and the documentation should be extended to mention this restriction.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/14748
Differential Revision:
D13356136
Pulled By: ezyang
fbshipit-source-id:
db85b8f0a98a5e24b3acd766132ab71c91794a82
Jie [Thu, 6 Dec 2018 16:57:39 +0000 (08:57 -0800)]
(#14580)
Summary:
Removes cast of half to float in torch.sum, with float16 input tensor and
float32 output tensor, instead we cast data when loading input in kernel.
This supposingly would save a kernel launch as well as a full global memory load
on promoted data type (float).
Pull Request resolved: https://github.com/pytorch/pytorch/pull/14580
Differential Revision:
D13356203
Pulled By: ezyang
fbshipit-source-id:
85e91225b880a65fe3ceb493371b9b36407fdf48
Ricardo Cuenca [Thu, 6 Dec 2018 16:57:31 +0000 (08:57 -0800)]
Consistent formatting in losses' docs
Summary: Pull Request resolved: https://github.com/pytorch/pytorch/pull/14739
Differential Revision:
D13356143
Pulled By: ezyang
fbshipit-source-id:
9ae8316dd8ba6e910247b64cec22db63df10e11c
Alex Şuhan [Thu, 6 Dec 2018 16:56:25 +0000 (08:56 -0800)]
Add (partial) autodiff support for nll_loss (#14305)
Summary:
Not ready yet, need some comments / help with this. It's good enough for https://github.com/pytorch/xla immediate goals (forward + backward trace fusion), but there are at least two issues with it:
1. If we don't allow it, `test/test_jit.py` fails to cover the change.
2. If we allow the weight to be set, running `test/test_jit.py TestJitGenerated.test_nn_nll_loss` fails with:
```
======================================================================
ERROR: test_nn_nll_loss (__main__.TestJitGenerated)
----------------------------------------------------------------------
Traceback (most recent call last):
File "test/test_jit.py", line 10001, in do_test
fn, f_args_variable, kwargs_variable, no_grad=no_grad)
File "test/test_jit.py", line 9360, in check_against_reference
outputs_test = self.runAndSaveRNG(func, recording_inputs, kwargs)
File "test/test_jit.py", line 425, in runAndSaveRNG
results = func(*inputs, **kwargs)
File "test/test_jit.py", line 9298, in script_fn
self.assertExportImport(CU.the_method.graph, tensors)
File "test/test_jit.py", line 415, in assertExportImport
self.assertExportImportModule(m, inputs)
File "test/test_jit.py", line 419, in assertExportImportModule
self.assertEqual(self.runAndSaveRNG(m.forward, inputs),
File "test/test_jit.py", line 425, in runAndSaveRNG
results = func(*inputs, **kwargs)
RuntimeError:
arguments for call are not valid:
for operator aten::nll_loss_backward(Tensor grad_output, Tensor self, Tensor target, Tensor? weight, int reduction, int ignore_index, Tensor total_weight, *, Tensor out) -> Tensor:
expected a value of type Tensor for argument 'total_weight' but found bool
<internally-created-node>
~ <--- HERE
for operator aten::nll_loss_backward(Tensor grad_output, Tensor self, Tensor target, Tensor? weight, int reduction, int ignore_index, Tensor total_weight) -> Tensor:
expected a value of type Tensor for argument 'total_weight' but found bool
<internally-created-node>
~ <--- HERE
for call at:
<internally-created-node>
~ <--- HERE
```
Pull Request resolved: https://github.com/pytorch/pytorch/pull/14305
Differential Revision:
D13356265
Pulled By: ezyang
fbshipit-source-id:
504d783b2d87f923e698a6a4efc0fd9935a94a41
svcscm [Thu, 6 Dec 2018 11:18:17 +0000 (03:18 -0800)]
Updating submodules
Reviewed By: yns88
fbshipit-source-id:
2adbb6f97d4b8f067a2538fec855063510b0ca3f
svcscm [Thu, 6 Dec 2018 10:53:28 +0000 (02:53 -0800)]
Updating submodules
Reviewed By: yns88
fbshipit-source-id:
e0509413215f3b7578b825c52365fec4da625bd5
projects / platform / upstream / pytorch.git / log