Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/55102
To avoid casting a tensor to `.long()`, we introduce support for int32 in `torch.repeat_interleave`.
Reviewed By: ezyang
Differential Revision: D27478235
fbshipit-source-id: 08b4cce65fe94ff10535ddc07e1ba2bacea6a2cf
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/53238
There is a tension for the Vitals design: (1) we want a macro based logging API for C++ and (2) we want a clean python API. Furthermore, we want to this to work with "print on destruction" semantics.
The unfortunate resolution is that there are (2) ways to define vitals:
(1) Use the macros for local use only within C++ - this keeps the semantics people enjoy
(2) For vitals to be used through either C++ or Python, we use a global VitalsAPI object.
Both these go to the same place for the user: printing to stdout as the globals are destructed.
The long history on this diff shows many different ways to try to avoid having 2 different paths... we tried weak pointers & shared pointers, verbose switch cases, etc. Ultimately each ran into an ugly trade-off and this cuts the difference better the alternatives.
Test Plan:
buck test mode/dev caffe2/test:torch -- --regex vital
buck test //caffe2/aten:vitals
Reviewed By: orionr
Differential Revision: D26736443
fbshipit-source-id: ccab464224913edd07c1e8532093f673cdcb789f
Summary:
#### Reason for relanding
Line 1607 of `torch/testing/_internal/common_methods_invocations.py` of https://github.com/pytorch/pytorch/issues/50999 had `dtype` instead of `dtype=torch.bool`, so 4 of the 9 sample inputs for `bool` had incorrect dtype. This bug was caught by https://github.com/pytorch/pytorch/issues/54949.
1. Added support for pow() on CPU for `float16` (`Half`) and `bfloat16` types.
Both `pow(Tensor, Scalar)` and `pow(Tensor, Tensor)` are now supported for the aforementioned types.
However autograd isn't supported for `Float16` on CPU yet, as `log_vml_cpu` can't be enabled for it.
2. heitorschueroff added `pow_tensor_scalar_optimized_kernel` to refactor & simplify `PowKernel.cpp`.
It provides a common path for all the complex types & floating point types (except Float16, due to lack of complete AVX2 vectorization support for it). It replaced code that had previously been duplicated for (float, double) and complex types,
so PowKernel.cpp looks a lot cleaner now.
3. Enabled (unskipped) some tests for `erf`, `erfc`,`erfinv`, `tan` and `linalg.vector.norm` which were being skipped earlier due to `pow()` not having been implemented for `float16` & `bfloat16`.
4. Added an OpInfo for `pow()` & enabled some test cases for `pow()`.
5. Extended the coverage of existing tests for `pow` in `test_binary_ufuncs.py` in order to enable comparison with `numpy`, even with discontiguous tensors, and added a test to ensure that a runtime error is raised for `pow`'s inplace variant if resizing the base tensor is required during its invocation.
6. Added `float16` & `bfloat16` to `square`'s dtype lists in its `UnaryUfuncInfo`.
7. Removed redundant `dtypesIfCPU` and `dtypesIfCUDA` from `OpInfo`s where they are equal to `dtypes`.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/55280
Reviewed By: jbschlosser
Differential Revision: D27591772
Pulled By: heitorschueroff
fbshipit-source-id: c7420811b32595bb3353149a61e54a73f2eb352b
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/55150
Somehow I forgot to add these checks. Now they're in here. Thanks
ngimel for noticing.
This is probably a slight efficiency hit on TensorIterator, which is
probably already doing all these checks. Would be good to follow up
on this, though it may not be easily fixable with the TI rewrite.
Signed-off-by: Edward Z. Yang <ezyang@fb.com>
Test Plan: Imported from OSS
Reviewed By: zhangguanheng66
Differential Revision: D27523879
Pulled By: ezyang
fbshipit-source-id: 458e617dbc6de6fcfa9e5841148b30b99f52e001
Summary:
Added the functionality desired in https://github.com/pytorch/pytorch/issues/50789.
1. Added support for pow() on CPU for `float16` (`Half`) and `bfloat16` types.
Both `pow(Tensor, Scalar)` and `pow(Tensor, Tensor)` are now supported for the aforementioned types.
However autograd isn't supported for `Float16` on CPU yet, as `log_vml_cpu` can't be enabled for it.
2. heitorschueroff added `pow_tensor_scalar_optimized_kernel` to refactor & simplify `PowKernel.cpp`.
It provides a common path for all the complex types & floating point types (except Float16, due to lack of complete AVX2 vectorization support for it). It replaced code that had previously been duplicated for (float, double) and complex types,
so PowKernel.cpp looks a lot cleaner now.
3. Enabled (unskipped) some tests for `erf`, `erfc`,`erfinv`, `linalg.norm` and `linalg.vector.norm` which were being skipped earlier due to `pow()` not having been implemented for `float16` & `bfloat16`.
4. Added an OpInfo for `pow()` & enabled some test cases for `pow()`.
5. Extended the coverage of existing tests for `pow` in `test_binary_ufuncs.py` in order to enable comparison with `numpy`, even with discontiguous tensors, and added a test to ensure that a runtime error is raised for `pow`'s inplace variant if resizing the base tensor is required during its invocation.
6. Added `float16` & `bfloat16` to `square`'s dtype lists in its `UnaryUfuncInfo`.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/50999
Reviewed By: zou3519
Differential Revision: D27478225
Pulled By: heitorschueroff
fbshipit-source-id: d309dd98d5a96d0cb9b08281757bb1c65266d011
Summary:
- Corrected a few errata in the SVD docs
- Made the notation more uniform (refer to `Vh` in `linalg.svd`, always use double tilts...)
- Wrote a better explanation about why the gradients of `U` and `V` are not well-defined when the input is complex or real but has repeated singular values. The previous one pointed to a somewhat obscure post on gauge theory.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/54002
Reviewed By: malfet
Differential Revision: D27459502
Pulled By: mruberry
fbshipit-source-id: f5c35eca02d35dadd2fc0eeadfacc8824f409400
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/54901
Some subtleties:
- Need to make sure not to clobber composite definitions when
deciding when to generate
- I was lazy and so I didn't make inplace on TensorList work,
nor did I make inplace functions that returned void work
- A few tests started complaining that these noop meta functions
weren't raising the errors they needed. This is tracked
in https://github.com/pytorch/pytorch/issues/54897
Signed-off-by: Edward Z. Yang <ezyang@fb.com>
Test Plan: Imported from OSS
Reviewed By: jbschlosser
Differential Revision: D27407232
Pulled By: ezyang
fbshipit-source-id: 5e706a267496368acdafd128942c310954e43d29
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/53973
Two parts to this PR; I had to put them together because adding support for X causes more test code to be exercised, which in turn may require a fix for Y.
The first part is restoring the concept of storage to meta tensors. Previously, meta tensors had a nullptr storage (e.g., `meta_tensor.storage()` is an error.) As I was increasing the coverage of meta tensors, I started running into test cases (specifically memory overlap tests) that were failing because not having storage meant I couldn't check for memory overlap. After some discussion, we decided that it would make sense for meta tensors to model this as well (we already model strides, so getting accurate view information also seems useful). This PR does that by:
* Rewrite all of the factory functions in MetaTensor.cpp to use the generic versions (which are very carefully written to not actually poke at the data pointer, so everything works out). The key idea here is we give meta tensors a special allocator, MetaAllocator, which always returns a nullptr even if you ask for a nonzero number of bytes. resize_ is also made generic; the normal variant can be used directly rather than having to instruct it to avoid resizing storage
* Turn on memory overlap checking in TensorIterator even for meta tensors
* Although meta tensors now have storage, the concept of meta storage is NOT exposed to Python land (as it would imply I would have to codegen MetaFloatStorage, MetaDoubleStorage, etc. classes). So `x.storage()` still raises an error and I have a cludge in `__deepcopy__` to break storage sharing upon deep copy (this is wrong, but no tests exercise this at the moment).
The second part is adding more support for the most used functions in the test suite.
* Inplace operations have very simple meta functions. I added `fill_`, `zero_`, `random_`, `uniform_` and `normal_`. In the case of random, I take advantage of pbelevich's templates for defining random kernels, so that I can reuse the common scaffolding, and then just register a noop stub that actually does the RNG. (Look, another structured kernels tiny variant!)
* `copy_` is now implemented. Copying into a meta tensor is always OK, but copying out of a meta tensor raises an error (as we don't know what the "correct" data to copy out is in this case)
* `empty_strided` usage from structured kernels now is implemented (TBH, this could have been done as soon as `empty_strided` was added)
* Meta was missing in a few places in TensorOptions/DispatchKey utility functions, so I added them
* Autograd engine now correctly homes meta tensors with CPU tensors (they have -1 device index so CUDA queues wouldn't work anyway)
* `apply_`, `map_` and `map2_` are special cased to no-op on meta tensor self. These count as inplace operations too but they are implemented a little differently.
Getting more meta function support triggers a number of bugs in the test suite, which I then fix:
- Linear algebra functions sometimes don't report NotImplementedError because they get swallowed by catch all try blocks. This is tracked in https://github.com/pytorch/pytorch/issues/53739
- dlpack obviously doesn't work with meta tensors, I just disabled the test
Signed-off-by: Edward Z. Yang <ezyang@fb.com>
Differential Revision: D27036572
Test Plan: Imported from OSS
Reviewed By: agolynski, bdhirsh
Pulled By: ezyang
fbshipit-source-id: 7005ecf4feb92a643c37389fdfbd852dbf00ac78
Summary:
```
index_add(Tensor self, int dim, Tensor index, Tensor source) -> Tensor
```
now becomes
```
index_add(Tensor self, int dim, Tensor index, Tensor source, Scalar alpha=1) -> Tensor
```
Generally, this sounds useful and harmless, and inside PyTorch, we are already needing this feature in `add_out_dense_sparse_cuda`, see the `SparseCUDATensorMath.cu` change in this PR.
**Test not added yet. Will add if after discussion we believe this is a good idea.**
- [ ] TODO: add test
Pull Request resolved: https://github.com/pytorch/pytorch/pull/54176
Reviewed By: ngimel
Differential Revision: D27319198
Pulled By: mruberry
fbshipit-source-id: fe43be082d1230c87c5313458213d5252be2ff23
Summary:
Added the support for half / bfloat / bool for `index_select`, as suggested by ngimel in
https://github.com/pytorch/pytorch/issues/49707#issuecomment-788140578
For the tests to pass, I also added the support for `index_add`.
I added `OpInfo` tests for `index_add` and more thorough forward tests for `index_select` to test these changes.
While doing so, I found that the support for scalar types in the derivative of `index_add` was not correct, so I corrected it.
Resolves https://github.com/pytorch/pytorch/issues/49707
It should also resolve similar issues that I encountered when porting `index_copy`, `take` and `put`.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/53898
Reviewed By: mruberry
Differential Revision: D27193294
Pulled By: ngimel
fbshipit-source-id: 5a0af2c62a0cf24f3cc9c74f230ab4f3712bbb7a
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/54079
Fixes https://github.com/pytorch/pytorch/issues/53815
Instead of testing if something is CUDA, we instead test if something
is not CPU. This in the general theming of "Don't be so darn CUDA
centric".
Intruigingly, we didn't have a is_cpu() method on Tensor. Which seems
like a big oversight and one of the reasons how we ended up in this
mess. So in it goes. Maybe we should also get this for Python bindings
as well (but in that case, should probably look into redoing all of the
is_X bindings so they aren't done manually).
Signed-off-by: Edward Z. Yang <ezyang@fb.com>
Test Plan: Imported from OSS
Reviewed By: ngimel
Differential Revision: D27109507
Pulled By: ezyang
fbshipit-source-id: abbe72c2e688c452ffe098d206cb79938b5824b1
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/53759Fixes#53587, see issue for in-depth explanation of the bug.
Signed-off-by: Edward Z. Yang <ezyang@fb.com>
Test Plan: Imported from OSS
Reviewed By: albanD
Differential Revision: D26971342
Pulled By: ezyang
fbshipit-source-id: 805983fed2658e27fb033f36a71fd30950a29328
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/53665
ngimel pointed out to me where we already test the behavior of the `Upsample` ops in `test_nn.py`. This PR deleting my bespoke tests in `test_torch.py` and updates those in `test_nn.py` to test memory format properly.
There were two reasons the original test didn't pick up on a memory format regression:
- They didn't test the memory format of the output tensor explicitly, i.e. `output.is_contiguous(memory_format=...)`
- Even with that change, the test tensors were to simple to fail the tests. From some trial and error, it looks like one of the first two dimensions in the inputs needs to be > 1 in order for the `channels_last` memory format to actually re-order the strides.
Test Plan: Imported from OSS
Reviewed By: ngimel
Differential Revision: D26929683
Pulled By: bdhirsh
fbshipit-source-id: d17bc660ff031e9b3e2c93c60a9e9308e56ea612
Summary:
Fixes https://github.com/pytorch/pytorch/issues/48841 for half datatype (it was fixed for other datatypes before).
The reason for https://github.com/pytorch/pytorch/issues/48841 happening for half was that `exponential_` for half was producing 0s.
Exponential distribution implementation on cuda is here e08aae2613/aten/src/ATen/native/cuda/DistributionTemplates.h (L535-L545)
with `transformation::exponential` defined here
e08aae2613/aten/src/ATen/core/TransformationHelper.h (L113-L123)
It takes a uniformly distributed random number and takes `log` of it. If necessary, the result is then converted to low precision datatype (half). To avoid 0's, before applying `log`, ones are replaced with std::nextafter(1,0). This seems fine, because log(1-eps) is still representable in half precision (`torch.tensor([1.], device="cuda").nextafter(torch.tensor([0.], device="cuda")).log().half()` produces 5.96e-8) , so casting to `scalar_t` should work. However, since fast log approximation is used (`__logf`), the log result is ~3e-9 instead of more accurate 5.96e-8, and underflows when casting to half. Using `::log` instead of fast approximation fixes it, however, it comes with ~20% perf penalty on exponential kernel for fp32 datatype, probably more for half.
Edit: alternative approach used now is to filter all small values returned by transformation. The result is equivalent to squashing of 1's to 1-eps that was used before, and computing correct log of 1-eps (which is -eps, exactly equal even for doubles). This doesn't incur noticeable performance hit.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/53480
Reviewed By: mruberry
Differential Revision: D26924622
Pulled By: ngimel
fbshipit-source-id: dc1329e4773bf91f26af23c8afa0ae845cfb0937
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/53535
During the port to structured kernels for upsample kernels, I missed that a subset of them explicitly pass `memory_format` information from the input to the output tensors.
Note 1:
I added the logic into the `meta` function of each op, which feels morally correct since this logic affects the output shape/metadata. One consequence is that all backend implementations will get the logic. I synced with fmassa that this seems reasonable.
Note 2:
This logic used to happen in the following operators, which this PR fixes:
- upsample_nearest3d
- upsample_trilinear3d
- upsample_nearest2d
- upsample_bilinear2d
I explicitly didn't patch the other upsample kernels, which look like they never forwarded memory_format information:
- `upsample_bicubic2d` (maybe this should though? `UpSampleBicubic2d.cpp` isn't currently written to do anything different for `channels_last` tensors)
- All of the `upsample_{mode}1d` operators. Probably because, afaik, channels_last isn't supported for 3d tensors
- The corresponding backwards operator for every upsample op.
Note 3:
I'm also wondering why memory_format isn't just directly a part of the `tensor::options()` method, which would cause all ops to universally forward memory_format information from input to output tensors, rather than just the upsample ops. My guess is:
- BC-breakage. I'm not sure whether this would really *break* people, but it's an API change
- performance. `tensor::options()` is called everywhere, and adding a call to `suggest_memory_format()` would probably noticeably hit microbenchmarks. We could probably deal with that by making `memory_format` a precomputed field on the tensor?
Test Plan: Imported from OSS
Reviewed By: H-Huang
Differential Revision: D26891540
Pulled By: bdhirsh
fbshipit-source-id: b3845f4dd5646b88bf738b9e41fe829be6b0e5cf
Summary:
Helps make master green by removing this hefty memory allocating from CPU test.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/53561
Reviewed By: malfet, albanD
Differential Revision: D26897941
Pulled By: janeyx99
fbshipit-source-id: 9f6c2d55f4eea1ab48665f7819fc113f21991036
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/53276
- One of the tests had a syntax error (but the test
wasn't fine grained enough to catch this; any error
was a pass)
- Doesn't work on ROCm
Signed-off-by: Edward Z. Yang <ezyang@fb.com>
Differential Revision: D26820048
Test Plan: Imported from OSS
Reviewed By: mruberry
Pulled By: ezyang
fbshipit-source-id: b02c4252d10191c3b1b78f141d008084dc860c45
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/53143
Meta is now an honest to goodness device type, like cpu, so you can use
device='meta' to trigger allocation of meta tensors. This way better
than empty_meta since we now have working API for most factory functions
(they don't necessarily work yet, though, because need to register Meta
versions of those functions.)
Some subtleties:
- I decided to drop the concept of CPU versus CUDA meta tensors; meta
tensors are device agnostic. It's hard to say exactly what the
correct level of abstraction here is, but in this particular case
implementation considerations trump semantic considerations: it
is way easier to have just a meta device, than to have a meta device
AND a cpu device AND a cuda device. This may limit the applicability
of meta tensors for tracing models that do explicit cpu()/cuda()
conversions (unless, perhaps, we make those operations no-ops on meta
tensors).
- I noticed that the DeviceType uppercase strings are kind of weird.
Are they really supposed to be all caps? That's weird.
- I moved the Meta dispatch key to live with the rest of the "device"
dispatch keys.
- I intentionally did NOT add a Backend for Meta. For now, I'm going to
hope meta tensors never exercise any of the Backend conversion code;
even if it does, better to fix the code to just stop converting to and
from Backend.
Signed-off-by: Edward Z. Yang <ezyang@fb.com>
Test Plan: Imported from OSS
Reviewed By: samestep
Differential Revision: D26763552
Pulled By: ezyang
fbshipit-source-id: 14633b6ca738e60b921db66a763155d01795480d
Summary:
Fixes https://github.com/pytorch/pytorch/issues/52213
Nans were previously inconsistently propagated due to std::min always returning first argument if one of the args in nan
when reduction functor was called on 2 `-inf` arguments, `std::min(x,y) - std::max(x,y)` resulted in `-inf - (-inf)` = nan, even though logcumsumexp is well defined for `-inf, -inf` pair.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/52947
Reviewed By: H-Huang
Differential Revision: D26718456
Pulled By: ngimel
fbshipit-source-id: a44433889da352cc959786dd15b6361a68fcfed7
Summary:
This PR adds functionality to skip a test based on CUDA version.
This way, we can be more specific when skipping a test, such as when the test only fails for a particular CUDA version.
This allows us to add back the skipped tests for CUDA 11.2 for other CUDA versions, such as 10.1 and 11.1.
I tested this locally (by using 11.0 instead of 11.2), but will run all the CI to make sure it works.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/52359
Reviewed By: walterddr
Differential Revision: D26487951
Pulled By: janeyx99
fbshipit-source-id: 45c71cc6105ffd9985054880009cf68ea5ef3f6a
Summary:
Fixes https://github.com/pytorch/pytorch/issues/51719, https://github.com/pytorch/pytorch/issues/28142
**Change**
- Update `torch.Tensor.unflatten` to support users pass`-1` as the inferred size for both tensors and named tensors.
- Examples of using `-1` in the `unflatten` function are added to the docs.
- Fix the rendered issue of original `unflatten` docs by removing a blank line between its example section.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/51955
Reviewed By: agolynski
Differential Revision: D26467198
Pulled By: zou3519
fbshipit-source-id: 6a3ede25561223187273796427ad0cb63f125364
Summary:
Reference: https://github.com/pytorch/pytorch/issues/50006
We should probably add aliases for these operators to be consistent with NumPy names i.e. `np.degrees` and `np.radians`.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/51283
Reviewed By: ngimel
Differential Revision: D26171163
Pulled By: mruberry
fbshipit-source-id: 1869604ed400820d95f6ff50a0e3cba1de1ffa84
Summary:
Adding CUDA 11.2 to Windows CI.
Disabled tests:
The following ran into `CUDA error: misaligned address` for CUDA 11.2: (issue linked below)
`test_where_scalar_valid_combination_cuda_complex128` in test_torch.py
`test_sgn_complex_cuda` in test_autograd.py
The following ran into `CUDA error: too many resources requested for launch` for CUDA 11.2: (https://github.com/pytorch/pytorch/issues/52002)
test_EmbeddingBag_per_sample_weights_and_new_offsets_cuda_int64_float64
test_EmbeddingBag_per_sample_weights_and_offsets_cuda_int64_float64
Pull Request resolved: https://github.com/pytorch/pytorch/pull/51598
Reviewed By: mrshenli
Differential Revision: D26344965
Pulled By: janeyx99
fbshipit-source-id: 3c9a4ed16d748969e96593220ec0a9f33e1ffcef
Summary:
Toward fixing https://github.com/pytorch/pytorch/issues/47624
~Step 1: add `TORCH_WARN_MAYBE` which can either warn once or every time in c++, and add a c++ function to toggle the value.
Step 2 will be to expose this to python for tests. Should I continue in this PR or should we take a different approach: add the python level exposure without changing any c++ code and then over a series of PRs change each call site to use the new macro and change the tests to make sure it is being checked?~
Step 1: add a python and c++ toggle to convert TORCH_WARN_ONCE into TORCH_WARN so the warnings can be caught in tests
Step 2: add a python-level decorator to use this toggle in tests
Step 3: (in future PRs): use the decorator to catch the warnings instead of `maybeWarnsRegex`
Pull Request resolved: https://github.com/pytorch/pytorch/pull/48560
Reviewed By: ngimel
Differential Revision: D26171175
Pulled By: mruberry
fbshipit-source-id: d83c18f131d282474a24c50f70a6eee82687158f
Summary:
Implements `np.diff` for single order differences only:
- method and function variants for `diff` and function variant for `diff_out`
- supports out variant, but not in-place since shape changes
- adds OpInfo entry, and test in `test_torch`
- automatic autograd because we are using the `Math` dispatch
_Update: we only support Tensors for prepend and append in this PR. See discussion below and comments for more details._
Currently there is a quirk in the c++ API based on how this is implemented: it is not possible to specify scalar prepend and appends without also specifying all 4 arguments.
That is because the goal is to match NumPy's diff signature of `diff(int n=1, int dim=-1, Union[Scalar, Tensor] prepend=None, Union[Scalar, Tensor] append)=None` where all arguments are optional, positional and in the correct order.
There are a couple blockers. One is c++ ambiguity. This prevents us from simply doing `diff(int n=1, int dim=-1, Scalar? prepend=None, Tensor? append=None)` etc for all combinations of {Tensor, Scalar} x {Tensor, Scalar}.
Why not have append, prepend not have default args and then write out the whole power set of {Tensor, Scalar, omitted} x {Tensor, Scalar, omitted} you might ask. Aside from having to write 18 overloads, this is actually illegal because arguments with defaults must come after arguments without defaults. This would mean having to write `diff(prepend, append, n, dim)` which is not desired. Finally writing out the entire power set of all arguments n, dim, prepend, append is out of the question because that would actually involve 2 * 2 * 3 * 3 = 36 combinations. And if we include the out variant, that would be 72 overloads!
With this in mind, the current way this is implemented is actually to still do `diff(int n=1, int dim=-1, Scalar? prepend=None, Tensor? append=None)`. But also make use of `cpp_no_default_args`. The idea is to only have one of the 4 {Tensor, Scalar} x {Tensor, Scalar} provide default arguments for the c++ api, and add `cpp_no_default_args` for the remaining 3 overloads. With this, Python api works as expected, but some calls such as `diff(prepend=1)` won't work on c++ api.
We can optionally add 18 more overloads that cover the {dim, n, no-args} x {scalar-tensor, tensor-scalar, scalar-scalar} x {out, non-out} cases for c++ api. _[edit: counting is hard - just realized this number is still wrong. We should try to count the cases we do cover instead and subtract that from the total: (2 * 2 * 3 * 3) - (3 + 2^4) = 17. 3 comes from the 3 of 4 combinations of {tensor, scalar}^2 that we declare to be `cpp_no_default_args`, and the one remaining case that has default arguments has covers 2^4 cases. So actual count is 34 additional overloads to support all possible calls]_
_[edit: thanks to https://github.com/pytorch/pytorch/issues/50767 hacky_wrapper is no longer necessary; it is removed in the latest commit]_
hacky_wrapper was also necessary here because `Tensor?` will cause dispatch to look for the `const optional<Tensor>&` schema but also generate a `const Tensor&` declaration in Functions.h. hacky_wrapper allows us to define our function as `const Tensor&` but wraps it in optional for us, so this avoids both the errors while linking and loading.
_[edit: rewrote the above to improve clarity and correct the fact that we actually need 18 more overloads (26 total), not 18 in total to complete the c++ api]_
Pull Request resolved: https://github.com/pytorch/pytorch/pull/50569
Reviewed By: H-Huang
Differential Revision: D26176105
Pulled By: soulitzer
fbshipit-source-id: cd8e77cc2de1117c876cd71c29b312887daca33f
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/51578https://github.com/pytorch/pytorch/pull/49710 introduced an edge case in which
drawing a single sample resulted in ignoring the `dtype` arg to `draw`. This
fixes this and adds a unit test to cover this behavior.
Test Plan: Unit tests
Reviewed By: danielrjiang
Differential Revision: D26204393
fbshipit-source-id: 441a44dc035002e7bbe6b662bf6d1af0e2cd88f4
Summary:
Performs the update that was suggested in https://github.com/pytorch/pytorch/issues/41489
Adjust the functionality to largely match that pf the scipy companion PR https://github.com/scipy/scipy/pull/10844/, including
- a new `draw_base2` method
- include zero as the first point in the (unscrambled) Sobol sequence
The scipy PR is also quite opinionated if the `draw` method doesn't get called with a base 2 number (for which the resulting sequence has nice properties, see the scipy PR for a comprehensive discussion of this).
Note that this update is a **breaking change** in the sense that sequences generated with the same parameters after as before will not be identical! They will have the same (better, arguably) distributional properties, but calling the engine with the same seed will result in different numbers in the sequence.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/49710
Test Plan:
```
from torch.quasirandom import SobolEngine
sobol = SobolEngine(3)
sobol.draw(4)
sobol = SobolEngine(4, scramble=True)
sobol.draw(5)
sobol = SobolEngine(4, scramble=True)
sobol.draw_base2(2)
```
Reviewed By: malfet
Differential Revision: D25657233
Pulled By: Balandat
fbshipit-source-id: 9df50a14631092b176cc692b6024aa62a639ef61
Summary:
Reference: https://github.com/pytorch/pytorch/issues/33152
Changes
* Enable complex support for masked_scatter
* Enable half support for masked_scatter CPU
* Enable complex autograd support for masked_scatter CPU and masked_select (both CPU and CUDA).
**Note**:
Complex Support for masked_scatter CUDA is disabled as it depends on `masked_fill` which is yet to be ported to ATen.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/51281
Reviewed By: ailzhang
Differential Revision: D26127561
Pulled By: anjali411
fbshipit-source-id: 6284926b934942213c5dfc24b5bcc8538d0231af
Summary:
Fixes https://github.com/pytorch/pytorch/issues/3307
Previously, `self.grad` was not ~cloned~ deepcopied to the returned tensor in `deepcopy`. Added a test and an implementation.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/50663
Reviewed By: heitorschueroff
Differential Revision: D26074811
Pulled By: albanD
fbshipit-source-id: 536dad36415f1d03714b4ce57453f406ad802b8c
Summary:
All these Unary operators have been an entry in OpInfo DB.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/50096
Reviewed By: zhangguanheng66
Differential Revision: D25870048
Pulled By: mruberry
fbshipit-source-id: b64e06d5b9ab5a03a202cda8c22fdb7e4ae8adf8
Summary:
Based on ngimel's (Thank you!) feedback, cpu half was only accidental, so I'm removing it.
This lets us ditch the old codepath for without replacement in favour of the new, better one.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/50063
Reviewed By: mruberry
Differential Revision: D25772449
Pulled By: ngimel
fbshipit-source-id: 608729c32237de4ee6d1acf7e316a6e878dac7f0
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/49552
This PR:
1. Migrates independent autograd test for `hstack`, `dstack`, `vstack`, `movedim`, `moveaxis` from `test_autograd.py` to the new `OpInfo` based tests.
2. Migrates autograd test for `gather`, `index_select` from the method_tests to the new `OpInfo` based tests.
2. Enables complex backward for `stack, gather, index_select, index_add_` and adds tests for complex autograd for all the above mentioned ops.
Test Plan: Imported from OSS
Reviewed By: mruberry
Differential Revision: D25682511
Pulled By: anjali411
fbshipit-source-id: 5d8f89db4a9ec340ab99a6196987d44a23e2c6c6
Summary:
**BC-breaking Note:**
This PR updates PyTorch's digamma function to be consistent with SciPy's special.digamma function. This changes the result of the digamma function on the nonpositive integers, where the gamma function is not defined. Since the gamma function is undefined at these points, the (typical) derivative of the logarithm of the gamma function is also undefined at these points, and for negative integers this PR updates digamma to return NaN. For zero, however, it returns -inf to be consistent with SciPy.
Interestingly, SciPy made a similar change, which was noticed by at least one user: https://github.com/scipy/scipy/issues/9663#issue-396587679.
SciPy's returning of negative infinity at zero is intentional:
59347ae8b8/scipy/special/cephes/psi.c (L163)
This change is consistent with the C++ standard for the gamma function:
https://en.cppreference.com/w/cpp/numeric/math/tgamma
**PR Summary:**
Reference https://github.com/pytorch/pytorch/issues/42515
Pull Request resolved: https://github.com/pytorch/pytorch/pull/48302
Reviewed By: ngimel
Differential Revision: D25664087
Pulled By: mruberry
fbshipit-source-id: 1168e81e218bf9fe5b849db0e07e7b22e590cf73
Summary:
**BC-Breaking Note:**
This PR updates PyTorch's angle operator to be consistent with NumPy's. Previously angle would return zero for all floating point values (including NaN). Now angle returns `pi` for negative floating point values, zero for non-negative floating point values, and propagates NaNs.
**PR Summary:**
Reference: https://github.com/pytorch/pytorch/issues/42515
TODO:
* [x] Add BC-Breaking Note (Prev all real numbers returned `0` (even `nan`)) -> Fixed to match the correct behavior of NumPy.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/49163
Reviewed By: ngimel
Differential Revision: D25681758
Pulled By: mruberry
fbshipit-source-id: 54143fe6bccbae044427ff15d8daaed3596f9685
Summary:
This replaces the narrow character set APIs with the wide character set ones in `THAllocator.cpp`. This fixes the potential crashes caused by passing non-ASCII characters in `torch::from_file` on Windows.
See: https://github.com/pytorch/pytorch/issues/47422
Pull Request resolved: https://github.com/pytorch/pytorch/pull/47905
Reviewed By: zhangguanheng66
Differential Revision: D25399146
Pulled By: ezyang
fbshipit-source-id: 0a183b65de171c48ed1718fa71e773224eaf196f
Summary:
Fixes https://github.com/pytorch/pytorch/issues/45964
Indexing operators e.g. `scatter`/`gather` use tensor restriding so the `TensorIterator` built in overlap checking needs to be disabled. This adds the missing overlap checks for these operators.
In addition, some indexing operators don't work will with `MemOverlapStatus::FULL` which is explicitly allowed by `assert_no_partial_overlap`. So, I've introduced `assert_no_overlap` that will raise an error on partial _or_ full overlap.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/48651
Reviewed By: zhangguanheng66
Differential Revision: D25401047
Pulled By: ngimel
fbshipit-source-id: 53abb41ac63c4283f3f1b10a0abb037169f20b89
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/48116
If you port kernels to be structured, you get Meta kernels automatically
generated for you. This is one payoff of structured kernels.
Code generation was mercifully really simple, although at risk of
"swiss cheese" syndrome: there's two new conditionals in the codegen
to tweak behavior when generating for meta keys. It's not too bad
right now but there's a risk of things getting out of hand. One
way to rationalize the logic here would be to transmit "TensorMeta-ness"
inside the TensorOptions (so tensor_from_meta can deal with it); then
the "Meta" kernel magic would literally just be generating empty
out_impls to call after all the scaffolding is done. But I didn't
do this because it seemed like it would be more annoying short term.
Also had to teach resize_ to work on meta tensors, since we use them
to implement the out kernels.
Signed-off-by: Edward Z. Yang <ezyang@fb.com>
Test Plan: Imported from OSS
Reviewed By: bhosmer, ailzhang
Differential Revision: D25056640
Pulled By: ezyang
fbshipit-source-id: f8fcfa0dbb58a94d9b4196748f56e155f83b1521
Summary:
Creates multiple new test suites to have fewer tests in test_torch.py, consistent with previous test suite creation like test_unary_ufuncs.py and test_linalg.py.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/47356
Reviewed By: ngimel
Differential Revision: D25202268
Pulled By: mruberry
fbshipit-source-id: 75fde3ca76545d1b32b86d432a5cb7a5ba8f5bb6
Summary:
Quiet errors from flake8. Only a couple of code changes for deprecated Python syntax from before 2.4. The rest is just adding noqa markers.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/48453
Reviewed By: mruberry
Differential Revision: D25181871
Pulled By: ngimel
fbshipit-source-id: f8d7298aae783b1bce2a46827b088fc390970641
Summary:
Adding Unary Ufunc Test entry for `erf` variants.
We use scipy functions for reference implementation.
We can later update the tests once these functions will update integer input to float.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/47155
Reviewed By: ngimel
Differential Revision: D25176654
Pulled By: mruberry
fbshipit-source-id: cb08efed1468b27650cec4f87a9a34e999ebd810
Summary:
The approach is to simply reuse `torch.repeat` but adding one more functionality to tile, which is to prepend 1's to reps arrays if there are more dimensions to the tensors than the reps given in input. Thus for a tensor of shape (64, 3, 24, 24) and reps of (2, 2) will become (1, 1, 2, 2), which is what NumPy does.
I've encountered some instability with the test on my end, where I could get a random failure of the test (due to, sometimes, random value of `self.dim()`, and sometimes, segfaults). I'd appreciate any feedback on the test or an explanation for this instability so I can this.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/47974
Reviewed By: ngimel
Differential Revision: D25148963
Pulled By: mruberry
fbshipit-source-id: bf63b72c6fe3d3998a682822e669666f7cc97c58
Summary:
Adds ldexp operator for https://github.com/pytorch/pytorch/issues/38349
I'm not entirely sure the changes to `NamedRegistrations.cpp` were needed but I saw other operators in there so I added it.
Normally the ldexp operator is used along with the frexp to construct and deconstruct floating point values. This is useful for performing operations on either the mantissa and exponent portions of floating point values.
Sleef, std math.h, and cuda support both ldexp and frexp but not for all data types. I wasn't able to figure out how to get the iterators to play nicely with a vectorized kernel so I have left this with just the normal CPU kernel for now.
This is the first operator I'm adding so please review with an eye for errors.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/45370
Reviewed By: mruberry
Differential Revision: D24333516
Pulled By: ranman
fbshipit-source-id: 2df78088f00aa9789aae1124eda399771e120d3f
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/48113
Fix is simple: just treat Meta as a backend covered by AutogradOther.
This semantically makes sense, since meta kernels are just like regular
CPU/CUDA kernels, they just don't do any compute.
Signed-off-by: Edward Z. Yang <ezyang@fb.com>
Test Plan: Imported from OSS
Reviewed By: zhangguanheng66
Differential Revision: D25056641
Pulled By: ezyang
fbshipit-source-id: 7b68911982352b3e0ee8616b38cd9c70bd58a740
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/47023
DeviceType pretty clearly only needs 1 byte. DeviceIndex only needs 1 byte given that machines don't have anywhere near 255 GPUs in them as far as I know.
ghstack-source-id: 116901430
Test Plan: Existing tests, added assertion to catch if my assumption about DeviceIndex is incorrect
Reviewed By: dzhulgakov
Differential Revision: D24605460
fbshipit-source-id: 7c9a89027fcf8eebd623b7cdbf6302162c981cd2
Summary:
Reference https://github.com/pytorch/pytorch/issues/38349
Delegates to `torch.transpose` (not sure what is the best way to alias)
TODO:
* [x] Add test
* [x] Add documentation
Pull Request resolved: https://github.com/pytorch/pytorch/pull/46041
Reviewed By: gchanan
Differential Revision: D25022816
Pulled By: mruberry
fbshipit-source-id: c80223d081cef84f523ef9b23fbedeb2f8c1efc5
Summary:
Now when https://github.com/pytorch/pytorch/pull/42553 is merged we can delete a bit of code from the tests and enable some of the skipped complex tests.
Unfortunately, `test_pinverse_complex_xfailed` and `test_symeig_complex_xfailed` had bugs and it wasn't caught automatically that these tests xpass. Need to be careful next time with `unittest.expectedFailure`.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/47910
Reviewed By: zhangguanheng66
Differential Revision: D25052130
Pulled By: mruberry
fbshipit-source-id: 29512995c024b882f9cb78b7bede77733d5762d0
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/48042
Moving scalar test to a separate method so the XLA team can continue to test for the other cases without failing. Requested here https://github.com/pytorch/xla/issues/2620#issuecomment-725696108
Test Plan: Imported from OSS
Reviewed By: zhangguanheng66
Differential Revision: D25055677
Pulled By: heitorschueroff
fbshipit-source-id: 5da66bac78ea197821fee0b9b8a213ff2dc19c67
Summary:
`torch.lu_solve` now works for complex inputs both on CPU and GPU.
I moved the existing tests to `test_linalg.py` and modified them to test complex dtypes, but I didn't modify/improve the body of the tests.
Ref. https://github.com/pytorch/pytorch/issues/33152
Pull Request resolved: https://github.com/pytorch/pytorch/pull/46862
Reviewed By: nikithamalgifb
Differential Revision: D24543682
Pulled By: anjali411
fbshipit-source-id: 165bde39ef95cafebf976c5ba4b487297efe8433
Summary:
Fixed test:
- `test_is_nonzero`, this is asserting exact match, which is flaky when `TORCH_SHOW_CPP_STACKTRACES=1`, I changed this to non-exact assert
- `test_pinverse` TF32
- `test_symeig` TF32
- `test_triangular_solve_batched_many_batches_cpu_float64` precision on CPU BLAS
- `test_qr` TF32, as well as the tensor factory forgets a `dtype=dtype`
- `test_lu` TF32
- `ConvTranspose2d` TF32
- `Conv3d_1x1x1_no_bias` TF32
- `Transformer*` TF32
Pull Request resolved: https://github.com/pytorch/pytorch/pull/46941
Reviewed By: heitorschueroff
Differential Revision: D24852725
Pulled By: mruberry
fbshipit-source-id: ccd4740cc643476178d81059d1c78da34e5082ed
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/42553
Ports `torch.bmm` and `torch.baddbmm` from TH to ATen, as well as adds support for complex dtypes. Also removes dead TH code for Level 2 functions.
Closes#24539
Test Plan: Imported from OSS
Reviewed By: ansley
Differential Revision: D24893511
Pulled By: anjali411
fbshipit-source-id: 0eba3f2aec99c48b3018a5264ee7789279cfab58
Summary:
`torch.triangular_solve` now works for complex inputs on GPU.
I moved the existing tests to `test_linalg.py` and modified them to test complex and float32 dtypes.
Ref. https://github.com/pytorch/pytorch/issues/33152
Pull Request resolved: https://github.com/pytorch/pytorch/pull/46916
Reviewed By: navahgar, agolynski
Differential Revision: D24706647
Pulled By: anjali411
fbshipit-source-id: fe780eac93d2ae1b2549539bb385e5fac25213b3
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/46398
This PR makes torch.einsum compatible with numpy.einsum except for the sublist input option as requested here https://github.com/pytorch/pytorch/issues/21412. It also fixed 2 performance issues linked below and adds a check for reducing to torch.dot instead of torch.bmm which is faster in some cases.
fixes#45854, #37628, #30194, #15671fixes#41467 with benchmark below
```python
import torch
from torch.utils.benchmark import Timer
a = torch.randn(10000, 100, 101, device='cuda')
b = torch.randn(10000, 101, 3, device='cuda')
c = torch.randn(10000, 100, 1, device='cuda')
d = torch.randn(10000, 100, 1, 3, device='cuda')
print(Timer(
stmt='torch.einsum("bij,bjf->bif", a, b)',
globals={'a': a, 'b': b}
).blocked_autorange())
print()
print(Timer(
stmt='torch.einsum("bic,bicf->bif", c, d)',
globals={'c': c, 'd': d}
).blocked_autorange())
```
```
<torch.utils.benchmark.utils.common.Measurement object at 0x7fa37c413850>
torch.einsum("bij,bjf->bif", a, b)
Median: 4.53 ms
IQR: 0.00 ms (4.53 to 4.53)
45 measurements, 1 runs per measurement, 1 thread
<torch.utils.benchmark.utils.common.Measurement object at 0x7fa37c413700>
torch.einsum("bic,bicf->bif", c, d)
Median: 63.86 us
IQR: 1.52 us (63.22 to 64.73)
4 measurements, 1000 runs per measurement, 1 thread
```
fixes#32591 with benchmark below
```python
import torch
from torch.utils.benchmark import Timer
a = torch.rand(1, 1, 16, 2, 16, 2, 16, 2, 2, 2, 2, device="cuda")
b = torch.rand(729, 1, 1, 2, 1, 2, 1, 2, 2, 2, 2, device="cuda")
print(Timer(
stmt='(a * b).sum(dim = (-3, -2, -1))',
globals={'a': a, 'b': b}
).blocked_autorange())
print()
print(Timer(
stmt='torch.einsum("...ijk, ...ijk -> ...", a, b)',
globals={'a': a, 'b': b}
).blocked_autorange())
```
```
<torch.utils.benchmark.utils.common.Measurement object at 0x7efe0de28850>
(a * b).sum(dim = (-3, -2, -1))
Median: 17.86 ms
2 measurements, 10 runs per measurement, 1 thread
<torch.utils.benchmark.utils.common.Measurement object at 0x7efe0de286a0>
torch.einsum("...ijk, ...ijk -> ...", a, b)
Median: 296.11 us
IQR: 1.38 us (295.42 to 296.81)
662 measurements, 1 runs per measurement, 1 thread
```
TODO
- [x] add support for ellipsis broadcasting
- [x] fix corner case issues with sumproduct_pair
- [x] update docs and add more comments
- [x] add tests for error cases
Test Plan: Imported from OSS
Reviewed By: malfet
Differential Revision: D24860367
Pulled By: heitorschueroff
fbshipit-source-id: 31110ee598fd598a43acccf07929b67daee160f9
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/45092
Adding two operators
1. at::float_to_half -> Converts FP32 tensor to FP16 tensor
2. at::half_to_float -> Converts FP16 tensor to FP32 tensor.
These operators internally use the kernel provided by FBGeMM. Both C2 and PT will use the same FBGeMM kernel underneath.
Test Plan:
buck test //caffe2/test:torch -- .*test_half_tensor.*
Run benchmark locally using
```
buck run //caffe2/benchmarks/operator_benchmark/pt:tensor_to_test
```
AI Bench results are pending. I expect that not to finish as we have large queue with jobs pending for 2+ days.
Benchmark for 512x512 tensor with FbGeMM implementation
```
# ----------------------------------------
# PyTorch/Caffe2 Operator Micro-benchmarks
# ----------------------------------------
# Tag : short
# Benchmarking PyTorch: FloatToHalfTensorConversionBenchmark
# Mode: Eager
# Name: FloatToHalfTensorConversionBenchmark_M512_N512_cpu
# Input: M: 512, N: 512, device: cpu
Forward Execution Time (us) : 1246.332
# Benchmarking PyTorch: HalfToFloatTensorConversionBenchmark
# Mode: Eager
# Name: HalfToFloatTensorConversionBenchmark_M512_N512_cpu
# Input: M: 512, N: 512, device: cpu
Forward Execution Time (us) : 1734.304
```
Benchmark for 512x512 tensor trunk with no FbGeMM integration.
```
# ----------------------------------------
# PyTorch/Caffe2 Operator Micro-benchmarks
# ----------------------------------------
# Tag : short
# Benchmarking PyTorch: FloatToHalfTensorConversionBenchmark
# Mode: Eager
# Name: FloatToHalfTensorConversionBenchmark_M512_N512_cpu
# Input: M: 512, N: 512, device: cpu
Forward Execution Time (us) : 169045.724
# Benchmarking PyTorch: HalfToFloatTensorConversionBenchmark
# Mode: Eager
# Name: HalfToFloatTensorConversionBenchmark_M512_N512_cpu
# Input: M: 512, N: 512, device: cpu
Forward Execution Time (us) : 152382.494
```
Reviewed By: ngimel
Differential Revision: D23824869
fbshipit-source-id: ef044459b6c8c6e5ddded72080204c6a0ab4582c
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/47305
Fixes https://github.com/pytorch/pytorch/issues/47127.
Ideally this would just use diag and sum (as the CUDA implementation does), but that seems to have performance problems, which I'll link in the github PR.
Test Plan: Imported from OSS
Reviewed By: zou3519
Differential Revision: D24729627
Pulled By: gchanan
fbshipit-source-id: 151b786b53e7b958f0929c803dbf8e95981c6884
