Summary:
`torch.tensor([True, False, True], dtype=torch.bool).sum()` should return **2** instead of **True** as it does now.
Tested via unit tests
Pull Request resolved: https://github.com/pytorch/pytorch/pull/21421
Differential Revision: D15674203
Pulled By: izdeby
fbshipit-source-id: b00e3d0ca809c9b92b750adc05632522dad50c74
Summary:
Something flaky is going on with `test_inplace_view_saved_output` on Windows.
With my PR #20598 applied, the test fails, even though there is no obvious reason it should be related, so the PR was reverted.
Based on commenting out various parts of my change and re-building, I think the problem is with the name -- renaming everything from `T` to `asdf` seems to make the test stop failing. I can't be sure that this is actually the case though, since I could just be seeing patterns in non-deterministic build output...
I spoke with colesbury offline and we agreed that it is okay to just disable this test on Windows for now and not block landing the main change. He will look into why it is failing.
**Test Plan:** I will wait to make sure the Windows CI suite passes before landing this.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/21175
Differential Revision: D15566970
Pulled By: umanwizard
fbshipit-source-id: edf223375d41faaab0a3a14dca50841f08030da3
Summary:
This PR improves performance of advanced indexing backward, partially solving #15245 (performance is still worse than gather, but not by such outrageous margins). Before, using benchmarking harness from #15245, cuda 10/V100:
```
Indexing is faster by at most -270.61607820767887 us on N: 16 D: 256 K: 1
Indexing is slower by at most 11127.466280784833 us on N: 16 D: 4096 K: 4096
```
after:
```
Indexing is faster by at most 23.524456737696028 us on N: 512 D: 4096 K: 4096
Indexing is slower by at most 186.24056029472553 us on N: 16 D: 1024 K: 4096
```
Strategy is to reuse embedding backward kernel, adapting it to handle unindexed dimensions in the beginning by launching additional threadblocks, and also allowing it to handle slices that are bigger than `65K*128`, that is hardly ever a problem for embedding. Still, integer indexing is baked in the kernel, and is important for performance, so for now bigger than 2G element tensors are not supported.
The main savings come from not having to expand index to all unindexed dimensions, and not sorting expanded index with incoming gradient values, but rather only sorting unexpanded index.
There are ways to make sorting overhead smaller (thanks mcarilli for suggestions) but I'll get to it when it becomes a real problem, or rather, when cuda graphs will force us to get rid of thrust::sort calls.
I've also added tests for indexing backward, before tests for index_put_ and indexing backward were non-existent.
This PR also fixes#20457 by casting indices to `self` backend.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/20557
Differential Revision: D15582434
Pulled By: ezyang
fbshipit-source-id: 91e8f2769580588ec7d18823d99a26f1c0da8e2a
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/21196
we'll add `quantize(quantizer)` as a tensor method later when we expose `quantizer` in Python frontend
Python
```
torch.quantize_linear(t, ...)
```
C++
```
at::quantize_linear(t, ...)
```
Differential Revision: D15577123
fbshipit-source-id: d0abeea488418fa9ab212f84b0b97ee237124240
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/21156
we'll add `quantize(quantizer)` as a tensor method later when we expose `quantizer` in Python frontend
Python
```
torch.quantize_linear(t, ...)
```
C++
```
at::quantize_linear(t, ...)
```
Differential Revision: D15558784
fbshipit-source-id: 0b194750c423f51ad1ad5e9387a12b4d58d969a9
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/20874
A criteria for what should go in Tensor method is whether numpy has it, for this one it does not
so we are removing it as a Tensor method, we can still call it as function.
Python
```
torch.quantize_linear(t, ...), torch.dequantize(t)
```
C++
```
at::quantize_linear(t, ...), at::dequantize(t)
```
Reviewed By: dzhulgakov
Differential Revision: D15477933
fbshipit-source-id: c8aa81f681e02f038d72e44f0c700632f1af8437
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/20869
Adding support for the functions listed in the title, by implementing the copy kernel.
Differential Revision: D15474060
fbshipit-source-id: 9264df6e442cca1cc5d952e3e5dcc9f4a426f317
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/21035
Fix the dtype error in `dequantize_linear`, it should accept the same dtype argument as `quantize_linear`
Differential Revision: D15521931
fbshipit-source-id: 0114c046a3f1046e42fca49c74c85e487fee8616
Summary:
This PR covers two important points with respect to the QR decomposition:
- batching of input matrices (#7500)
- adding `some` as an option in `torch.qr` akin to NumPy's `mode` option (#10538)
Changelog:
- Enable batching for inputs to `torch.qr`
- Move QR decomposition implementation to ATen (CPU and CUDA)
- Remove existing implementations in TH/THC
- Add a `some` option to `torch.qr` that will enable users to switch between complete and reduced decomposition
- Modify doc strings
Pull Request resolved: https://github.com/pytorch/pytorch/pull/20689
Differential Revision: D15529230
Pulled By: soumith
fbshipit-source-id: 16af82b1d2db8a3a758fa8a5f798d83f5f950efb
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/20938
Dequantize_linear need not be exposed to the front end users.
It will only be used for the jit passes for q-dq insertion and op
substitution.
Differential Revision: D15446097
fbshipit-source-id: a5fbcf2bb72115122c9653e5089d014e2a2e891d
Summary:
Bug reported internally at FB:
```python
>>> t=torch.from_numpy(np.empty((0,4)))
>>> t[:,1::2]*=1
Traceback (most recent call last):
File "<stdin>", line 1, in <module>
RuntimeError: Trying to resize storage that is not resizable at ../aten/src/TH/THStorageFunctions.cpp:76
```
This happens because the storage offset of `t[:, 1::2]` is 1, and it has 0 elements. We can fix this by avoiding resizing the storage for no-element arrays.
(We could *also* have avoided it by not modifying the storage index in this case, but I felt this way was more semantically correct -- in general, we should not be assuming it's okay to do anything to the storage when it has zero elements).
Pull Request resolved: https://github.com/pytorch/pytorch/pull/20914
Differential Revision: D15497860
Pulled By: umanwizard
fbshipit-source-id: 6af61d73a05edfc5c07ce8be9e530f15bf72e6a9
Summary:
This PR also moves Device::validate into the header file, which makes
statements like `Device d = kCPU` effectively free.
Device includes the device's index, so TensorIterator::compute_types
now implicitly checks that all CUDA inputs are on the same GPU.
Previously, this was done ad-hoc in places like TensorIterator::binary_op.
Note that zero-dim Tensor (scalars) are NOT required to be on the
same device as other inputs because they behave almost like Python numbers.
TensorIterator handles copying zero-dim Tensors to the common device.
Prior to this PR, TensorIterator would copy zero-dim Tensors between CPU
and GPU, but not between different GPUs (because Backend didn't encode
the GPU index). This removes that restriction.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/20690
Differential Revision: D15414826
Pulled By: colesbury
fbshipit-source-id: 1d0ad1f7d663252af36dd4590bcda418c2f7a09f
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/20740
Provide a way to assemble quantized Tensor from int8 Tensor, scale and zero point.
Differential Revision: D15232416
fbshipit-source-id: c3a3d9d7214b1dc569214c019440c2779fbd063b
Summary:
CUDA 8 is no longer supported and removed from CI, so these checks are irrelevant
Pull Request resolved: https://github.com/pytorch/pytorch/pull/20482
Differential Revision: D15393438
Pulled By: ezyang
fbshipit-source-id: ac0979bf660b3314eec502c745e34ce4940bda0e
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/19932
In preparation to add int8_t data type for QTensor
Reviewed By: zafartahirov
Differential Revision: D15137838
fbshipit-source-id: 59462c36d6fc5982986d4196bf3f32f49bb294d7
Summary:
#19975 was separated by 2 PRs.
This one:
Introduce MemoryFormat argument to the `x.is_contiguous(memory_format=torch.channels_last)` and to the `y = x.contiguous(memory_format=torch.channels_last)` functions.
At this moment both functions just operate with strides and doesn't store any tensor state.
(Original RFC #19092)
-----
Expands functionality of two tensor functions `.is_contiguous` and `.contiguous` (both python and c++ api).
Note: We had several complaints about `.to(memory_format)` function, and decided not to support it.
1. `.contiguous` now support optional keyword-only argument - `memory_format`, which can be either `torch.contiguous_format` or `torch.channels_last`.
- Using `torch.contiguous_format` will preserve existing `.contiguous()` behavior.
- Calling `x.contiguous(memory_format=torch.channels_last)` returns new tensor which maintain same semantical layout (NCHW), but have different memory allocation pattern.
`x.contiguous(memory_format=torch.channels_last)` expects input tensor to be 3d, 4d or 5d; and fails otherwise.
2. `.is_contiguous` now support optional keyword-only argument - `memory_format`, which can be either `torch.contiguous_format` or `torch.channels_last`.
- `x.is_contiguous(memory_format=torch.contiguous_format)` preserves same functionality as `x.is_contiguous()` and remains unchanged.
- `x.is_contiguous(memory_format=torch.channels_last)` returns true if A) input tensor is contiguous in memory AND B) allocated in the memory in NWHC (or similar for 3d,5d) format.
Note: By the end of the phase one `x.is_contiguous(memory_format=torch.channels_last)` will calculate state of the Tensor on every call. This functionality going to be updated later.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/20455
Differential Revision: D15341577
Pulled By: VitalyFedyunin
fbshipit-source-id: bbb6b4159a8a49149110ad321109a3742383185d
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/19816
We need this for quantization for bias
add third argument of ScalarType to `quantize_linear`
Differential Revision: D15094174
fbshipit-source-id: f19ec8f4716cf5fe0aa21b38d45af6d27c9ab377
Summary:
The current variance kernels compute mean at the same time. Many times we want both statistics together, so it seems reasonable to have a kwarg/function that allows us to get both values without launching an extra kernel.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/18731
Differential Revision: D14726082
Pulled By: ifedan
fbshipit-source-id: 473cba0227b69eb2240dca5e61a8f4366df0e029
Summary:
Add automatic translations for a few argument names that commonly differ between PyTorch and NumPy.
For now, they are as follows:
* `keepdim` -> `keepdims`
* `dim` -> `axis`
* `input` -> (any of `a`, `x`, `x1`)
* `other` -> `x2`
Basic examples:
```python
>>> t=torch.randn(10,10)
>>> torch.sum(x=t, axis=1)
tensor([ 0.5199, -0.3768, 4.3619, -0.9105, 1.1804, 1.0837, -0.9036, 0.2365,
1.1171, -0.0999])
```
```python
>>> torch.add(x1=5, x2=6)
tensor(11)
```
The additional overhead is zero when using traditional PyTorch argument names, and a few (usually 1) extra PyDict lookups when using NumPy argument names.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/20451
Differential Revision: D15337521
Pulled By: umanwizard
fbshipit-source-id: 7a7d389786f4ccf5c86a14ecb2002c61730c51b5
Summary:
This addresses #18436
The logic replicates the essence of closing file descriptors in numpy:
bf20e30340/numpy/core/include/numpy/npy_3kcompat.h (L278)
This stores the position of the file descriptor before resetting it to the Python handle offset, then resets to the original position before exit. The Python-side handle is then updated to reflect the new position. Also added somewhat more demanding tests to cover this.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/20270
Differential Revision: D15275902
Pulled By: soumith
fbshipit-source-id: 5ca8a52b61c7718d2e69571f72f80b1350b0acdb
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/19513
Add support for printing a QTensor in python frontend
Differential Revision: D15017168
fbshipit-source-id: 312d1f18e6ca3c9eb4a5b8bb1c64f7cc8bc1dcf5
Summary:
log_normal_ and geometric_ were disabled for CPU by mistake in [this PR](bc53805f2e), this PR fixes it.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/19938
Differential Revision: D15143404
Pulled By: izdeby
fbshipit-source-id: 41c7bd29f046b5a3ac6d601de8c64ab553771d19
Summary:
Added deprecation warnings for the masked methods and enabled them for a bool tensor.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/19140
Differential Revision: D14888021
Pulled By: izdeby
fbshipit-source-id: 0e42daf8f3732ca29f36d10485402bfc502716ad
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/19676
Make copy work with QTensor, enable assignment of QTensor in pytorch frontend.
Differential Revision: D15064710
fbshipit-source-id: 04f2dc02a825695d41fa1114bfca49e92108fef3
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/19530
Make copy work with QTensor, enable assignment of QTensor in pytorch frontend.
Differential Revision: D15008160
fbshipit-source-id: 5f1166246d768b23f009cde1fa03e8952368a332
Summary:
Add base support for torch.logspace. See #19220 for details.
SsnL can you feedback? Thanks a lot.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/19542
Differential Revision: D15028484
Pulled By: soumith
fbshipit-source-id: fe5a58a203b279103abbc192c754c25d5031498e
Summary:
Changelog:
- Rename `potri` to `cholesky_inverse` to remain consistent with names of `cholesky` methods (`cholesky`, `cholesky_solve`)
- Fix all callsites
- Rename all tests
- Create a tentative alias for `cholesky_inverse` under the name `potri` and add a deprecation warning to not promote usage
Pull Request resolved: https://github.com/pytorch/pytorch/pull/19498
Differential Revision: D15029901
Pulled By: ezyang
fbshipit-source-id: 2074286dc93d8744cdc9a45d54644fe57df3a57a
Summary:
Attempt fix for #14057 . This PR fixes the example script in the issue.
The old behavior is a bit confusing here. What happened to pickling is python2 failed to recognize `torch.float32` is in module `torch`, thus it's looking for `torch.float32` in module `__main__`. Python3 is smart enough to handle it.
According to the doc [here](https://docs.python.org/2/library/pickle.html#object.__reduce__), it seems `__reduce__` should return `float32` instead of the old name `torch.float32`. In this way python2 is able to find `float32` in `torch` module.
> If a string is returned, it names a global variable whose contents are pickled as normal. The string returned by __reduce__() should be the object’s local name relative to its module
Pull Request resolved: https://github.com/pytorch/pytorch/pull/18045
Differential Revision: D14990638
Pulled By: ailzhang
fbshipit-source-id: 816b97d63a934a5dda1a910312ad69f120b0b4de
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/18960
empty_affine_quantized creates an empty affine quantized Tensor from scratch.
We might need this when we implement quantized operators.
Differential Revision: D14810261
fbshipit-source-id: f07d8bf89822d02a202ee81c78a17aa4b3e571cc
Summary:
This adds checks for `mul_`, `add_`, `sub_`, `div_`, the most common
binops. See #17935 for more details.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/19317
Differential Revision: D14972399
Pulled By: zou3519
fbshipit-source-id: b9de331dbdb2544ee859ded725a5b5659bfd11d2
Summary:
Make it possible to construct a pinned memory tensor without creating a storage first and without calling pin_memory() function. It is also faster, as copy operation is unnecessary.
Supported functions:
```python
torch.rand_like(t, pin_memory=True)
torch.randn_like(t, pin_memory=True)
torch.empty_like(t, pin_memory=True)
torch.full_like(t, 4, pin_memory=True)
torch.zeros_like(t, pin_memory=True)
torch.ones_like(t, pin_memory=True)
torch.tensor([10,11], pin_memory=True)
torch.randn(3, 5, pin_memory=True)
torch.rand(3, pin_memory=True)
torch.zeros(3, pin_memory=True)
torch.randperm(3, pin_memory=True)
torch.empty(6, pin_memory=True)
torch.ones(6, pin_memory=True)
torch.eye(6, pin_memory=True)
torch.arange(3, 5, pin_memory=True)
```
Part of the bigger: `Remove Storage` plan.
Now compatible with both torch scripts:
` _1 = torch.zeros([10], dtype=6, layout=0, device=torch.device("cpu"), pin_memory=False)`
and
` _1 = torch.zeros([10], dtype=6, layout=0, device=torch.device("cpu"))`
Same checked for all similar functions `rand_like`, `empty_like` and others
It is fixed version of #18455
Pull Request resolved: https://github.com/pytorch/pytorch/pull/18952
Differential Revision: D14801792
Pulled By: VitalyFedyunin
fbshipit-source-id: 8dbc61078ff7a637d0ecdb95d4e98f704d5450ba
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/18546
We'll expose all combinations of various ways of quantization in the top level dispatch key, that is we have AffineCPUTensor, PerChannelAffineCUDATensor, etc.
QTensor method added:
- is_quantized()
- item()
Differential Revision: D14637671
fbshipit-source-id: 346bc6ef404a570f0efd34e8793056ad3c7855f5
Summary:
I've been messing around with vectorizing the fusion compiler in JIT, and noticed that these ops were pathologically slow. I moved them to use TensorIterator + Vec256<> and got some speed wins.
Benchmark script:
```
import torch, time
ops = ['abs', 'neg', 'reciprocal', 'frac']
x = torch.rand(1024, 1024)
NITER = 10000
print('op', 'time per iter (ms)', 'gops/s', 'GB/s', sep='\t')
for op in ops:
s = time.time()
for i in range(NITER):
getattr(x, op)()
elapsed_sec = ((time.time() - s) / NITER)
print(op, elapsed_sec * 1000, (1024*1024/elapsed_sec)/1e9, (1024*1024*4*2) / elapsed_sec / 1e9, sep='\t')
```
Before this change (on my mac with a skylake):
```
op time per iter (ms) gops/s GB/s
abs 0.9730974197387695 1.0775652866097343 8.620522292877874
neg 1.0723679780960083 0.9778136063534356 7.822508850827485
reciprocal 1.2610594034194946 0.8315040490215421 6.6520323921723366
frac 1.1681334018707275 0.8976509004200546 7.181207203360437
```
After this change:
```
op time per iter (ms) gops/s GB/s
abs 0.5031076192855835 2.084198210889721 16.673585687117768
neg 0.4433974027633667 2.3648672578256087 18.91893806260487
reciprocal 0.47145988941192624 2.2241043693195985 17.79283495455679
frac 0.5036592721939087 2.0819154096627024 16.65532327730162
```
So, after this change it looks like we are hitting machine peak for bandwidth and are bandwidth bound.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/19041
Differential Revision: D14862037
Pulled By: jamesr66a
fbshipit-source-id: e2032ac0ca962dbf4120bb36812277c260e22912
Summary:
Changelog:
- Rename `btrisolve` to `lu_solve` to remain consistent with names of solve methods (`cholesky_solve`, `triangular_solve`, `solve`)
- Fix all callsites
- Rename all tests
- Create a tentative alias for `lu_solve` under the name `btrisolve` and add a deprecation warning to not promote usage
Pull Request resolved: https://github.com/pytorch/pytorch/pull/18726
Differential Revision: D14726237
Pulled By: zou3519
fbshipit-source-id: bf25f6c79062183a4153015e0ec7ebab2c8b986b
Summary:
Partial fix of: https://github.com/pytorch/pytorch/issues/394
- `gels` and `triangular_solve` now returns namedtuple
- refactor test for namedtuple API for better coverage and maintainability
Pull Request resolved: https://github.com/pytorch/pytorch/pull/17195
Differential Revision: D14851875
Pulled By: ezyang
fbshipit-source-id: 9b2cba95564269d2c3a15324ba48751d68ed623c
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/18832
ghimport-source-id: fde4ad90541ba52dfa02bdd83466f17e6541e535
Stack from [ghstack](https://github.com/ezyang/ghstack):
* #18833 [STACK] Cache device on TensorImpl; clean up TensorImpl constructors.
* **#18832 [STACK] Disallow changing the device of a tensor via set_.**
* #18831 [STACK] Stop swapping in Storages of the wrong device for Tensors.
This is necessary to cache the device on a TensorImpl.
Differential Revision: D14766231
fbshipit-source-id: bba61634b2d6252ac0697b96033c9eea680956e8
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/18831
ghimport-source-id: 2741e0d70ebe2c2217572c3af54ddd9d2047e342
Stack from [ghstack](https://github.com/ezyang/ghstack):
* #18833 [STACK] Cache device on TensorImpl; clean up TensorImpl constructors.
* #18832 [STACK] Disallow changing the device of a tensor via set_.
* **#18831 [STACK] Stop swapping in Storages of the wrong device for Tensors.**
This is necessary to support device caching, see https://github.com/pytorch/pytorch/pull/18751 and https://github.com/pytorch/pytorch/pull/18578.
In library code, we potentially swap in Storages with the wrong device when device_guard is False. This happens as follows with "view-like" operations.
1) We allocate a tensor on the 'wrong' device (because device_guard is false).
2) We swap out the 'wrong' storage with the 'right' storage using e.g. THCTensor_setStorage.
Instead, we can just construct the Tensor with the correct Storage from the beginning. This is what we do with 'view'.
Note there are two other "view-like" cases where this happens:
1) unfold
2) set_()
Because these aren't performance critical, I just added the device_guard instead of applying the above correction.
For completeness, this also includes a test that all `device_guard: false` functions behave properly under these conditions.
Reviewed By: dzhulgakov
Differential Revision: D14766232
fbshipit-source-id: 0865c3ddae3f415df5da7a9869b1ea9f210e81bc
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/18648
ghimport-source-id: 1cf4a8fe91492621e02217f38cae5d7e0699fb05
Stack from [ghstack](https://github.com/ezyang/ghstack):
* #18661 Step 7: remove _unique
* #18655 Step 6: Rename _unique2 to unique and add int? dim
* #18654 Step 5: remove _unque_dim in favor of unique_dim
* #18651 Step 4: add support for unique with dim=None
* #18650 Step 3: Add support for return_counts to torch.unique for dim not None
* #18649 Step 2: Rename _unique_dim2_temporary_will_remove_soon to unique_dim
* **#18648 Step 1: Secretly add return_counts to unique, and refactor unique_dim for performance**
`unique` is fragile, previously I tried to change it in #18391 and #17097, they all pass OSS tests but finally get reverted due to internal failure. My previous work of refactoring unique #18459 is based on #18391, and after #18391 get reverted, I could not work on #18459. To continue working on #18459, #18391, and #17097 without worrying about internal failures, I am suggesting the following steps for the improvements of `unique` and `unique_dim`. soumith Please take this and there is no need to put #18391 back.
The motivation is basically to move forward as much as possible without causing any internal failures. So I will try to divide it into steps and sort from low probability of internal failure to high probability. (I don't know what the internal failure is, so I have to guess). Let's merge these PR stack one by one until we enounter internal failure.
Step 1: Create two new ATen operators, `_unique2_temporary_will_remove_soon` and `_unique_dim2_temporary_will_remove_soon` and keep `_unique` and `_unique_dim` unchanged. The backend of these two functions and `_unique` and `_unique_dim` are all the same, the only difference is the temporary ones support `return_counts` but not the `_unique` and `_unique_dim`. Step one is mostly #18391 + #18459. The cuda8 errors has been fixed. At this point, there is no user visible API change, so no docs are updated. `torch.unique` does not support `return_counts` yet, and `return_counts` is tested through the newly added temporary operators. This step just added two new ATen operators, so there shouldn't be any internal failure.
Step 2: Rename `_unique_dim2_temporary_will_remove_soon` to `unique_dim`. This should cause no internal failure either, because no change to existing operators. The only thing to worry about is to delete `unique_dim` from python side because we don't want users to use it. At this point, C++ users now have `return_counts` support for `unique_dim`.
Step 3: Update the docs of `torch.unique` and use `unique_dim` inside `torch.unique` to support `return_counts` In the docs, we should say `torch.unique` with None dim support does not support `return_counts` yet. This might cause internal failure.
Step 4: Rename `_unique2_temporary_will_remove_soon` to `_unique2` and use `_unique2` inside `torch.unique` to support `return_counts`. Update the docs saying that `torch.unique` with None dim now support `return_counts`. This might cause internal failure.
Step 5: Remove `_unique_dim`. This might cause internal failure.
Step 6: Rename `_unique2` to `unique`, add optional `dim` argument to make it looks like the signature of Python's `torch.unique`. Inside `torch.unique`, use `unique` and get rid of `unique_dim`. Unbind `unique_dim` totally from Python at codegen. This is likely to cause internal fail.
Step 7: Remove `_unique`. This is very likely to cause internal failure.
This PR
======
This PR is for step 1. This create two new ATen operators, `_unique2_temporary_will_remove_soon` and `_unique_dim2_temporary_will_remove_soon` and implement `return_counts` inside them and do refactor for performance improvements.
Please review ngimel VitalyFedyunin. They are mostly copied from #18391 and #18459, so the review should be easy.
Below is a benchmark on a tensor of shape `torch.Size([15320, 2])`:
Before
---------
```python
print(torch.__version__)
%timeit a.unique(dim=0, sorted=True, return_inverse=False); torch.cuda.synchronize()
%timeit a.unique(dim=0, sorted=True, return_inverse=True); torch.cuda.synchronize()
```
```
1.0.1
192 µs ± 1.61 µs per loop (mean ± std. dev. of 7 runs, 1000 loops each)
548 ms ± 3.39 ms per loop (mean ± std. dev. of 7 runs, 1 loop each)
```
```python
print(torch.__version__)
%timeit a.unique(sorted=True, return_inverse=False); torch.cuda.synchronize()
%timeit a.unique(sorted=True, return_inverse=True); torch.cuda.synchronize()
```
```
1.0.1
226 µs ± 929 ns per loop (mean ± std. dev. of 7 runs, 1000 loops each)
302 µs ± 7.06 µs per loop (mean ± std. dev. of 7 runs, 1000 loops each)
```
After
-------
```python
print(torch.__version__)
%timeit a.unique(dim=0, sorted=True, return_inverse=False); torch.cuda.synchronize()
%timeit a.unique(dim=0, sorted=True, return_inverse=True); torch.cuda.synchronize()
%timeit torch._unique_dim2_temporary_will_remove_soon(a, dim=0, sorted=True, return_inverse=False, return_counts=True); torch.cuda.synchronize()
%timeit torch._unique_dim2_temporary_will_remove_soon(a, dim=0, sorted=True, return_inverse=True, return_counts=True); torch.cuda.synchronize()
```
```
1.1.0a0+83ab8ac
190 µs ± 2.14 µs per loop (mean ± std. dev. of 7 runs, 1000 loops each)
237 µs ± 1.23 µs per loop (mean ± std. dev. of 7 runs, 1000 loops each)
219 µs ± 2.3 µs per loop (mean ± std. dev. of 7 runs, 1000 loops each)
263 µs ± 1.15 µs per loop (mean ± std. dev. of 7 runs, 1000 loops each)
```
```python
print(torch.__version__)
%timeit a.unique(sorted=True, return_inverse=False); torch.cuda.synchronize()
%timeit a.unique(sorted=True, return_inverse=True); torch.cuda.synchronize()
%timeit torch._unique2_temporary_will_remove_soon(a, sorted=True, return_inverse=False, return_counts=True); torch.cuda.synchronize()
%timeit torch._unique2_temporary_will_remove_soon(a, sorted=True, return_inverse=True, return_counts=True); torch.cuda.synchronize()
```
```
1.1.0a0+83ab8ac
232 µs ± 2.21 µs per loop (mean ± std. dev. of 7 runs, 1000 loops each)
301 µs ± 1.65 µs per loop (mean ± std. dev. of 7 runs, 1000 loops each)
264 µs ± 7.67 µs per loop (mean ± std. dev. of 7 runs, 1000 loops each)
339 µs ± 9.2 µs per loop (mean ± std. dev. of 7 runs, 1000 loops each)
```
Differential Revision: D14730905
fbshipit-source-id: 10026b4b98628a8565cc28a13317d29adf1225cc
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/18230
Implementing minimum qtensor API to unblock other workstreams in quantization
Changes:
- Added Quantizer which represents different quantization schemes
- Added qint8 as a data type for QTensor
- Added a new ScalarType QInt8
- Added QTensorImpl for QTensor
- Added following user facing APIs
- quantize_linear(scale, zero_point)
- dequantize()
- q_scale()
- q_zero_point()
Reviewed By: dzhulgakov
Differential Revision: D14524641
fbshipit-source-id: c1c0ae0978fb500d47cdb23fb15b747773429e6c
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/18749
ghimport-source-id: 9026a037f5e11cdb9ccd386f4b6b5768b9c3259b
Stack from [ghstack](https://github.com/ezyang/ghstack):
* #18751 Disallow changing the device of a tensor via set_.
* #18750 Use non-legacy constructors for tensor deserialization.
* **#18749 Add device and dtype to storage.**
The goal here is to fix our serialization, which currently depends on the legacy constructors. Having dtype and device on Storage allows us to use the non-legacy constructors.
This fits somewhat along our goal of removing Storage, my having Storage act like a Tensor.
Differential Revision: D14729516
fbshipit-source-id: bf4a3e8669ad4859931f4a3fa56df605cbc08dcb
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/18166
ghimport-source-id: a8e2ba2d966e49747a55701c4f6863c5e24d6f14
Stack from [ghstack](https://github.com/ezyang/ghstack):
* **#18166 Bool Tensor for CUDA**
* #18165 Resolved comments from Bool Tensor for CPU PR
------
This PR enables bool tensor creation and some basic operations for the CPU backend. This is a part of Bool Tensor feature implementation work. The whole plan looks like this:
1. Storage Implementation [Done]
2. Tensor Creation.
a) CPU [Done]
b) CUDA [This PR]
3. Tensor Conversions.
4. Tensor Indexing.
5. Tensor Operations.
6. Back compatibility related changes.
Change:
Enable bool tensor in CUDA with the following operations:
torch.zeros
torch.tensor
torch.ones
torch.rand/rand_like/randint/randint_like
torch.full
torch.full_like
torch.empty
torch.empty_like
Tested via unit tests and local scripts.
Differential Revision: D14605104
fbshipit-source-id: b7d7340a7d70edd03a109222d271e68becba762c
Summary:
Argument dim=-1 doesn't work for torch.cross. The signature of the torch.cross has been changed to c10::optional<int64_t> dim instead of int64_t. So based on document "If dim is not given, it defaults to the first dimension found with the size 3." and if dim is specified (even negative) it will use the correspondent dim.
Fixes#17229
Pull Request resolved: https://github.com/pytorch/pytorch/pull/17582
Differential Revision: D14483063
Pulled By: ifedan
fbshipit-source-id: f9699093ec401cb185fd33ca4563c8a46cdcd746
Summary:
Make it possible to construct a pinned memory tensor without creating a storage first and without calling pin_memory() function. It is also faster, as copy operation is unnecessary.
Supported functions:
```python
torch.rand_like(t, pin_memory=True)
torch.randn_like(t, pin_memory=True)
torch.empty_like(t, pin_memory=True)
torch.full_like(t, 4, pin_memory=True)
torch.zeros_like(t, pin_memory=True)
torch.ones_like(t, pin_memory=True)
torch.tensor([10,11], pin_memory=True)
torch.randn(3, 5, pin_memory=True)
torch.rand(3, pin_memory=True)
torch.zeros(3, pin_memory=True)
torch.randperm(3, pin_memory=True)
torch.empty(6, pin_memory=True)
torch.ones(6, pin_memory=True)
torch.eye(6, pin_memory=True)
torch.arange(3, 5, pin_memory=True)
```
Part of the bigger: `Remove Storage` plan.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/18455
Reviewed By: ezyang
Differential Revision: D14672084
Pulled By: VitalyFedyunin
fbshipit-source-id: 9d0997ec00f59500ee018f8b851934d334012124
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/18598
ghimport-source-id: c74597e5e7437e94a43c163cee0639b20d0d0c6a
Stack from [ghstack](https://github.com/ezyang/ghstack):
* **#18598 Turn on F401: Unused import warning.**
This was requested by someone at Facebook; this lint is turned
on for Facebook by default. "Sure, why not."
I had to noqa a number of imports in __init__. Hypothetically
we're supposed to use __all__ in this case, but I was too lazy
to fix it. Left for future work.
Be careful! flake8-2 and flake8-3 behave differently with
respect to import resolution for # type: comments. flake8-3 will
report an import unused; flake8-2 will not. For now, I just
noqa'd all these sites.
All the changes were done by hand.
Signed-off-by: Edward Z. Yang <ezyang@fb.com>
Differential Revision: D14687478
fbshipit-source-id: 30d532381e914091aadfa0d2a5a89404819663e3
Summary:
Changelog:
- Renames `btriunpack` to `lu_unpack` to remain consistent with the `lu` function interface.
- Rename all relevant tests, fix callsites
- Create a tentative alias for `lu_unpack` under the name `btriunpack` and add a deprecation warning to not promote usage.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/18529
Differential Revision: D14683161
Pulled By: soumith
fbshipit-source-id: 994287eaa15c50fd74c2f1c7646edfc61e8099b1
Summary:
Changelog:
- Renames `btrifact` and `btrifact_with_info` to `lu`to remain consistent with other factorization methods (`qr` and `svd`).
- Now, we will only have one function and methods named `lu`, which performs `lu` decomposition. This function takes a get_infos kwarg, which when set to True includes a infos tensor in the tuple.
- Rename all tests, fix callsites
- Create a tentative alias for `lu` under the name `btrifact` and `btrifact_with_info`, and add a deprecation warning to not promote usage.
- Add the single batch version for `lu` so that users don't have to unsqueeze and squeeze for a single square matrix (see changes in determinant computation in `LinearAlgebra.cpp`)
Pull Request resolved: https://github.com/pytorch/pytorch/pull/18435
Differential Revision: D14680352
Pulled By: soumith
fbshipit-source-id: af58dfc11fa53d9e8e0318c720beaf5502978cd8
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/18507
ghimport-source-id: 1c3642befad2da78a7e5f39d6d58732b85c76267
Stack from [ghstack](https://github.com/ezyang/ghstack):
* **#18507 Upgrade flake8-bugbear to master, fix the new lints.**
It turns out Facebobok is internally using the unreleased master
flake8-bugbear, so upgrading it grabs a few more lints that Phabricator
was complaining about but we didn't get in open source.
A few of the getattr sites that I fixed look very suspicious (they're
written as if Python were a lazy language), but I didn't look more
closely into the matter.
Signed-off-by: Edward Z. Yang <ezyang@fb.com>
Differential Revision: D14633682
fbshipit-source-id: fc3f97c87dca40bbda943a1d1061953490dbacf8
Summary:
This depend on https://github.com/pytorch/pytorch/pull/16039
This prevent people (reviewer, PR author) from forgetting adding things to `tensors.rst`.
When something new is added to `_tensor_doc.py` or `tensor.py` but intentionally not in `tensors.rst`, people should manually whitelist it in `test_docs_coverage.py`.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/16057
Differential Revision: D14619550
Pulled By: ezyang
fbshipit-source-id: e1c6dd6761142e2e48ec499e118df399e3949fcc
Summary:
More ops for https://github.com/pytorch/pytorch/issues/394. ~~Also need to rebase after landing #16186, because we need to update the whitelist of the new unit test added in #16186.~~
cc: ezyang
Pull Request resolved: https://github.com/pytorch/pytorch/pull/17093
Differential Revision: D14620068
Pulled By: ezyang
fbshipit-source-id: deec5ffc9bf7624e0350c85392ee59789bad4237
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/18165
ghimport-source-id: 55cb3fb63a25c2faab1725b4ec14c688bf45bd38
Stack from [ghstack](https://github.com/ezyang/ghstack):
* #18166 Bool Tensor for CUDA
* **#18165 Resolved comments from Bool Tensor for CPU PR**
-------
------------
This is a follow up PR that resolves some additional feedback on one the of previous Bool Tensor PRs.
gchanan, here is a list of almost all the comments from the original PR with respective fixes and replies:
**[utils/python_scalars.h]** why is this converting from uint8_t and not bool? (comment?)
When i was adding this, i was testing by creating a tensor and then calling its .tolist(). it worked for bool and uint8_t equally good so i left uint8_t as thought it makes more sense as we are calling PyBool_FromLong. �Changing it to bool.
**[ATen/Dispatch.h]**better name?.
fixed.
**[test/test_torch.py]** what about other factories, such as full? (and more).
There is a test that goes through the factory methods - test_tensor_factories_empty. i added some bool cases above it and added a comment that once CUDA will be done, i will unite them and it will iterate not just between CUDA and CPU but also all types. ��Adding all bool cases now. Will unite in CUDA PR.
**[generic/THTensorMath.h]** any changes in this file actually needed?
Bad merge. Fixed.
**[TH/THTensor.h]** this generates code for random, clampedRandom, and cappedRandom -- do we have tests for all of these with bool?
Added
**[c10/core/ScalarType.h]** I'm not very confident about the lack of Bool here -- can you look at the call sites and see what makes sense to do here?
Added bool to the macro and created a similar one without for a single case which fails the build with errors:
_./torch/csrc/jit/symbolic_variable.h:79:20: error: ambiguous overload for ‘operator*’ (operand types are ‘const torch::jit::SymbolicVariable’ and ‘torch::jit::Value*’)
return (*this) * insertConstant(rhs);_
Differential Revision: D14605105
fbshipit-source-id: abf82d50e8f8c50b386545ac068268651b28496d
