Lu Fang [Thu, 7 Mar 2019 19:03:57 +0000 (11:03 -0800)]
update of fbcode/onnx to
96c58ceeacf0f2b73d752e413e4fd78787a12da3 (#17676)
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/17676
Previous import was
e18bb41d255a23daf368ffd62a2645db55db4c72
Included changes:
- **[96c58ce](https://github.com/onnx/onnx/commit/96c58ce)**: Fix shape inference when auto_pad is notset again (#1830) <Li-Wen Chang>
- **[873ddbb](https://github.com/onnx/onnx/commit/873ddbb)**: More extendable Runner (#1809) <Michał Karzyński>
Reviewed By: zrphercule
Differential Revision:
D14321241
fbshipit-source-id:
12de9021afc61f5435f1b719cccf7b0f4ad73a84
Lu Fang [Thu, 7 Mar 2019 18:51:29 +0000 (10:51 -0800)]
Set the default ONNX opset to the latest stable opset (i.e., 9) (#17736)
Summary:
1) The changes in the new opset won't affect internal pipeline.
2) The CI won't be affected by the ONNX changes.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/17736
Reviewed By: zrphercule
Differential Revision:
D14358710
Pulled By: houseroad
fbshipit-source-id:
4ef15d2246b50f6875ee215ce37ecf92d555ca6a
David Riazati [Thu, 7 Mar 2019 18:41:13 +0000 (10:41 -0800)]
Add module attributes (#17309)
Summary:
Similar to `nn.Parameter`s, this PR lets you store any `IValue` on a module as an attribute on a `ScriptModule` (only from the Python front-end currently). To mark something as an attribute, it should wrapped in `jit.Attribute(value, type)` (ex. `self.table = torch.jit.Attribute(table, Dict[str, torch.Tensor])`)
Followup Work:
* (de)serializing for use in C++
* change `self.training` to be a `bool` attribute instead of a buffer
* mutable attributes
* string frontend support
* documentation
Pull Request resolved: https://github.com/pytorch/pytorch/pull/17309
Differential Revision:
D14354316
Pulled By: driazati
fbshipit-source-id:
67e08ab5229366b67fbc837e67b58831a4fb3318
Spandan Tiwari [Thu, 7 Mar 2019 18:06:17 +0000 (10:06 -0800)]
- refactoring serialization of ONNX initializers to be name-based (#17420)
Summary:
Currently, serialization of model parameters in ONNX export depends on the order in which they are stored in a container (`list` on Python side and `std::vector` on C++ side). This has worked fine till now, but if we need to do any pass on that graph that mutates the parameter list, then strictly order-based serialization may not work.
This PR is the first in a set to bring in more passes (such as constant folding) related to ONNX export. This PR lays the groundwork by moving the serialization in ONNX export from order-based to name based approach, which is more amenable to some of the passes.
houseroad - As discussed this change uses a map for export, and removes the code from `export.cpp` that relies on the order to compute initializer names.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/17420
Differential Revision:
D14361993
Pulled By: houseroad
fbshipit-source-id:
da93e945d55755c126de06641f35df87d1648cc4
Lara Haidar-Ahmad [Thu, 7 Mar 2019 17:59:28 +0000 (09:59 -0800)]
ONNX Export for Max and Average Pooling in CEIL_MODE
Summary: Pull Request resolved: https://github.com/pytorch/pytorch/pull/16769
Differential Revision:
D14362175
Pulled By: houseroad
fbshipit-source-id:
65cfb1dfba6a43d39cc85374add368fe8e4e5645
Elias Ellison [Thu, 7 Mar 2019 17:12:35 +0000 (09:12 -0800)]
use flake8-mypy (#17721)
Summary:
Use flake8 installed with mypy checks so that our linter matches fbcode. Mypy type errors also provide valuable signal
Pull Request resolved: https://github.com/pytorch/pytorch/pull/17721
Differential Revision:
D14357778
Pulled By: eellison
fbshipit-source-id:
d8c9ea3fe3b5f550c3b70fe259e0eabf95e4c92d
Jongsoo Park [Thu, 7 Mar 2019 10:17:42 +0000 (02:17 -0800)]
use fp16<->fp32 intrinsic (#17496)
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/17496
As title.
Reviewed By: hyuen
Differential Revision:
D14222907
fbshipit-source-id:
d5d6c032e725ca8b52aca2be7401ec3c59f6a242
Ahmed Aly [Thu, 7 Mar 2019 09:03:51 +0000 (01:03 -0800)]
Implement a Caffe2 standalone LSTM operator (#17726)
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/17726
Pull Request resolved: https://github.com/pytorch/pytorch/pull/17725
Pull Request resolved: https://github.com/pytorch/pytorch/pull/17461
Implementing a standalone LSTM Operator in Caffe2 adopted from this Aten implementation: diffusion/FBS/browse/master/fbcode/caffe2/aten/src/ATen/native/RNN.cpp. The most tricky thing in this exercise was that caffe2::Tensor has no copy constructor that made it necessary to implement a custom templated copy constructor for the different Tensor containers used in the code. Also there was no way to use off-the-shelf C2 operators in my code easily so I had to copy some code that is doing basic matmul, cat, split, transpose and linear as utility functions.
Two things missing:
- Profiling this implementation against the current ONNXified LSTM op
- Make this operator available to use in PyTorch
Reviewed By: dzhulgakov
Differential Revision:
D14351575
fbshipit-source-id:
3b99b53212cf593c7a49e45580b5a07b90809e64
Sebastian Messmer [Thu, 7 Mar 2019 07:50:14 +0000 (23:50 -0800)]
caffe2:libtorch_cuda depends on caffe2:caffe2_gpu (#17729)
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/17729
When doing "import torch" in fbcode, previously the caffe2 cuda kernels weren't loaded because libcaffe2_gpu.so wasn't loaded.
Once you also did "from caffe2.python import workspace", then the cuda kernels were loaded because that triggered a runtime mechanism for loading libcaffe2_gpu.so.
We want the cuda kernels to always be available, so this diff adds a dependency from caffe2:libtorch_cuda to caffe2:caffe2_gpu.
Reviewed By: ezyang
Differential Revision:
D14353498
fbshipit-source-id:
76a9fe69f231b308ab40eac393bb216c6fad3658
Jongsoo Park [Thu, 7 Mar 2019 07:26:27 +0000 (23:26 -0800)]
add tensor and cost inference functions (#17684)
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/17684
Adding tensor and cost inference functions to more int8 operators.
Reviewed By: yinghai
Differential Revision:
D14174746
fbshipit-source-id:
dfad975fa75899565c8fb61f1b7747a9206ebd22
Lara Haidar [Thu, 7 Mar 2019 06:35:12 +0000 (22:35 -0800)]
ONNX Export Narrow op
Summary: Pull Request resolved: https://github.com/pytorch/pytorch/pull/17550
Differential Revision:
D14350401
Pulled By: houseroad
fbshipit-source-id:
4d88079bb7a8bbd270b0272009826eb3b202cc33
Yinghai Lu [Thu, 7 Mar 2019 03:55:39 +0000 (19:55 -0800)]
Keep the dim_type of hinted shape as BATCH if possible (#17734)
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/17734
If input is not BATCH, we will skip adjust its batch size during onnxifi transformation. So when we take hints, we take it as CONSTANT but later need to change it to BATCH if possible.
Reviewed By: jackm321
Differential Revision:
D14355983
fbshipit-source-id:
63eb54a44afb1565c71486fdd73db07ca0ac4fd4
jwu [Thu, 7 Mar 2019 03:37:03 +0000 (19:37 -0800)]
fix different round behavior on CPU and GPU #16498 (#17443)
Summary:
xxtemp, colesbury, bhushan23, zou3519, convert gpu round behavior to half-to-even, consistent with torch cpu version and numpy. You feedback are welcomed.
See #16498
Pull Request resolved: https://github.com/pytorch/pytorch/pull/17443
Differential Revision:
D14261786
Pulled By: VitalyFedyunin
fbshipit-source-id:
98156436b545d72769831a89e2775d43ad913ebc
zou3519 [Thu, 7 Mar 2019 01:37:13 +0000 (17:37 -0800)]
Warn about memory overlaps on expanded tensors (#17576)
Summary:
Eventually we should remove these when we're certain that all our ops
handle memory overlaps correctly.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/17576
Differential Revision:
D14349990
Pulled By: zou3519
fbshipit-source-id:
c3a09f6113b9b1bf93e7f13c0b426c45b2cdf21f
Tongzhou Wang [Wed, 6 Mar 2019 23:35:25 +0000 (15:35 -0800)]
fix exp fam. formula
Summary: Pull Request resolved: https://github.com/pytorch/pytorch/pull/17719
Differential Revision:
D14349029
Pulled By: soumith
fbshipit-source-id:
cf016756a9319436f7379e8377f8bd1e1b672b40
Sebastian Messmer [Wed, 6 Mar 2019 23:08:44 +0000 (15:08 -0800)]
refactor caffe2 operator constructors - 10/9 (#17659)
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/17659
clangr codemod
Reviewed By: ezyang
Differential Revision:
D14304675
fbshipit-source-id:
45fbd84c50651a70ae29bf46df3322715e99d225
Lu Fang [Wed, 6 Mar 2019 22:59:16 +0000 (14:59 -0800)]
Improve ONNX symbolic for logsoftmax and softmax (#17672)
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/17672
support dtype in the onnx symbolic
Reviewed By: zrphercule
Differential Revision:
D14313987
fbshipit-source-id:
e9364621b3f795191d880599711dfbcb220d0e31
peter [Wed, 6 Mar 2019 22:40:05 +0000 (14:40 -0800)]
Enable using CMD when building cpp extensions on Windows
Summary: Pull Request resolved: https://github.com/pytorch/pytorch/pull/17706
Differential Revision:
D14346482
Pulled By: ezyang
fbshipit-source-id:
7c85e51c701f6c0947ad324ef19fafda40ae1cb9
Yinghai Lu [Wed, 6 Mar 2019 22:24:02 +0000 (14:24 -0800)]
Do not rename net boundary inputs/outputs during ssaRewrite. (#17545)
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/17545
This diff avoids renaming boundary inputs of net during onnxifi transform.
It also removes adding mappings for the initializer during onnxifi op creation.
Thus gets read of the mapped ws creation during onnxifi op creation.
Reviewed By: zrphercule
Differential Revision:
D14243161
fbshipit-source-id:
6eafa920c45f6a6bfacbbb443e8e84cf9778644c
Sebastian Messmer [Wed, 6 Mar 2019 21:47:27 +0000 (13:47 -0800)]
Reapply
D14078519 (#17596)
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/17596
Was reverted before, now fixed version.
Reviewed By: ezyang
Differential Revision:
D14270288
fbshipit-source-id:
c72490b5d02cc6098cb60145fa9a842b3c9a24c5
eellison [Wed, 6 Mar 2019 21:41:13 +0000 (13:41 -0800)]
Batch of expect file removals (#17581)
Summary:
Another batch of removing expect files.
One note - I removed the Batched expect files without adding equivalent tests since they are already being tested in another ways, and we are no longer actively maintaining that project.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/17581
Differential Revision:
D14343578
Pulled By: eellison
fbshipit-source-id:
ce0b1fd2b5b4ec80ad9003bab1b58f41645d3da6
jiej [Wed, 6 Mar 2019 21:36:14 +0000 (13:36 -0800)]
(#14267)
Summary:
- Summary:
Added synchronized batch normalization, allows synchronization of stats across mini-batches between processes within a process group.
Current implementation uses a mixture of extended ATen native functions (cpp cuda extension) + torch.nn.modules (c10d python API)
- User-facing api:
1. torch.nn.utils.convert_sync_batchnorm(modules, process_group=None)
2. torch.nn.SyncBatchNorm(num_features, eps=1e-5, momentum=0.1, affine=True, track_running_stats=True, ***process_group=None***)
- supported use case:
DistributedDataParallel with ***single-gpu multi-process***
a. User creates model containing `torch.nn.SyncBatchNorm` layers through one of the ways listed below:
1. use layers directly:
torch.nn.SyncBatchNorm(...)
similar API as with torch.nn.BatchNormXd(...)
with added argument `process_group` which is used to limit the scope of
synchronization within each process group. Default value is None, which
implies synchronization across all GPUs
2. use torch.nn.utils.convert_sync_batchnorm(modules, process_group)
recursively convert all `torch.nn.BatchNormXd` into `torch.nn.SyncBatchNorm`
preserving values of parameters/buffers.
the utility function also allows user to specify process_group value to all
converted layers.
b. user wraps their model with
`torch.distributed.parallel.DataParallelDistributed`, from this point, user
should follow the general guidelines for DDP use guide
- Error checking
For use cases not supported, we error out:
1. Application launched without ddp:
> import torch
> sbn = torch.nn.SyncBatchNorm(10).cuda()
> inp = torch.randn(5, 10, 3, 3).cuda()
> sbn(inp) --> Error!
> AttributeError: SyncBatchNorm is only supported within torch.nn.parallel.DistributedDataParallel
2. Application launched using DDP with multi-GPU per-process:
> ddp_module = nn.parallel.DistributedDataParallel(module, device_ids=device_ids, output_device=args.local_rank)
> ValueError: SyncBatchNorm is only supported for DDP with single GPU per process
Pull Request resolved: https://github.com/pytorch/pytorch/pull/14267
Differential Revision:
D14270035
Pulled By: ezyang
fbshipit-source-id:
4956d8fa565c32e9df5408d53719ff9f945f4d6d
Tongzhou Wang [Wed, 6 Mar 2019 21:06:41 +0000 (13:06 -0800)]
Update ModuleDict doc about order
Summary: Pull Request resolved: https://github.com/pytorch/pytorch/pull/17717
Differential Revision:
D14346557
Pulled By: ezyang
fbshipit-source-id:
2484c7d8105f9aa8bce5567d1fa2d4f587cc9cc2
Pieter Noordhuis [Wed, 6 Mar 2019 20:30:05 +0000 (12:30 -0800)]
Update CODEOWNERS (#17720)
Summary:
teng-li is passing the baton to mrshenli. Thanks for all your work on distributed teng-li!! :tada:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/17720
Differential Revision:
D14350120
Pulled By: pietern
fbshipit-source-id:
edfe784520c54630203cc8fbb296455d3dbf341b
Lara Haidar-Ahmad [Wed, 6 Mar 2019 20:05:48 +0000 (12:05 -0800)]
ONNX Export Argmin and Argmax ops
Summary: Pull Request resolved: https://github.com/pytorch/pytorch/pull/17382
Differential Revision:
D14338811
Pulled By: houseroad
fbshipit-source-id:
be07548d8063d1aa94f1801c18137738365b85fb
Lu Fang [Wed, 6 Mar 2019 20:02:34 +0000 (12:02 -0800)]
Turn atol to 1e-5 when comparing the end to end results (#17708)
Summary:
results smaller than 1e-5 don't make sense.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/17708
Differential Revision:
D14348893
Pulled By: houseroad
fbshipit-source-id:
5e07c38e5b58b27b61fae63bfc3c21e2fe5629fe
Elias Ellison [Wed, 6 Mar 2019 19:42:19 +0000 (11:42 -0800)]
remove loop expects (#17695)
Summary:
Replace loop unrolling expect files with assertions on the output IR
Pull Request resolved: https://github.com/pytorch/pytorch/pull/17695
Differential Revision:
D14347105
Pulled By: eellison
fbshipit-source-id:
1703b4ca32bc1c67c01fc4330b0e6eb66feaa103
youkaichao [Wed, 6 Mar 2019 19:31:50 +0000 (11:31 -0800)]
typo fix
Summary: Pull Request resolved: https://github.com/pytorch/pytorch/pull/17653
Differential Revision:
D14302003
Pulled By: ezyang
fbshipit-source-id:
8ad90985a392b07127c7e315d4e74ce77962b573
Deepali Chourasia [Wed, 6 Mar 2019 19:25:26 +0000 (11:25 -0800)]
omit group conv NHWC test for GPU (#17715)
Summary:
Observed the test `TestGroupConvolution.test_group_convolution` to fail with the following error:
```
Falsifying example: test_group_convolution(self=<caffe2.python.operator_test.group_conv_test.TestGroupConvolution testMethod=test_group_convolution>, stride=3, pad=0, kernel=5, size=8, group=4, input_channels_per_group=7, output_channels_per_group=8, batch_size=2, order='NHWC', engine='', use_bias=False, gc=, dc=[, device_type: 1])
You can reproduce this example by temporarily adding reproduce_failure('3.59.1', b'AAAA') as a decorator on your test case
```
This example generated by hypothesis has `group=2, order='NHWC' and dc=[, device_type: 1])`.
I think this example should be skipped.
I have mimicked the change corresponding to [PR#13554](https://github.com/pytorch/pytorch/pull/13554) to skip this example.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/17715
Differential Revision:
D14346642
Pulled By: ezyang
fbshipit-source-id:
b1f1fef09f625fdb43d31c7213854e61a96381ba
Elias Ellison [Wed, 6 Mar 2019 19:21:09 +0000 (11:21 -0800)]
fix tuple matching (#17687)
Summary:
Check for Tuple Matching in isSubvalueOf, since they may contain container types that need to be recursed within isSubvalueOf
Fix for https://github.com/pytorch/pytorch/issues/17650
Pull Request resolved: https://github.com/pytorch/pytorch/pull/17687
Differential Revision:
D14324642
Pulled By: eellison
fbshipit-source-id:
7f1e019875286b2640a3b9c003d1635dda8cf543
Spandan Tiwari [Wed, 6 Mar 2019 19:03:32 +0000 (11:03 -0800)]
Temporarily disable Upsample operator tests in pytorch-onnx tests (#17696)
Summary:
In discussion with houseroad, because Upsample op is being updated in ONNX https://github.com/onnx/onnx/pull/1773 and these tests are blocking it. These tests will be updated once the ONNX PR goes in.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/17696
Differential Revision:
D14338845
Pulled By: houseroad
fbshipit-source-id:
cfaf8cf1ab578ae69dd3bf21b1c0681b572b9b6f
peter [Wed, 6 Mar 2019 18:41:20 +0000 (10:41 -0800)]
Add check for x64 Python before setup (#17707)
Summary:
Fixes https://github.com/pytorch/pytorch/issues/17657.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/17707
Differential Revision:
D14346705
Pulled By: ezyang
fbshipit-source-id:
5daafacdb99eb9a9c6517263d10f20c79f920d24
Edward Yang [Wed, 6 Mar 2019 18:32:38 +0000 (10:32 -0800)]
Replace caffe2::DeviceGuard with c10::cuda::CUDAGuard (#17623)
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/17623
Despite it's generic sounding name, caffe2::DeviceGuard actually
only worked on CUDA devices. Rename it to something that more
clearly spells out its applicability.
I'm not sure if it's the right call, but in this patch I added
'using CUDAGuard = c10::cuda::CUDAGuard', as this seems to be more
in-line with how the Caffe2 codebase is currently written. More
idiomatic c10 namespace style would be to say cuda::CUDAGuard.
Willing to change this if people shout.
This is a respin of
D13156470 (#14284)
Reviewed By: dzhulgakov
Differential Revision:
D14285504
fbshipit-source-id:
93b8ab938b064572b3b010c307e1261fde0fff3d
Duc Ngo [Wed, 6 Mar 2019 18:31:00 +0000 (10:31 -0800)]
Remove nomscheduler (#17693)
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/17693
Remove nomscheduler tool
Reviewed By: yinghai
Differential Revision:
D14328168
fbshipit-source-id:
674d0e18596a4dc2bbb6b8d321f4066c4fc454ab
bhushan [Wed, 6 Mar 2019 18:28:49 +0000 (10:28 -0800)]
index operation support for torch.HalfTensor (#17645)
Summary:
- Test cases added
1. indexing for half tensor
2. setting for half tensor
fixes #17161
Pull Request resolved: https://github.com/pytorch/pytorch/pull/17645
Differential Revision:
D14302069
Pulled By: ezyang
fbshipit-source-id:
100f141c07046f200c904e27c5882a9417bccda0
Soumith Chintala [Wed, 6 Mar 2019 16:41:42 +0000 (08:41 -0800)]
Revert
D14160172: Implement a Caffe2 standalone LSTM operator
Differential Revision:
D14160172
Original commit changeset:
c33e3f9e8aea
fbshipit-source-id:
cffe35d93f0ac75ca93aa98a3b82af3d372f2fc1
Tongzhou Wang [Wed, 6 Mar 2019 07:14:25 +0000 (23:14 -0800)]
fix typo in hub doc
Summary: Pull Request resolved: https://github.com/pytorch/pytorch/pull/17705
Differential Revision:
D14338380
Pulled By: ailzhang
fbshipit-source-id:
d53eece30bede88a642e718ee6f829ba29c7d1c4
Ailing Zhang [Wed, 6 Mar 2019 04:47:02 +0000 (20:47 -0800)]
fix dropout AD & rename range to rangelist (#17691)
Summary:
fixes #17669
Address apaszke 's comments in #17523
Pull Request resolved: https://github.com/pytorch/pytorch/pull/17691
Differential Revision:
D14328083
Pulled By: ailzhang
fbshipit-source-id:
9ec4a54f13bfd1aaf4b1821dd00c31793ac07a44
Chaitanya Sri Krishna Lolla [Wed, 6 Mar 2019 02:41:20 +0000 (18:41 -0800)]
enable use of MIOpen for depthwise convolutions (#17685)
Summary:
* added miopen conv mode to be used for setConvDescriptor
* added miopen depthwise convolutions
Pull Request resolved: https://github.com/pytorch/pytorch/pull/17685
Differential Revision:
D14327811
Pulled By: bddppq
fbshipit-source-id:
d5bdc1abafd5f39694fadf3f9275b9d880c5b115
Ahmed Aly [Wed, 6 Mar 2019 01:31:51 +0000 (17:31 -0800)]
Implement a Caffe2 standalone LSTM operator (#17461)
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/17461
Implementing a standalone LSTM Operator in Caffe2 adopted from this Aten implementation: diffusion/FBS/browse/master/fbcode/caffe2/aten/src/ATen/native/RNN.cpp. The most tricky thing in this exercise was that caffe2::Tensor has no copy constructor that made it necessary to implement a custom templated copy constructor for the different Tensor containers used in the code. Also there was no way to use off-the-shelf C2 operators in my code easily so I had to copy some code that is doing basic matmul, cat, split, transpose and linear as utility functions.
Two things missing:
- Profiling this implementation against the current ONNXified LSTM op
- Make this operator available to use in PyTorch
Reviewed By: dzhulgakov
Differential Revision:
D14160172
fbshipit-source-id:
c33e3f9e8aeae578b64d97593cb031a251216029
Soumith Chintala [Tue, 5 Mar 2019 22:26:20 +0000 (14:26 -0800)]
Fix nll_loss crash on cpu where ignore_index is out of bounds (#17328)
Summary:
Fixes https://github.com/pytorch/pytorch/issues/15508
Pull Request resolved: https://github.com/pytorch/pytorch/pull/17328
Differential Revision:
D14322629
Pulled By: soumith
fbshipit-source-id:
7d02f372be78794782c18affcfc109ce30b1e91c
Johannes M Dieterich [Tue, 5 Mar 2019 20:49:25 +0000 (12:49 -0800)]
Add '--hip-clang-launch' to favor <<<>>>-based launch. (#17686)
Summary:
hip-clang uses triple chevron kernel dispatch syntax. Add an option to the hipification script to skip translating triple chevron to hipLaunchKernelGGL.
Once we switch to hip-clang, this option will be default and subsequently removed.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/17686
Differential Revision:
D14327810
Pulled By: bddppq
fbshipit-source-id:
5e1512325077dd3ebb8fb9b5bf35fd1f8d9a4dc3
Sam Gross [Tue, 5 Mar 2019 17:38:23 +0000 (09:38 -0800)]
Improve caching allocator for Pascal and newer GPUs. (#17120)
Summary:
```
NVIDIA changed the CUDA allocation behavior on Pascal GPUs. The
page size increased from 1MB to 2MB and allocations larger than 1MB
are now always page-aligned. Previously, allocations larger than 1MB
were aligned to 128KB boundaries.
This interacted poorly with the caching allocator. The remaining
memory in a page could only be filled by small cudaMalloc calls, but
the caching allocator never cudaMalloc's a chunk smaller than 1MB.
This behavior could also cause a large discrepancy between the memory
usage reported by nvidia-smi and the memory usage reported by
PyTorch, because nvidia-smi counts a partially used page as "full",
while PyTorch only counts the actual memory requested.
This PR makes a few changes to the caching allocator to better support
Pascal and Volta GPUs:
- All cudaMalloc calls are now multiples of 2MB (the page size)
- Requests between 1-10MB allocate (and split) a 20MB block to
reduce wasted space due to rounding
- Small requests are now packed into 2MB blocks (instead of 1MB)
This improves Mask R-CNN memory usage by 10-20% in internal tests on
Volta GPUs. Maxwell performance seems to be largely unchanged, but
it's possible that some use cases suffer slightly.
```
Pull Request resolved: https://github.com/pytorch/pytorch/pull/17120
Differential Revision:
D14301536
Pulled By: colesbury
fbshipit-source-id:
a8282315ea8f7b8ca149b5066fdeaecd0d404edf
Davide Libenzi [Tue, 5 Mar 2019 15:24:27 +0000 (07:24 -0800)]
Turn the Half::from_bits into a constexpr function to avoid unresolve… (#17661)
Summary:
…d symbol errors when building in DEBUG mode.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/17661
Differential Revision:
D14319610
Pulled By: soumith
fbshipit-source-id:
6c508a37155e29260f403d7174f343aa1ff32385
Elias Ellison [Tue, 5 Mar 2019 06:38:41 +0000 (22:38 -0800)]
Remove Expect Files from python / tracing / script interop
Summary: Pull Request resolved: https://github.com/pytorch/pytorch/pull/17622
Differential Revision:
D14308307
Pulled By: eellison
fbshipit-source-id:
bda249d38ac2570000a12b0ca328c26233ecefe8
peterjc123 [Tue, 5 Mar 2019 05:50:53 +0000 (21:50 -0800)]
Enable apex on Windows
Summary: Pull Request resolved: https://github.com/pytorch/pytorch/pull/17675
Differential Revision:
D14320473
Pulled By: soumith
fbshipit-source-id:
cb696984f5196f9b8b50722b4fe927bb6407c322
Soumith Chintala [Tue, 5 Mar 2019 04:28:06 +0000 (20:28 -0800)]
bump docker build to upgrade magma to 2.5.0 (#17674)
Summary:
upgrades magma in docker build.
vishwakftw
Pull Request resolved: https://github.com/pytorch/pytorch/pull/17674
Differential Revision:
D14320187
Pulled By: soumith
fbshipit-source-id:
7887f65fb703b802fc6231408b55ad9c4039882b
Sebastian Messmer [Mon, 4 Mar 2019 23:56:21 +0000 (15:56 -0800)]
refactor caffe2 operator constructors - 1/9 (#17082)
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/17082
clangr codemod
Reviewed By: ezyang
Differential Revision:
D14078498
fbshipit-source-id:
f7f65d6d81c7942293f53fdaa61f756d8b7360c1
Sebastian Messmer [Mon, 4 Mar 2019 22:53:55 +0000 (14:53 -0800)]
Expose cuda kernel for caffe2::GenerateProposals
Summary: Pull Request resolved: https://github.com/pytorch/pytorch/pull/17066
Reviewed By: ezyang, wat3rBro
Differential Revision:
D14071130
fbshipit-source-id:
6fe26503f6069c36ec31d6c09b549b932d5db242
Jongsoo Park [Mon, 4 Mar 2019 22:25:19 +0000 (14:25 -0800)]
print warnings when DNNLOWP_16 or DNNLOWP_ROWWISE_16 engine is used (#17176)
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/17176
As title
Reviewed By: csummersea
Differential Revision:
D14111616
fbshipit-source-id:
1282cb2452c4ad385fd2dc6d3f8c19e9fec715ff
Sebastian Messmer [Mon, 4 Mar 2019 22:17:11 +0000 (14:17 -0800)]
Fix XOutput/XOutputTensor for ivalue based c2 operators (#17599)
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/17599
XOutput/XOutputTensor was broken for ivalue based operators. This diff fixes that.
Reviewed By: ezyang
Differential Revision:
D14274003
fbshipit-source-id:
b99f020244c66c4e2551dbd32ae0f665cc91b338
Sebastian Messmer [Mon, 4 Mar 2019 22:17:11 +0000 (14:17 -0800)]
Fix InputSize/OutputSize for ivalue based operators (#17579)
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/17579
These methods previously just returned 0 when it was not a legacy operator,
making it impossible to convert some operators.
Reviewed By: dzhulgakov
Differential Revision:
D14253094
fbshipit-source-id:
72bfdcf6da291a4ab80d1e0ceb20984b86edc408
Wanchao Liang [Mon, 4 Mar 2019 21:04:53 +0000 (13:04 -0800)]
Fix clamp fusion on missing limits (#17533)
Summary:
Fixes #17449
Context: before #17186, we don't fuse `clamp` for the case when `min/max` are missing inputs, because they are `prim::None` node, after #17186, we make None a `prim::Constant` node which enables the fusion for `clamp`. But codegen.cpp does not handle the case when `prim::Constant` is not a Double/Int/Bool, this PR makes it so that missing inputs are handled correctly, it is done in the following way:
1. emit nothing when you see `type? = prim::Constant()`
2. when emitRHS, do special casing for aten::clamp
Pull Request resolved: https://github.com/pytorch/pytorch/pull/17533
Differential Revision:
D14238450
Pulled By: wanchaol
fbshipit-source-id:
61a272154754b13e89021bb86002927f02cde19c
Jie [Mon, 4 Mar 2019 21:02:40 +0000 (13:02 -0800)]
int32 indexing for Tensor Iterator Reduction (#17428)
Summary:
1. Enabling int32 indexing for cases where TI cannot accumulate in output due to
incompatible data types (e.g. Welford).
2. Updating Welford kernel to use int32 instead of int64 indexing on GPU.
This change improves performance for torch.var / torch.std
Implementation:
1. Allocated extra buffer to handle accumulation between sub Tensor Iterators.
2. Removed int64 indexing in gpu_reduce_kernel
3. WelfordOps now supports index type / combination typeas a template parameter.
While GPU uses int32_t and float, CPU implementation uses int64_t and double.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/17428
Differential Revision:
D14264608
Pulled By: umanwizard
fbshipit-source-id:
3eb54451de925b469dbc1127e5ea7443c4431036
Iurii Zdebskyi [Mon, 4 Mar 2019 20:43:28 +0000 (12:43 -0800)]
Removed all usages of TH_Index_Base (#17591)
Summary:
TH_Index_Base is hard coded to 0 and can be removed from the code base.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/17591
Differential Revision:
D14269273
Pulled By: izdeby
fbshipit-source-id:
d844e261f4af7297bad8a81e7d6dcf0a391b94e6
Dmytro Dzhulgakov [Mon, 4 Mar 2019 19:30:43 +0000 (11:30 -0800)]
PyTorch/Caffe2 tensor interop in Python (#17190)
Summary:
Because of two separate python extensions with different pybind
instances I have to go through void* conversion. Since it's hidden from
user, it's fine.
New APIs added on C2 side:
- workspace.FetchTorch('blob')
- workspace.Workspace.current.blobs['blob'].to_torch()
- workspace.FeedBlob('blob', pytorch_tensor)
Works on CPU an GPU.
The only glitches are with resizing because of variable/tensor split.
But data sharing works properly.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/17190
Reviewed By: ezyang
Differential Revision:
D14163882
Pulled By: dzhulgakov
fbshipit-source-id:
d18e5b8fcae026f393c842a1149e972515732de2
wkcn [Mon, 4 Mar 2019 18:08:04 +0000 (10:08 -0800)]
Fixed typo in aten/src/ATen/native_parse.py (#17641)
Summary:
Hi, there.
There is a typo in aten/src/ATen/native_parse.py, and I fix it.
`std::aray` -> `std::array`
Pull Request resolved: https://github.com/pytorch/pytorch/pull/17641
Differential Revision:
D14301981
Pulled By: ezyang
fbshipit-source-id:
a37859cdedcbf6c29333b954486dfa086d6c2176
Martin Schatz [Mon, 4 Mar 2019 17:55:05 +0000 (09:55 -0800)]
Remove GPU dependency from ProfileObserver (#17592)
Summary:
Remove GPU dependency and register ProfileObserver.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/17592
Reviewed By: ezyang
Differential Revision:
D14265801
Pulled By: mdschatz
fbshipit-source-id:
f98c0c32653c64a8b087c58ece4f864dfbe1d4b8
Brennan Vincent [Mon, 4 Mar 2019 06:13:27 +0000 (22:13 -0800)]
Don't make factory methods create a tensor and then immediately copy it (#17565)
Summary:
Create a `make_variable` override that moves out of a tensor instead of going through `shallow_copy_and_detach`. Call this override from factory methods like `empty` that create a brand new tensor, do nothing with it, and then copy it into a variable.
Will update this with actual numbers, but it seems to get rid of around 20-40% of the overhead of calling `torch.empty(0)`
Pull Request resolved: https://github.com/pytorch/pytorch/pull/17565
Differential Revision:
D14266130
Pulled By: umanwizard
fbshipit-source-id:
f57d5f2ca3f80ee8ee96d50f905e852fd10db941
Jack Richter-Powell [Sun, 3 Mar 2019 18:05:36 +0000 (10:05 -0800)]
Fixed typo in torch/functional.py w/r/t broadcast_tensors (#17642)
Summary:
In reference to #17574
Pull Request resolved: https://github.com/pytorch/pytorch/pull/17642
Differential Revision:
D14297177
Pulled By: ezyang
fbshipit-source-id:
968176ea3b46a0153da0fd9e6b40db314d29e51c
Bryan He [Sun, 3 Mar 2019 09:01:26 +0000 (01:01 -0800)]
Change fake tqdm constructor to match real tqdm (#17636)
Summary:
Currently, the fake tqdm implementation requires an input (whereas real tqdm does not).
This caused a problem in torchvision (https://github.com/pytorch/vision/pull/770), and seems likely to cause minor irritations elsewhere.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/17636
Differential Revision:
D14296530
Pulled By: ezyang
fbshipit-source-id:
bc077d898773c93dab34c985a7b30525a43e558a
Christian Puhrsch [Sun, 3 Mar 2019 02:14:02 +0000 (18:14 -0800)]
Mark native_functions as matched if uncaptured by JIT (#17631)
Summary:
Various functions aren't used by the JIT, so they're jit-compliant w.r.t. their schema by default.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/17631
Differential Revision:
D14295559
Pulled By: cpuhrsch
fbshipit-source-id:
a2ecdcb5df47eb67c54ec642d88d42e985515142
Christian Puhrsch [Sat, 2 Mar 2019 03:18:47 +0000 (19:18 -0800)]
Ban std::array from native_functions.yaml
Summary: Pull Request resolved: https://github.com/pytorch/pytorch/pull/17629
Differential Revision:
D14292941
Pulled By: cpuhrsch
fbshipit-source-id:
3c3eed57a5505a4e1da3aea682092677ab0e73e3
Christian Puhrsch [Sat, 2 Mar 2019 03:12:08 +0000 (19:12 -0800)]
Remove more usages of BoolTensor and IndexTensor from native_functions.yaml
Summary: Pull Request resolved: https://github.com/pytorch/pytorch/pull/16468
Differential Revision:
D14095405
Pulled By: cpuhrsch
fbshipit-source-id:
ea4d6bb7a4e81c05fe9861190ddbf52201612bbf
Thomas Viehmann [Sat, 2 Mar 2019 02:57:02 +0000 (18:57 -0800)]
Implement kthvalue in ATen (#17544)
Summary:
The CPU version is based on the TH version.
The GPU version is based on #8406 by Pararth Shah (thank you).
CPU quickselect based on that in TH's THTensorMoreMath.cpp, but with C++ (quickselectnoindex will be achieved by a different swap)
CPU kthvalue is based on the THTensor function in the same file.
The dim_apply function is a C++ replacement for TH_TENSOR_DIM_APPLYx macros.
The CUDA kernel uses functions adapted from the THCTensorSortK implementation.
In particular radixSelect is from THCTensorTopK.cuh.
The CUDA launcher code replaces a bunch of macros with C++. It will be re-used in one of the following patches.
Plan for further PRs:
- This
- Sort
- TopK + Mode + Median in any order
- Rip out THC stuff.
There may be utility functions / structs in the SortingCommon.cuh that come into
relevance only with sort.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/17544
Differential Revision:
D14286934
Pulled By: ezyang
fbshipit-source-id:
35dbea050b097e88777ac5fa5c0f499d5e23c738
Christian Puhrsch [Sat, 2 Mar 2019 00:53:23 +0000 (16:53 -0800)]
Change vml.h to support sizes greater than 2**32 - 1
Summary: Pull Request resolved: https://github.com/pytorch/pytorch/pull/17280
Differential Revision:
D14154997
Pulled By: cpuhrsch
fbshipit-source-id:
c19b15d18da59c9ee87e82765d3244d2a4ef6729
Grigory Arutyunov [Fri, 1 Mar 2019 23:07:18 +0000 (15:07 -0800)]
msvc_fixes (#17201)
Summary:
Fixing MSVC errors
```
D:\pytorch-scripts\caffe2_builders\v141\pytorch\aten\src\THC/THCReduce.cuh(144): error C4002: too many actual paramet
ers for macro 'C10_LAUNCH_BOUNDS_1' [D:\pytorch-scripts\caffe2_builders\v141\pytorch\build\Debug\caffe2\caffe2_gpu.vcxp
roj]
D:\pytorch-scripts\caffe2_builders\v141\pytorch\aten\src\THC/THCReduce.cuh(259): error C4002: too many actual paramet
ers for macro 'C10_LAUNCH_BOUNDS_1' [D:\pytorch-scripts\caffe2_builders\v141\pytorch\build\Debug\caffe2\caffe2_gpu.vcxp
roj]
D:/pytorch-scripts/caffe2_builders/v141/pytorch/aten/src/THCUNN/SpatialDilatedMaxPooling.cu(51): error C4002: too man
y actual parameters for macro 'C10_LAUNCH_BOUNDS_1' [D:\pytorch-scripts\caffe2_builders\v141\pytorch\build\Debug\caffe2
\caffe2_gpu.vcxproj]
```
on variadic C10_LAUNCH_BOUNDS as well as Debug linking issues with at::Half in pool_op_cudnn.cc like this one
```
pool_op_cudnn.obj : error LNK2019: unresolved external symbol "public: bool __cdecl caffe2::MaxPoolFunctor<class caff
e2::CUDAContext>::GlobalPoolingBackward<struct c10::Half,2>(int,int,int,struct c10::Half const *,struct c10::Half const
,struct c10::Half const ,struct c10::Half ,class caffe2::CUDAContext )const " (??$GlobalPoolingBackward@UHalf@c10@
@$01@?$MaxPoolFunctor@VCUDAContext@caffe2@@caffe2@QEBA_NHHHPEBUHalf@c10@00PEAU23@PEAVCUDAContext@1@Z) referenced in
function "public: bool __cdecl caffe2::`anonymous namespace'::CuDNNMaxPoolFunctor::GlobalPoolingBackward<struct c10::H
alf,2>(int,int,int,struct c10::Half const ,struct c10::Half const ,struct c10::Half const ,struct c10::Half ,class
caffe2::CUDAContext *)const " (??$GlobalPoolingBackward@UHalf@c10@@$01@CuDNNMaxPoolFunctor@?A0xb936404a@caffe2@QEBA_NH
HHPEBUHalf@c10@00PEAU34@PEAVCUDAContext@2@Z) [D:\pytorch-scripts\caffe2_builders\v141\pytorch\build\Debug\caffe2\caff
e2_gpu.vcxproj]
```
Pull Request resolved: https://github.com/pytorch/pytorch/pull/17201
Differential Revision:
D14165732
Pulled By: ezyang
fbshipit-source-id:
875fd9a5b2db6f83fc483f6d750d2c011260eb8b
Jithun Nair [Fri, 1 Mar 2019 23:00:30 +0000 (15:00 -0800)]
Hipify fixes for Masquerade logic (#17598)
Summary:
ezyang Please review.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/17598
Differential Revision:
D14287724
Pulled By: ezyang
fbshipit-source-id:
46e5083854a827370bb4c81b82e5a4ede511e473
Wanchao Liang [Fri, 1 Mar 2019 23:00:01 +0000 (15:00 -0800)]
Rename prim::Undefined to prim::AutogradZero (#17611)
Summary:
supersedes #17245
Pull Request resolved: https://github.com/pytorch/pytorch/pull/17611
Differential Revision:
D14283581
Pulled By: wanchaol
fbshipit-source-id:
8022d02b8a021ea2fee9a18a2c8920eb123200c5
Roy Li [Fri, 1 Mar 2019 22:18:58 +0000 (14:18 -0800)]
Add python test for extension backend tensor.device (#17602)
Summary:
Adding a test for #17361
Pull Request resolved: https://github.com/pytorch/pytorch/pull/17602
Differential Revision:
D14287373
Pulled By: li-roy
fbshipit-source-id:
544ecf17eb310aed22ba0ea5f86f46b8e3bb69b5
Edward Yang [Fri, 1 Mar 2019 22:14:02 +0000 (14:14 -0800)]
Revert
D13935403: Call c10 cuda op from test_torch
Differential Revision:
D13935403
Original commit changeset:
b2915ec8a366
fbshipit-source-id:
0f3409d5c102d719bc1f0483695aee93e7d613c9
Amy Yang [Fri, 1 Mar 2019 21:53:11 +0000 (13:53 -0800)]
add command line option to use hive filler; add README (#17619)
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/17619
--filler hive --iter -1 will let debugger exhaust all batches from a hive partition before exiting.
add README that summarizes command line options and usage.
Reviewed By: yinghai
Differential Revision:
D14220166
fbshipit-source-id:
daa23b7e8a9184481c6d7b67acf1599e5c99d74a
Thomas Viehmann [Fri, 1 Mar 2019 20:32:47 +0000 (12:32 -0800)]
Remove TH(CU)NN Sparse Linear (#17610)
Summary:
Sparse Linear in TH(CU)NN implements sparse linear layers without
using sparse matrices.
It is currently not documented in PyTorch and there is no functional or
module interface. This means it is unused from a PyTorch point of view.
The reason for removing it is twofold:
- The module uses sort, which I would like to move to ATen.
- When we implement a SparseLinear layer, we would want to do it
using sparse tensors, so it's not all that useful, anyway.
I checked this on slack with soumith, I hope the above is an accurate
representation. All bad ideas are my own.
This is part of the ongoing work to move
sort/topk/mode/median/kthvalue to ATen.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/17610
Differential Revision:
D14280663
Pulled By: gchanan
fbshipit-source-id:
289231d2c20626855ce2ceecd4f204b460c32378
ZhuBaohe [Fri, 1 Mar 2019 19:26:54 +0000 (11:26 -0800)]
Correct docstring of vision/init functions
Summary: Pull Request resolved: https://github.com/pytorch/pytorch/pull/17351
Differential Revision:
D14276355
Pulled By: soumith
fbshipit-source-id:
9b572b6a04eeb1e44cd93961edac76ed10f7b24e
Sebastian Messmer [Fri, 1 Mar 2019 18:51:34 +0000 (10:51 -0800)]
Call c10 cuda op from test_torch
Summary: Pull Request resolved: https://github.com/pytorch/pytorch/pull/16692
Reviewed By: ezyang
Differential Revision:
D13935403
fbshipit-source-id:
b2915ec8a3664bb6e918ed357908cc33d8f9449a
peter [Fri, 1 Mar 2019 18:33:58 +0000 (10:33 -0800)]
Revert #17191 and #17215 that no longer apply on Windows (#17567)
Summary:
They are previously merged to resolve #17051. However, since it was resolved by the upstream, and it was causing some issues like https://github.com/abjer/tsds/issues/8, I think it's time to revert these changes.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/17567
Differential Revision:
D14265241
Pulled By: kostmo
fbshipit-source-id:
7fa2b7dd4ebc5148681acb439cf82d983898694e
Michael Suo [Fri, 1 Mar 2019 18:00:19 +0000 (10:00 -0800)]
usertype -> class (#17528)
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/17528
as title. register_prim_ops is messy because someone ruined clang-format, but I figured it's okay to include here since this is such a mechanical change
Reviewed By: driazati
Differential Revision:
D14236943
fbshipit-source-id:
c2b22845837b7f830015510e48ec2ee5202fa407
Michael Suo [Fri, 1 Mar 2019 18:00:19 +0000 (10:00 -0800)]
alias analysis refactor take 2 (#17594)
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/17594
The original version of this broke things because a concurrent change raced with it in CI.
Reviewed By: ezyang
Differential Revision:
D14266663
fbshipit-source-id:
e8ac5dfcb7349b4f2c425d9f0eabbfc964314063
peter [Fri, 1 Mar 2019 17:56:43 +0000 (09:56 -0800)]
Fix the missing Windows CPU job in the build status section (#17608)
Summary:
It will be better to split the CPU job on CI. But unluckily, we are out of Windows machines.
cc, davidbrownellWork yf225
Pull Request resolved: https://github.com/pytorch/pytorch/pull/17608
Differential Revision:
D14281393
Pulled By: soumith
fbshipit-source-id:
ae9a6140b7207ce56cfb2da3d812bc3fe060764a
peter [Fri, 1 Mar 2019 17:45:38 +0000 (09:45 -0800)]
Update magma to 2.5.0 for Windows
Summary: Pull Request resolved: https://github.com/pytorch/pytorch/pull/17607
Differential Revision:
D14281291
Pulled By: yf225
fbshipit-source-id:
51209c5540932871e45e54ba6d61b3b7d264aa8c
bhushan [Fri, 1 Mar 2019 16:38:06 +0000 (08:38 -0800)]
Adding support for 0-d tensor for transpose (.t()) (#17535)
Summary:
- Test updates
1. test_torch: added 0-d test case and t_() test cases
2. test_jit : updated error message for TestAsync.test_async_script_error
- Updating documentation for torch.t()
Adding information regarding new support of 0-D and 1-D tenso
Fixes #17520
Pull Request resolved: https://github.com/pytorch/pytorch/pull/17535
Differential Revision:
D14269984
Pulled By: gchanan
fbshipit-source-id:
38b723f31484be939261c88edb33575d242eca65
svcscm [Fri, 1 Mar 2019 09:33:59 +0000 (01:33 -0800)]
Updating submodules
Reviewed By: yns88
fbshipit-source-id:
05fafcfb34c76f425ac5c8ef24a5f920641c2cf7
Junjie Bai [Fri, 1 Mar 2019 08:02:56 +0000 (00:02 -0800)]
Mark cudaGetLastError return value unused in C10_CUDA_CHECK
Summary: Pull Request resolved: https://github.com/pytorch/pytorch/pull/17605
Reviewed By: xw285cornell
Differential Revision:
D14277586
Pulled By: bddppq
fbshipit-source-id:
38879208f2ab83cf39d8a8a61b288cd09fcafd9a
Huan Gui [Fri, 1 Mar 2019 07:17:35 +0000 (23:17 -0800)]
add dropout during eval (#17549)
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/17549
Currently Dropout is only enabled in training, we enable the option of having dropout in Eval.
This is to follow [1]. This functionality would be used for uncertainty estimation in exploration project.
[1] Gal, Yarin, and Zoubin Ghahramani. "Dropout as a bayesian approximation: Representing model uncertainty in deep learning." international conference on machine learning. 2016.
Reviewed By: Wakeupbuddy
Differential Revision:
D14216216
fbshipit-source-id:
87c8c9cc522a82df467b685805f0775c86923d8b
Johannes M Dieterich [Fri, 1 Mar 2019 06:53:34 +0000 (22:53 -0800)]
Adjust launch_bounds annotation for AMD hardware. (#17555)
Summary:
The max pooling backwards kernel is currently annotated with launch bounds (256,8).
Adjust the number of waves to 4 (4 times 64 is 256) for ROCm. This improves training performance for torchvision models by up to 15% (AlexNet) on a gfx906 GPU.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/17555
Differential Revision:
D14277744
Pulled By: bddppq
fbshipit-source-id:
2a62088f7b8a87d1e350c432bf655288967c7883
Sebastian Messmer [Fri, 1 Mar 2019 00:26:49 +0000 (16:26 -0800)]
Fix verbose compiler warning in flat_hash_map (#17562)
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/17562
fixes https://github.com/pytorch/pytorch/issues/17332
Reviewed By: ezyang
Differential Revision:
D14254499
fbshipit-source-id:
9d5d7408c2ce510ac20cd438c6514dc2bbe3a854
Sebastian Messmer [Fri, 1 Mar 2019 00:26:49 +0000 (16:26 -0800)]
Fix diagnostic pragmas (#17561)
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/17561
The push at the top of the file was missing a corresponding pop
Reviewed By: ezyang
Differential Revision:
D14254500
fbshipit-source-id:
ff20359b563d6d6dcc68273dc754ab31aa8fad12
Sebastian Messmer [Fri, 1 Mar 2019 00:25:37 +0000 (16:25 -0800)]
Allow dispatch based on tensor list args (#17522)
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/17522
Dispatch is still based on the first tensor arg, but that first "tensor arg" is now allowed to be a tensor list.
That is, the first argument that is either Tensor or TensorList will be the deciding factor for dispatch.
If it is a TensorList, then that TensorList must not be empty or dispatch will fail.
Reviewed By: ezyang
Differential Revision:
D14235840
fbshipit-source-id:
266c18912d56ce77aa84306c5605c4191f3d882b
Sebastian Messmer [Fri, 1 Mar 2019 00:25:37 +0000 (16:25 -0800)]
Allow exposing caffe2 operators with variable number of input tensors to c10 (#17491)
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/17491
Before, there was no way to expose a caffe2 operator that had a variable number of inputs.
Now, this is allowed by giving the operator one tensor list input.
Note that the tensor list must be the first input, and that any other tensor inputs will be ignored and inaccessible in this case.
Reviewed By: ezyang
Differential Revision:
D14220705
fbshipit-source-id:
7f921bfb581caf46b229888c409bbcc40f7dda80
Syed Tousif Ahmed [Fri, 1 Mar 2019 00:17:37 +0000 (16:17 -0800)]
blacklist fft algorithms for strided dgrad (#17016)
Summary:
Applies https://github.com/pytorch/pytorch/pull/16626 from v1.0.1
Pull Request resolved: https://github.com/pytorch/pytorch/pull/17016
Differential Revision:
D14270100
Pulled By: ezyang
fbshipit-source-id:
1137899dd1551d33d16f39e8dde76cad8192af46
Sebastian Messmer [Thu, 28 Feb 2019 22:58:12 +0000 (14:58 -0800)]
Revert
D14078519: [codemod][caffe2] [clangr] refactor caffe2 operator constructors - 5/9
Differential Revision:
D14078519
Original commit changeset:
b0ca31a52e4a
fbshipit-source-id:
713ae108d3dd6f33abdbf98a5f213e57e2b64642
David Riazati [Thu, 28 Feb 2019 22:43:05 +0000 (14:43 -0800)]
Add generic list/dict custom op bindings (#17587)
Summary:
Fixes #17017
Sandcastle refuses to land #17037, so trying fresh here
Pull Request resolved: https://github.com/pytorch/pytorch/pull/17587
Differential Revision:
D14265402
Pulled By: driazati
fbshipit-source-id:
b942721aa9360ac6b3862f552ac95529eb0cf52c
Sebastian Messmer [Thu, 28 Feb 2019 22:12:37 +0000 (14:12 -0800)]
refactor caffe2 operator constructors - 8/9 (#17089)
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/17089
clangr codemod
Reviewed By: ezyang
Differential Revision:
D14078539
fbshipit-source-id:
9ca196af4af7f26fc82e6cf82b35d478d0597752
Sebastian Messmer [Thu, 28 Feb 2019 22:12:14 +0000 (14:12 -0800)]
refactor caffe2 operator constructors - 6/9 (#17087)
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/17087
clangr codemod
Reviewed By: ezyang
Differential Revision:
D14078525
fbshipit-source-id:
7cc03b30b0d4eb99818e35406be4119b27bdb1bc
Sebastian Messmer [Thu, 28 Feb 2019 22:06:51 +0000 (14:06 -0800)]
refactor caffe2 operator constructors - 2/9 (#17083)
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/17083
clangr codemod
Reviewed By: ezyang
Differential Revision:
D14078504
fbshipit-source-id:
34dddb035eee2fca3150e47c57489614b91b6725
Sebastian Messmer [Thu, 28 Feb 2019 22:04:06 +0000 (14:04 -0800)]
refactor caffe2 operator constructors - 7/9 (#17088)
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/17088
clangr codemod
also manually moved the constructor of a class from the .cpp file to the .h file.
Reviewed By: ezyang
Differential Revision:
D14078531
fbshipit-source-id:
2adb4ac0ce523742da6cce3bc3b6c177b816c299
Sebastian Messmer [Thu, 28 Feb 2019 22:03:34 +0000 (14:03 -0800)]
refactor caffe2 operator constructors - 4/9 (#17085)
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/17085
clangr codemod
Reviewed By: ezyang
Differential Revision:
D14078515
fbshipit-source-id:
aaa48ae10892e3f47063f2133e026fea46f3240b
Sebastian Messmer [Thu, 28 Feb 2019 22:01:30 +0000 (14:01 -0800)]
refactor caffe2 operator constructors - 3/9 (#17084)
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/17084
clangr codemod
Reviewed By: ezyang
Differential Revision:
D14078507
fbshipit-source-id:
ed02d772890b30196302b6830f541f054b7e95c8
Edward Yang [Thu, 28 Feb 2019 21:32:22 +0000 (13:32 -0800)]
Make HIPStream also masquerade as CUDA. (#17469)
Summary:
HIPGuard interfaces that interacted with HIPStream were previously
totally busted (because the streams had the wrong device type).
This fixes it, following along the same lines of MasqueardingAsCUDA.
Along the way I beefed up the explanatory comment.
Signed-off-by: Edward Z. Yang <ezyang@fb.com>
cc jithunnair-amd iotamudelta bddppq
Pull Request resolved: https://github.com/pytorch/pytorch/pull/17469
Differential Revision:
D14243396
Pulled By: ezyang
fbshipit-source-id:
972455753a62f8584ba9ab194f9c785db7bb9bde
Alex Şuhan [Thu, 28 Feb 2019 21:28:17 +0000 (13:28 -0800)]
Fix Python device type property for XLA and MSNPU
Summary: Pull Request resolved: https://github.com/pytorch/pytorch/pull/17361
Differential Revision:
D14243546
Pulled By: soumith
fbshipit-source-id:
b7498968f72e3d97de5bf6e5b44c5a59b6913acb