Summary:
Adds native_dropout to have a reasonable target for torchscript in auto diff. native_dropout has scale and train as arguments in its signature, this makes native_dropout more consistent with other operators and removes conditionals in the autodiff definition.
cc gmagogsfm
Pull Request resolved: https://github.com/pytorch/pytorch/pull/63937
Reviewed By: mruberry
Differential Revision: D32477657
Pulled By: ngimel
fbshipit-source-id: d37b137a37acafa50990f60c77f5cea2818454e4
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/63568
This PR adds the first solver with structure to `linalg`. This solver
has an API compatible with that of `linalg.solve` preparing these for a
possible future merge of the APIs. The new API:
- Just returns the solution, rather than the solution and a copy of `A`
- Removes the confusing `transpose` argument and replaces it by a
correct handling of conj and strides within the call
- Adds a `left=True` kwarg. This can be achieved via transposes of the
inputs and the result, but it's exposed for convenience.
This PR also implements a dataflow that minimises the number of copies
needed before calling LAPACK / MAGMA / cuBLAS and takes advantage of the
conjugate and neg bits.
This algorithm is implemented for `solve_triangular` (which, for this, is
the most complex of all the solvers due to the `upper` parameters).
Once more solvers are added, we will factor out this calling algorithm,
so that all of them can take advantage of it.
Given the complexity of this algorithm, we implement some thorough
testing. We also added tests for all the backends, which was not done
before.
We also add forward AD support for `linalg.solve_triangular` and improve the
docs of `linalg.solve_triangular`. We also fix a few issues with those of
`torch.triangular_solve`.
Resolves https://github.com/pytorch/pytorch/issues/54258
Resolves https://github.com/pytorch/pytorch/issues/56327
Resolves https://github.com/pytorch/pytorch/issues/45734
cc jianyuh nikitaved pearu mruberry walterddr IvanYashchuk xwang233 Lezcano
Test Plan: Imported from OSS
Reviewed By: zou3519, JacobSzwejbka
Differential Revision: D32283178
Pulled By: mruberry
fbshipit-source-id: deb672e6e52f58b76536ab4158073927a35e43a8
Summary:
### Create `linalg.cross`
Fixes https://github.com/pytorch/pytorch/issues/62810
As discussed in the corresponding issue, this PR adds `cross` to the `linalg` namespace (**Note**: There is no method variant) which is slightly different in behaviour compared to `torch.cross`.
**Note**: this is NOT an alias as suggested in mruberry's [https://github.com/pytorch/pytorch/issues/62810 comment](https://github.com/pytorch/pytorch/issues/62810#issuecomment-897504372) below
> linalg.cross being consistent with the Python Array API (over NumPy) makes sense because NumPy has no linalg.cross. I also think we can implement linalg.cross without immediately deprecating torch.cross, although we should definitely refer users to linalg.cross. Deprecating torch.cross will require additional review. While it's not used often it is used, and it's unclear if users are relying on its unique behavior or not.
The current default implementation of `torch.cross` is extremely weird and confusing. This has also been reported multiple times previously. (See https://github.com/pytorch/pytorch/issues/17229, https://github.com/pytorch/pytorch/issues/39310, https://github.com/pytorch/pytorch/issues/41850, https://github.com/pytorch/pytorch/issues/50273)
- [x] Add `torch.linalg.cross` with default `dim=-1`
- [x] Add OpInfo and other tests for `torch.linalg.cross`
- [x] Add broadcasting support to `torch.cross` and `torch.linalg.cross`
- [x] Remove out skip from `torch.cross` OpInfo
- [x] Add docs for `torch.linalg.cross`. Improve docs for `torch.cross` mentioning `linalg.cross` and the difference between the two. Also adds a warning to `torch.cross`, that it may change in the future (we might want to deprecate it later)
---
### Additional Fixes to `torch.cross`
- [x] Fix Doc for Tensor.cross
- [x] Fix torch.cross in `torch/overridres.py`
While working on `linalg.cross` I noticed these small issues with `torch.cross` itself.
[Tensor.cross docs](https://pytorch.org/docs/stable/generated/torch.Tensor.cross.html) still mentions `dim=-1` default which is actually wrong. It should be `dim=None` after the behaviour was updated in PR https://github.com/pytorch/pytorch/issues/17582 but the documentation for the `method` or `function` variant wasn’t updated. Later PR https://github.com/pytorch/pytorch/issues/41850 updated the documentation for the `function` variant i.e `torch.cross` and also added the following warning about the weird behaviour.
> If `dim` is not given, it defaults to the first dimension found with the size 3. Note that this might be unexpected.
But still, the `Tensor.cross` docs were missed and remained outdated. I’m finally fixing that here. Also fixing `torch/overrides.py` for `torch.cross` as well now, with `dim=None`.
To verify according to the docs the default behaviour of `dim=-1` should raise, you can try the following.
```python
a = torch.randn(3, 4)
b = torch.randn(3, 4)
b.cross(a) # this works because the implementation finds 3 in the first dimension and the default behaviour as shown in documentation is actually not true.
>>> tensor([[ 0.7171, -1.1059, 0.4162, 1.3026],
[ 0.4320, -2.1591, -1.1423, 1.2314],
[-0.6034, -1.6592, -0.8016, 1.6467]])
b.cross(a, dim=-1) # this raises as expected since the last dimension doesn't have a 3
>>> RuntimeError: dimension -1 does not have size 3
```
Please take a closer look (particularly the autograd part, this is the first time I'm dealing with `derivatives.yaml`). If there is something missing, wrong or needs more explanation, please let me know. Looking forward to the feedback.
cc mruberry Lezcano IvanYashchuk rgommers
Pull Request resolved: https://github.com/pytorch/pytorch/pull/63285
Reviewed By: gchanan
Differential Revision: D32313346
Pulled By: mruberry
fbshipit-source-id: e68c2687c57367274e8ddb7ef28ee92dcd4c9f2c
Summary:
Adds `torch.argwhere` as an alias to `torch.nonzero`
Currently, `torch.nonzero` is actually provides equivalent functionality to `np.argwhere`.
From NumPy docs,
> np.argwhere(a) is almost the same as np.transpose(np.nonzero(a)), but produces a result of the correct shape for a 0D array.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/64257
Reviewed By: qihqi
Differential Revision: D32049884
Pulled By: saketh-are
fbshipit-source-id: 016e49884698daa53b83e384435c3f8f6b5bf6bb
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/64430
The functionalization pass needs `{view}_scatter` versions of the slice/select/diagonal ops in order to correctly propagate mutations from a view to its base. On top of that, the implementations need to be primitive w.r.t. autograd, because they look something like `...slice().copy_()`, and the functionalization pass can't use views + mutations inside of it's own alias-removal machinery!
I added some basic tests that I tried to base off of existing tests for views (particularly around testing the derivative formulas), but I'm wondering if I should add something more comprehensive.
Also, as_strided fits into this category - the functionalization pass will need an `as_strided_scatter` op that's primitive w.r.t. autograd. I didn't add it for now, because it'll involve duplicating a bunch of logic from the current `as_strided_backward()` function, and also writing a derivative formula that I wasn't sure how to write :)
Test Plan: Imported from OSS
Reviewed By: albanD
Differential Revision: D31942092
Pulled By: bdhirsh
fbshipit-source-id: c702a57c2748a7c771c14e4bcc3e996b48fcc4c8
Summary:
Adds mixed precision autocasting support between fp32/fp16 to torchscript/JIT. More in depth descriptoin can be found at [torch/csrc/jit/JIT-AUTOCAST.md](https://github.com/pytorch/pytorch/pull/63939/files#diff-1f1772aaa508841c5bb58b74ab98f49a1e577612cd9ea5c386c8714a75db830b)
This PR implemented an autocast optimization pass that inserts casting ops per AMP rule (torch/csrc/jit/passes/autocast.cpp), that mimics the behavior of eager autocast. The pass also takes into consideration the context of `torch.cuda.amp.autocast` and only inserts casting ops within the enabled context manager, giving feature parity as with eager amp autocast.
We currently provide JIT AMP autocast as a prototyping feature, so it is default off and could be turned on via `torch._C._jit_set_autocast_mode(True)`
The JIT support for autocast is subject to different constraints compared to the eager mode implementation (mostly related to the fact that TorchScript is statically typed), restriction on the user facing python code is described in doc torch/csrc/jit/JIT-AUTOCAST.md
This is a prototype, there are also implementation limitation that's necessary to keep this PR small and get something functioning quickly on upstream, so we can iterate on designs.
Few limitation/challenge that is not properly resolved in this PR:
1. Autocast inserts cast operation, which would have impact on scalar type of output tensor feeding downstream operations. We are not currently propagating the updated scalar types, this would give issues/wrong results on operations in promotion rules.
2. Backward for autodiff in JIT misses the casting of dgrad to input scalar type, as what autograd does in eager. This forces us to explicitly mark the casting operation for certain operations (e.g. binary ops), otherwise, we might be feeding dgrad with mismatch scalar type to input. This could potentially break gradient function consuming dgrad. (e.g. gemm backwards, which assumes grad_output to be of same scalar type as input')
3. `torch.autocast` api has an optional argument `dtype` which is not currently supported in the JIT autocast and we require a static value.
Credit goes mostly to:
tlemo
kevinstephano
Pull Request resolved: https://github.com/pytorch/pytorch/pull/63939
Reviewed By: navahgar
Differential Revision: D31093381
Pulled By: eellison
fbshipit-source-id: da6e26c668c38b01e296f304507048d6c1794314
Summary:
Adds `torch.argwhere` as an alias to `torch.nonzero`
Currently, `torch.nonzero` is actually provides equivalent functionality to `np.argwhere`.
From NumPy docs,
> np.argwhere(a) is almost the same as np.transpose(np.nonzero(a)), but produces a result of the correct shape for a 0D array.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/64257
Reviewed By: dagitses
Differential Revision: D31474901
Pulled By: saketh-are
fbshipit-source-id: 335327a4986fa327da74e1fb8624cc1e56959c70
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/62030
Remove dtype tracking from Python Storage interface, remove all the different `<type>Storage` classes except for `ByteStorage`, and update serialization accordingly, while maintaining as much FC/BC as possible
Fixes https://github.com/pytorch/pytorch/issues/47442
* **THE SERIALIZATION FORMAT IS FULLY FC/BC.** We worked very hard to make sure this is the case. We will probably want to break FC at some point to make the serialization structure of tensors make more sense, but not today.
* There is now only a single torch.ByteStorage class. Methods like `Tensor.set_` no longer check that the dtype of storage is appropriate.
* As we no longer know what dtype of a storage is, we've **removed** the size method from Storage, replacing it with nbytes. This is to help catch otherwise silent errors where you confuse number of elements with number of bytes.
* `Storage._new_shared` takes a `nbytes` kwarg and will reject previous positional only calls. `Storage._new_with_file` and `_set_from_file` require explicit element size arguments.
* It's no longer possible to convert storages to different types using the float/double/etc methods. Instead, do the conversion using a tensor.
* It's no longer possible to allocate a typed storage directly using FloatStorage/DoubleStorage/etc constructors. Instead, construct a tensor and extract its storage. The classes still exist but they are used purely for unpickling.
* The preexisting serialization format stores dtype with storage, and in fact this dtype is used to determine the dtype of the tensor overall.
To accommodate this case, we introduce a new TypedStorage concept that exists only during unpickling time which is used to temporarily store the dtype so we can construct a tensor. **If you overrode the handling of pickling/unpickling, you MUST add handling for TypedStorage** or your serialization code will degrade to standard file-based serialization.
Original pull request: https://github.com/pytorch/pytorch/pull/59671
Reviewed By: soulitzer, ngimel
Differential Revision: D29466819
Pulled By: ezyang
fbshipit-source-id: 4a14e5d3c2b08e06e558683d97f7378a3180b00e
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/65621
Add a new attribute to the FusedMovingAvgObsFakeQuantize that controls if the Fake Quant operation should be applied at the output of a particular layer. The motivation is to give the users additional control to control the numerics of the fake_quant operators during training. It defaults to always fake quant the output (True).
Note: We will still observer the tensors as before (only the fake_quant operation is controlled using this flag)
For example
```
input model
x -> fc1 -> fc2 -> non_quantizable_op -> fc3
After fake_quant
x -> fake_quant(x) -> fc1 -> fake_quant(fc1) -> fc2 -> fake_quant(fc2) -> non_quantizable_op -> fake_quant() -> fc3 -> fake_quantize(fc3)
With output_fake_quant disabled at the output of fc2 and fc3 (since their outputs are non-quantizable)
x -> fake_quant(x) -> fc1 -> fake_quant(fc1) -> fc2 -> non_quantizable_op -> fake_quant() -> fc3
```
Test Plan: ./buck-out/gen/caffe2/test/quantization_fx\#binary.par -r test_disable_output_fake_quant
Reviewed By: jerryzh168
Differential Revision: D31174526
fbshipit-source-id: bffe776216d041fb09133a6fb09bfc2c0bb46b89
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/65340
I thought about a few possible ways of doing this. The main hazard is
that if I create a CPU tensor that doesn't have any real storage, the
moment I actually try to access the data on the tensor I will segfault.
So I don't want to use _make_subclass on a "cpu meta tensor" because
the CPU meta tensor (with no subclass) is radioactive: printing it
will immediately cause a segfault. So instead, I have to create
the CPU meta tensor AND subclass all in one go, and that means I need
another function for it. One downside to doing it this way is
I need another overload for explicit strides, and in general it is
difficult to get the view relationships to all work out properly;
tracked at https://github.com/pytorch/pytorch/issues/65339
Fixes https://github.com/pytorch/pytorch/issues/62972
Fixes https://github.com/pytorch/pytorch/issues/62730
Signed-off-by: Edward Z. Yang <ezyang@fb.com>
Test Plan: Imported from OSS
Reviewed By: albanD
Differential Revision: D31057231
Pulled By: ezyang
fbshipit-source-id: 73522769e093ae8a1bf0c7f7e594659bfb827b28
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/62671
Very crude first implementation of `torch.nanmean`. The current reduction kernels do not have good support for implementing nan* variants. Rather than implementing new kernels for each nan* operator, I will work on new reduction kernels with support for a `nan_policy` flag and then I will port `nanmean` to use that.
**TODO**
- [x] Fix autograd issue
Test Plan: Imported from OSS
Reviewed By: malfet
Differential Revision: D30515181
Pulled By: heitorschueroff
fbshipit-source-id: 303004ebd7ac9cf963dc4f8e2553eaded5f013f0
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/64689
This brings it in line with the C++ implementation.
Fixes https://github.com/pytorch/pytorch/issues/64687
Signed-off-by: Edward Z. Yang <ezyang@fb.com>
Test Plan: Imported from OSS
Reviewed By: albanD
Differential Revision: D30816215
Pulled By: ezyang
fbshipit-source-id: ed36af6c35467ae678d9548197efd97c36d38dec
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/63552
In this PR, we want to exclude these 2 cases in the `Autocast` weight cache usages:
- Using `torch.jit.trace` under the `Autocast`
As report in https://github.com/pytorch/pytorch/issues/50231 and several other discussions, using `torch.jit.trace` under the `Autocast`, the trace process would hit Autocast's weight cache and fails. So we should disable weight cache under the trace process.
- Using `Autocast` with `Grad mode`
- Usually we are using `Grad mode` for training. Since in the training phase, the weight will change in every step. So we doesn't need to cache the weight.
- For the recommended `Autocast` training case in the [doc](https://pytorch.org/docs/stable/amp.html), `Autocast` will clear the cache every step leaving the context. We should disable it to save the clear operations.
```
model = Net().cuda()
optimizer = optim.SGD(model.parameters(), ...)
for input, target in data:
optimizer.zero_grad()
with autocast():
output = model(input)
loss = loss_fn(output, target)
loss.backward()
optimizer.step()
```
Test Plan: Imported from OSS
Reviewed By: mrshenli
Differential Revision: D30644913
Pulled By: ezyang
fbshipit-source-id: ad7bc87372e554e7aa1aa0795e9676871b3974e7
Summary:
Fixes https://github.com/pytorch/pytorch/issues/62811
Add `torch.linalg.matmul` alias to `torch.matmul`. Note that the `linalg.matmul` doesn't have a `method` variant.
Also cleaning up `torch/_torch_docs.py` when formatting is not needed.
cc IvanYashchuk Lezcano mruberry rgommers
Pull Request resolved: https://github.com/pytorch/pytorch/pull/63227
Reviewed By: mrshenli
Differential Revision: D30770235
Pulled By: mruberry
fbshipit-source-id: bfba77dfcbb61fcd44f22ba41bd8d84c21132403
Summary:
Fixes https://github.com/pytorch/pytorch/issues/61767
## Changes
- [x] Add `torch.concat` alias to `torch.cat`
- [x] Add OpInfo for `cat`/`concat`
- [x] Fix `test_out` skips (Use `at::native::resize_output` or `at::native::resize_output_check`)
- [x] `cat`/`concat`
- [x] `stack`
- [x] `hstack`
- [x] `dstack`
- [x] `vstack`/`row_stack`
- [x] Remove redundant tests for `cat`/`stack`
~I've not added `cat`/`concat` to OpInfo `op_db` yet, since cat is a little more tricky than other OpInfos (should have a lot of tests) and currently there are no OpInfos for that. I can try to add that in a subsequent PR or maybe here itself, whatever is suggested.~
**Edit**: cat/concat OpInfo has been added.
**Note**: I've added the named tensor support for `concat` alias as well, maybe that's out of spec in `array-api` but it is still useful for consistency in PyTorch.
Thanks to krshrimali for guidance on my first PR :))
cc mruberry rgommers pmeier asmeurer leofang AnirudhDagar asi1024 emcastillo kmaehashi heitorschueroff krshrimali
Pull Request resolved: https://github.com/pytorch/pytorch/pull/62560
Reviewed By: saketh-are
Differential Revision: D30762069
Pulled By: mruberry
fbshipit-source-id: 6985159d1d9756238890488a0ab3ae7699d94337
Summary:
This PR implements the necessary hooks/stubs/enums/etc for complete ONNX Runtime (ORT) Eager Mode integration. The actual extension will live out of tree at https://github.com/pytorch/ort.
We have been [working on this at Microsoft](https://github.com/microsoft/onnxruntime-pytorch/tree/eager-ort/torch_onnxruntime) for the last few months, and are finally ready to contribute the PyTorch core changes upstream (nothing major or exciting, just the usual boilerplate for adding new backends).
The ORT backend will allow us to ferry [almost] all torch ops into granular ONNX kernels that ORT will eagerly execute against any devices it supports (therefore, we only need a single ORT backend from a PyTorch perspective).
Pull Request resolved: https://github.com/pytorch/pytorch/pull/58248
Reviewed By: astaff
Differential Revision: D30344992
Pulled By: albanD
fbshipit-source-id: 69082b32121246340d686e16653626114b7714b2
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/59077Fixes#58549
`from_buffer` constructs a tensor object from an already allocated buffer through
CPython's buffer protocol. Besides the standard `dtype`, `count`, and `offset` parameters,
this function also accepts:
- `device`: where the buffer lives
- `requires_grad`: should autograd record operations on the new tensor
A new test file _test_buffer_protocol.py_ was created. Currently, only CPU tests were
implemented. That's because neither PyTorch nor Numba implements CPython's buffer
protocol. Therefore, there's no way to create a CUDA buffer with the existing
dependencies (could use PyCUDA for that, though).
At the moment, if `device` differs from the device the buffer actually lives, two things
may happen:
- `RuntimeError`, if `device='cuda'`
- Segmentation fault (not tested -- see above), if `device='cpu'`
Test Plan: Imported from OSS
Reviewed By: jbschlosser
Differential Revision: D29870914
Pulled By: mruberry
fbshipit-source-id: 9fa8611aeffedfe39c9af74558178157a11326bb
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/61570
Fused operator that computes moving average min/max values (in-place) of the input tensor and fake-quantizes it.
It expects the qmin/qmax values to reflect the range of the quantized tensor (instead of reduce_range)
Motivation for adding this operator is for performance reasons, since moving the computation from python to C++/CUDA can increase the performance of QAT.
Test Plan:
python test/test_quantization.py TestFusedObsFakeQuant
Imported from OSS
Reviewed By: vkuzo
Differential Revision: D29682762
fbshipit-source-id: 28e4c50e77236d6976fe4b326c9a12103ed95840
Summary:
This PR un-reverts https://github.com/pytorch/pytorch/issues/61475 + fixes compilation with MSVC, that does not recognize alternative operator spellings (i.e. using `or` instead of `||` )
Pull Request resolved: https://github.com/pytorch/pytorch/pull/61937
Reviewed By: albanD
Differential Revision: D29805941
Pulled By: malfet
fbshipit-source-id: 01e5963c6717c1b44b260300d87ba0bf57f26ce9
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/56058
User facing changes:
1. Adds a negative bit and corresponding new API (`is_neg()`,`resolve_neg()`)
2. `tensor.conj().imag` now returns a floating point tensor with neg bit set to 1 instead of a tensor with no notion of negative bit. Note that imag is still a view and all the view properties still hold for imag.
Non user facing changes:
1. Added a new Negative dispatch key and a backend fallback to handle it
2. Updated copy kernel to handle negative bit
3. Merged conjugate and negative bit fallback kernel
4. fixed https://github.com/pytorch/pytorch/issues/60478 (caused due to https://github.com/pytorch/pytorch/pull/54987)
Testing:
1. Added a new OpInfo based test `test_neg_view` (verifies that out-of-place and in-place operations work correctly for all operations when the input is a neg view tensor by checking the result against an actually negated tensor, verifies that autograd returns the same output for both neg view and actually negated tensors as well as it works fine when grad_out is a neg view).
2. Added a new test class containing `test_conj_view`, `test_neg_view`.
Test Plan: Imported from OSS
Reviewed By: soulitzer
Differential Revision: D29636403
fbshipit-source-id: 12214c9dc4806c51850f4a72a109db9527c0ca63
Summary:
Based from https://github.com/pytorch/pytorch/pull/50466
Adds the initial implementation of `torch.cov` similar to `numpy.cov`. For simplicity, we removed support for many parameters in `numpy.cov` that are either redundant such as `bias`, or have simple workarounds such as `y` and `rowvar`.
cc PandaBoi
closes https://github.com/pytorch/pytorch/issues/19037
Pull Request resolved: https://github.com/pytorch/pytorch/pull/58311
Reviewed By: jbschlosser
Differential Revision: D29431651
Pulled By: heitorschueroff
fbshipit-source-id: 167dea880f534934b145ba94291a9d634c25b01b
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/58059
Add CUDA.used vital sign which is true only if CUDA was "used" which technically means the context was created.
Also adds the following features:
- Force vitals to be written even if vitals are disabled, to enable testing when the env variable is not set from the start of execution
- Add a read_vitals call for python to read existing vital signs.
Test Plan: buck test mode/dbg caffe2/test:torch -- --regex basic_vitals
Reviewed By: xuzhao9
Differential Revision: D28357615
fbshipit-source-id: 681bf9ef63cb1458df9f1c241d301a3ddf1e5252
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/59760
See https://github.com/pytorch/pytorch/issues/59049
There are some moving parts to this PR, I'll structure this explanation so the straightforward parts go first, and then the less straightforward parts.
**The actual dispatch to Python.** The core logic of dispatch to Python lives in `concrete_dispatch_fn` in `torch/csrc/autograd/python_variable.cpp`. It takes the input IValue stack, scans all the arguments for Tensor arguments, and defers most of the heavy lifting to `handle_torch_function_no_python_arg_parser` which actually does all of the logic for calling out to torch dispatch (in particular, this function handles multiple dispatch situations for you). Because we have a different function name than regular `__torch_function__` handling, `handle_torch_function_no_python_arg_parser` is generalized to accept a magic method name to look for when testing if Tensors have custom handling or not. Unlike `__torch_function__`, by default there is no `__torch_dispatch__` on Tensor classes.
**Maintaining the Python dispatch key.** In order to get to the dispatch to Python logic, we must tag Tensors with the `__torch_dispatch__` magic method with the newly added Python dispatch key (separated from PythonFuncTorch to allow for a transitional period while they migrate to this mechanism). We expose a new private property `_is_python_dispatch` that assists in debugging if a Tensor is participating in Python dispatch or not. We apply the Python dispatch key the first time a PyObject for a Tensor is constructed (THPVariable_NewWithVar), testing if `__torch_dispatch__` exists with then newly added `check_has_torch_dispatch`.
**Shallow copy and detach.** For the simple examples tested in this PR, most creations of Tensor route through the dispatcher. The exception to this is `shallow_copy_and_detach`, which bypasses the dispatcher and is used when saving tensors for backwards. When a Tensor is Python dispatch, we override the behavior of `shallow_copy_and_detach` to instead directly call into `__torch_dispatch__` to perform a `detach` operation (in the same way it would be invoked if you called `detach` directly). Because this Python call is triggered directly from c10::TensorImpl, it must be indirected through `PyInterpreter::detach`, which is the general mechanism for dynamic dispatching to the Python interpreter associated with a TensorImpl.
**torchdeploy compatibility.** The dispatch to Python logic cannot be directly registered to the dispatcher as it is compiled in the Python library, which will get loaded multiple times per torchdeploy interpreter. Thus, we must employ a two phase process. First, we register a fallback inside a non-Python library (aten/src/ATen/core/PythonFallbackKernel.cpp). Its job is to determine the appropriate PyInterpreter to handle the Python dispatch by going through all of the arguments and finding the first argument that has a PyObject/PyInterpreter. With this PyInterpreter, it makes another dynamic dispatch via "dispatch" which will go to the correct torchdeploy interpreter to handle dispatching to actual Python.
**Testing.** We provide a simple example of a LoggingTensor for testing, which can be used to generate TorchScript-like traces to observe what operations are being called when a Tensor is invoked. Although a LoggingTensor would be better implemented via an is-a relationship rather than a has-a relationship (as is done in the test), we've done it this way to show that arbitrarily complex compositions of tensors inside a tensor work properly.
**Known limitations.**
* We haven't adjusted any operator code, so some patterns may not work (as they lose the Python subclass in an unrecoverable way)
* `__torch_function__` must be explicitly disabled with `_disabled_torch_function_impl` otherwise things don't work quite correctly (in particular, what is being disabled is default subclass preservation behavior.)
* We don't ever populate kwargs, even when an argument is kwarg-only
Signed-off-by: Edward Z. Yang <ezyang@fb.com>
Differential Revision:
D29017912
D29017912
Test Plan: Imported from OSS
Reviewed By: bdhirsh
Pulled By: ezyang
fbshipit-source-id: a67714d9e541d09203a8cfc85345b8967db86238
Summary:
Reference https://github.com/pytorch/pytorch/issues/50345
`zeta` was already present in the codebase to support computation of `polygamma`.
However, `zeta` only had `double(double, double)` signature **for CPU** before the PR (which meant that computation `polygamma` were always upcasted to `double` for zeta part).
With this PR, float computations will take place in float and double in double.
Have also refactored the code and moved the duplicate code from `Math.cuh` to `Math.h`
**Note**: For scipy, q is optional, and if it is `None`, it defaults `1` which corresponds to Reimann-Zeta. However, for `torch.specia.zeta`, I made it mandatory cause for me it feels odd without `q` this is Reimann-Zeta and with `q` it is the general Hurwitz Zeta. I think sticking to just general made more sense as passing `1` for q sounds trivial.
Verify:
* [x] Docs https://14234587-65600975-gh.circle-artifacts.com/0/docs/special.html#torch.special.zeta
Pull Request resolved: https://github.com/pytorch/pytorch/pull/59623
Reviewed By: ngimel
Differential Revision: D29348269
Pulled By: mruberry
fbshipit-source-id: a3f9ebe1f7724dbe66de2b391afb9da1cfc3e4bb
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/60464
Fixes https://github.com/szagoruyko/pytorchviz/issues/65
An alternate implementation of this PR would be to remove the
__torch_function__ interposition points for these accessors entirely.
In the end, I decided to opt for extra expressivity. See
torch.overrides for the criterion on how I decided which accessors
should get the nowrap treatment.
Signed-off-by: Edward Z. Yang <ezyang@fb.com>
Test Plan: Imported from OSS
Reviewed By: albanD
Differential Revision: D29302835
Pulled By: ezyang
fbshipit-source-id: fbe0ac4530a6cc9d6759a3fdf5514d4d7b1f7690
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/60050
It doesn't work to put torch.Tensor.prop.__get__ in the ignored
list. Now it does. (Not exercised here, see next diff in stack).
Signed-off-by: Edward Z. Yang <ezyang@fb.com>
Test Plan: Imported from OSS
Reviewed By: zou3519
Differential Revision: D29171464
Pulled By: ezyang
fbshipit-source-id: e7354668b481f9275f2eb5bb3a6228d1815fecea
Summary:
Fixes https://github.com/pytorch/pytorch/issues/3025
## Background
This PR implements a function similar to numpy's [`isin()`](https://numpy.org/doc/stable/reference/generated/numpy.isin.html#numpy.isin).
The op supports integral and floating point types on CPU and CUDA (+ half & bfloat16 for CUDA). Inputs can be one of:
* (Tensor, Tensor)
* (Tensor, Scalar)
* (Scalar, Tensor)
Internally, one of two algorithms is selected based on the number of elements vs. test elements. The heuristic for deciding which algorithm to use is taken from [numpy's implementation](fb215c7696/numpy/lib/arraysetops.py (L575)): if `len(test_elements) < 10 * len(elements) ** 0.145`, then a naive brute-force checking algorithm is used. Otherwise, a stablesort-based algorithm is used.
I've done some preliminary benchmarking to verify this heuristic on a devgpu, and determined for a limited set of tests that a power value of `0.407` instead of `0.145` is a better inflection point. For now, the heuristic has been left to match numpy's, but input is welcome for the best way to select it or whether it should be left the same as numpy's.
Tests are adapted from numpy's [isin and in1d tests](7dcd29aaaf/numpy/lib/tests/test_arraysetops.py).
Note: my locally generated docs look terrible for some reason, so I'm not including the screenshot for them until I figure out why.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/53125
Test Plan:
```
python test/test_ops.py # Ex: python test/test_ops.py TestOpInfoCPU.test_supported_dtypes_isin_cpu_int32
python test/test_sort_and_select.py # Ex: python test/test_sort_and_select.py TestSortAndSelectCPU.test_isin_cpu_int32
```
Reviewed By: soulitzer
Differential Revision: D29101165
Pulled By: jbschlosser
fbshipit-source-id: 2dcc38d497b1e843f73f332d837081e819454b4e
Summary:
Based from https://github.com/pytorch/pytorch/pull/50466
Adds the initial implementation of `torch.cov` similar to `numpy.cov`. For simplicity, we removed support for many parameters in `numpy.cov` that are either redundant such as `bias`, or have simple workarounds such as `y` and `rowvar`.
cc PandaBoi
TODO
- [x] Improve documentation
Pull Request resolved: https://github.com/pytorch/pytorch/pull/58311
Reviewed By: mruberry
Differential Revision: D28994140
Pulled By: heitorschueroff
fbshipit-source-id: 1890166c0a9c01e0a536acd91571cd704d632f44
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/59710
This is the exact same PR as before.
The version that landed was actually outdated compared to the github PR and that's why it failed on master... Sorry for the noise.
Test Plan: Imported from OSS
Reviewed By: zou3519
Differential Revision: D28995764
Pulled By: albanD
fbshipit-source-id: 8f7ae3356a886d45787c5e6ca53a4e7b033e306e
Summary:
Fixes https://github.com/pytorch/pytorch/issues/35379
- Adds `retains_grad` attribute backed by cpp as a native function. The python bindings for the function are skipped to be consistent with `is_leaf`.
- Tried writing it without native function, but the jit test `test_tensor_properties` seems to require that it be a native function (or alternatively maybe it could also work if we manually add a prim implementation?).
- Python API now uses `retain_grad` implementation from cpp
Pull Request resolved: https://github.com/pytorch/pytorch/pull/59362
Reviewed By: jbschlosser
Differential Revision: D28969298
Pulled By: soulitzer
fbshipit-source-id: 335f2be50b9fb870cd35dc72f7dadd6c8666cc02
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/54987
Based off of ezyang (https://github.com/pytorch/pytorch/pull/44799) and bdhirsh (https://github.com/pytorch/pytorch/pull/43702) 's prototype:
Here's a summary of the changes in this PR:
This PR adds a new dispatch key called Conjugate. This enables us to make conjugate operation a view and leverage the specialized library functions that fast path with the hermitian operation (conj + transpose).
1. Conjugate operation will now return a view with conj bit (1) for complex tensors and returns self for non-complex tensors as before. This also means `torch.view_as_real` will no longer be a view on conjugated complex tensors and is hence disabled. To fill the gap, we have added `torch.view_as_real_physical` which would return the real tensor agnostic of the conjugate bit on the input complex tensor. The information about conjugation on the old tensor can be obtained by calling `.is_conj()` on the new tensor.
2. NEW API:
a) `.conj()` -- now returning a view.
b) `.conj_physical()` -- does the physical conjugate operation. If the conj bit for input was set, you'd get `self.clone()`, else you'll get a new tensor with conjugated value in its memory.
c) `.conj_physical_()`, and `out=` variant
d) `.resolve_conj()` -- materializes the conjugation. returns self if the conj bit is unset, else returns a new tensor with conjugated values and conj bit set to 0.
e) `.resolve_conj_()` in-place version of (d)
f) `view_as_real_physical` -- as described in (1), it's functionally same as `view_as_real`, just that it doesn't error out on conjugated tensors.
g) `view_as_real` -- existing function, but now errors out on conjugated tensors.
3. Conjugate Fallback
a) Vast majority of PyTorch functions would currently use this fallback when they are called on a conjugated tensor.
b) This fallback is well equipped to handle the following cases:
- functional operation e.g., `torch.sin(input)`
- Mutable inputs and in-place operations e.g., `tensor.add_(2)`
- out-of-place operation e.g., `torch.sin(input, out=out)`
- Tensorlist input args
- NOTE: Meta tensors don't work with conjugate fallback.
4. Autograd
a) `resolve_conj()` is an identity function w.r.t. autograd
b) Everything else works as expected.
5. Testing:
a) All method_tests run with conjugate view tensors.
b) OpInfo tests that run with conjugate views
- test_variant_consistency_eager/jit
- gradcheck, gradgradcheck
- test_conj_views (that only run for `torch.cfloat` dtype)
NOTE: functions like `empty_like`, `zero_like`, `randn_like`, `clone` don't propagate the conjugate bit.
Follow up work:
1. conjugate view RFC
2. Add neg bit to re-enable view operation on conjugated tensors
3. Update linalg functions to call into specialized functions that fast path with the hermitian operation.
Test Plan: Imported from OSS
Reviewed By: VitalyFedyunin
Differential Revision: D28227315
Pulled By: anjali411
fbshipit-source-id: acab9402b9d6a970c6d512809b627a290c8def5f
Summary:
Adds `is_inference` as a native function w/ manual cpp bindings.
Also changes instances of `is_inference_tensor` to `is_inference` to be consistent with other properties such as `is_complex`.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/58729
Reviewed By: mruberry
Differential Revision: D28874507
Pulled By: soulitzer
fbshipit-source-id: 0fa6bcdc72a4ae444705e2e0f3c416c1b28dadc7
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/56017Fixes#55686
This patch is seemingly straightforward but some of the changes are very
subtle. For the general algorithmic approach, please first read the
quoted issue. Based on the algorithm, there are some fairly
straightforward changes:
- New boolean on TensorImpl tracking if we own the pyobj or not
- PythonHooks virtual interface for requesting deallocation of pyobj
when TensorImpl is being released and we own its pyobj, and
implementation of the hooks in python_tensor.cpp
- Modification of THPVariable to MaybeOwned its C++ tensor, directly
using swolchok's nice new class
And then, there is python_variable.cpp. Some of the changes follow the
general algorithmic approach:
- THPVariable_NewWithVar is simply adjusted to handle MaybeOwned and
initializes as owend (like before)
- THPVariable_Wrap adds the logic for reverting ownership back to
PyObject when we take out an owning reference to the Python object
- THPVariable_dealloc attempts to resurrect the Python object if
the C++ tensor is live, and otherwise does the same old implementation
as before
- THPVariable_tryResurrect implements the resurrection logic. It is
modeled after CPython code so read the cited logic and see if
it is faithfully replicated
- THPVariable_clear is slightly updated for MaybeOwned and also to
preserve the invariant that if owns_pyobj, then pyobj_ is not null.
This change is slightly dodgy: the previous implementation has a
comment mentioning that the pyobj nulling is required to ensure we
don't try to reuse the dead pyobj. I don't think, in this new world,
this is possible, because the invariant says that the pyobj only
dies if the C++ object is dead too. But I still unset the field
for safety.
And then... there is THPVariableMetaType. colesbury explained in the
issue why this is necessary: when destructing an object in Python, you
start off by running the tp_dealloc of the subclass before moving up
to the parent class (much in the same way C++ destructors work). The
deallocation process for a vanilla Python-defined class does irreparable
harm to the PyObject instance (e.g., the finalizers get run) making it
no longer valid attempt to resurrect later in the tp_dealloc chain.
(BTW, the fact that objects can resurrect but in an invalid state is
one of the reasons why it's so frickin' hard to write correct __del__
implementations). So we need to make sure that we actually override
the tp_dealloc of the bottom most *subclass* of Tensor to make sure
we attempt a resurrection before we start finalizing. To do this,
we need to define a metaclass for Tensor that can override tp_dealloc
whenever we create a new subclass of Tensor. By the way, it was totally
not documented how to create metaclasses in the C++ API, and it took
a good bit of trial error to figure it out (and the answer is now
immortalized in https://stackoverflow.com/q/67077317/23845 -- the things
that I got wrong in earlier versions of the PR included setting
tp_basicsize incorrectly, incorrectly setting Py_TPFLAGS_HAVE_GC on
the metaclass--you want to leave it unset so that it inherits, and
determining that tp_init is what actually gets called when you construct
a class, not tp_call as another not-to-be-named StackOverflow question
suggests).
Aside: Ordinarily, adding a metaclass to a class is a user visible
change, as it means that it is no longer valid to mixin another class
with a different metaclass. However, because _C._TensorBase is a C
extension object, it will typically conflict with most other
metaclasses, so this is not BC breaking.
The desired new behavior of a subclass tp_dealloc is to first test if
we should resurrect, and otherwise do the same old behavior. In an
initial implementation of this patch, I implemented this by saving the
original tp_dealloc (which references subtype_dealloc, the "standard"
dealloc for all Python defined classes) and invoking it. However, this
results in an infinite loop, as it attempts to call the dealloc function
of the base type, but incorrectly chooses subclass type (because it is
not a subtype_dealloc, as we have overridden it; see
b38601d496/Objects/typeobject.c (L1261) )
So, with great reluctance, I must duplicate the behavior of
subtype_dealloc in our implementation. Note that this is not entirely
unheard of in Python binding code; for example, Cython
c25c3ccc4b/Cython/Compiler/ModuleNode.py (L1560)
also does similar things. This logic makes up the bulk of
THPVariable_subclass_dealloc
To review this, you should pull up the CPython copy of subtype_dealloc
b38601d496/Objects/typeobject.c (L1230)
and verify that I have specialized the implementation for our case
appropriately. Among the simplifications I made:
- I assume PyType_IS_GC, because I assume that Tensor subclasses are
only ever done in Python and those classes are always subject to GC.
(BTW, yes! This means I have broken anyone who has extend PyTorch
tensor from C API directly. I'm going to guess no one has actually
done this.)
- I don't bother walking up the type bases to find the parent dealloc;
I know it is always THPVariable_dealloc. Similarly, I can get rid
of some parent type tests based on knowledge of how
THPVariable_dealloc is defined
- The CPython version calls some private APIs which I can't call, so
I use the public PyObject_GC_UnTrack APIs.
- I don't allow the finalizer of a Tensor to change its type (but
more on this shortly)
One alternative I discussed with colesbury was instead of copy pasting
the subtype_dealloc, we could transmute the type of the object that was
dying to turn it into a different object whose tp_dealloc is
subtype_dealloc, so the stock subtype_dealloc would then be applicable.
We decided this would be kind of weird and didn't do it that way.
TODO:
- More code comments
- Figure out how not to increase the size of TensorImpl with the new
bool field
- Add some torture tests for the THPVariable_subclass_dealloc, e.g.,
involving subclasses of Tensors that do strange things with finalizers
- Benchmark the impact of taking the GIL to release C++ side tensors
(e.g., from autograd)
- Benchmark the impact of adding a new metaclass to Tensor (probably
will be done by separating out the metaclass change into its own
change)
- Benchmark the impact of changing THPVariable to conditionally own
Tensor (as opposed to unconditionally owning it, as before)
- Add tests that this actually indeed preserves the Python object
Signed-off-by: Edward Z. Yang <ezyang@fb.com>
Test Plan: Imported from OSS
Reviewed By: albanD
Differential Revision: D27765125
Pulled By: ezyang
fbshipit-source-id: 857f14bdcca2900727412aff4c2e2d7f0af1415a
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/57386
Here is the PR for what's discussed in the RFC https://github.com/pytorch/pytorch/issues/55374 to enable the autocast for CPU device. Currently, this PR only enable BF16 as the lower precision datatype.
Changes:
1. Enable new API `torch.cpu.amp.autocast` for autocast on CPU device: include the python API, C++ API, new Dispatchkey etc.
2. Consolidate the implementation for each cast policy sharing between CPU and GPU devices.
3. Add the operation lists to corresponding cast policy for cpu autocast.
Test Plan: Imported from OSS
Reviewed By: soulitzer
Differential Revision: D28572219
Pulled By: ezyang
fbshipit-source-id: db3db509973b16a5728ee510b5e1ee716b03a152
Summary:
This adds the methods `Tensor.cfloat()` and `Tensor.cdouble()`.
I was not able to find the tests for `.float()` functions. I'd be happy to add similar tests for these functions once someone points me to them.
Fixes https://github.com/pytorch/pytorch/issues/56014
Pull Request resolved: https://github.com/pytorch/pytorch/pull/58137
Reviewed By: ejguan
Differential Revision: D28412288
Pulled By: anjali411
fbshipit-source-id: ff3653cb3516bcb3d26a97b9ec3d314f1f42f83d
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/58039
The new function has the following signature
`inv_ex(Tensor inpit, *, bool check_errors=False) -> (Tensor inverse, Tensor info)`.
When `check_errors=True`, an error is thrown if the matrix is not invertible; `check_errors=False` - responsibility for checking the result is on the user.
`linalg_inv` is implemented using calls to `linalg_inv_ex` now.
Resolves https://github.com/pytorch/pytorch/issues/25095
Test Plan: Imported from OSS
Reviewed By: ngimel
Differential Revision: D28405148
Pulled By: mruberry
fbshipit-source-id: b8563a6c59048cb81e206932eb2f6cf489fd8531
Summary:
Fixes https://github.com/pytorch/pytorch/issues/56608
- Adds binding to the `c10::InferenceMode` RAII class in `torch._C._autograd.InferenceMode` through pybind. Also binds the `torch.is_inference_mode` function.
- Adds context manager `torch.inference_mode` to manage an instance of `c10::InferenceMode` (global). Implemented in `torch.autograd.grad_mode.py` to reuse the `_DecoratorContextManager` class.
- Adds some tests based on those linked in the issue + several more for just the context manager
Issues/todos (not necessarily for this PR):
- Improve short inference mode description
- Small example
- Improved testing since there is no direct way of checking TLS/dispatch keys
-
Pull Request resolved: https://github.com/pytorch/pytorch/pull/58045
Reviewed By: agolynski
Differential Revision: D28390595
Pulled By: soulitzer
fbshipit-source-id: ae98fa036c6a2cf7f56e0fd4c352ff804904752c
Summary:
Backward methods for `torch.lu` and `torch.lu_solve` require the `torch.lu_unpack` method.
However, while `torch.lu` is a Python wrapper over a native function, so its gradient is implemented via `autograd.Function`,
`torch.lu_solve` is a native function, so it cannot access `torch.lu_unpack` as it is implemented in Python.
Hence this PR presents a native (ATen) `lu_unpack` version. It is also possible to update the gradients for `torch.lu` so that backward+JIT is supported (no JIT for `autograd.Function`) with this function.
~~The interface for this method is different from the original `torch.lu_unpack`, so it is decided to keep it hidden.~~
Pull Request resolved: https://github.com/pytorch/pytorch/pull/46913
Reviewed By: albanD
Differential Revision: D28355725
Pulled By: mruberry
fbshipit-source-id: 281260f3b6e93c15b08b2ba66d5a221314b00e78
Summary:
Backward methods for `torch.lu` and `torch.lu_solve` require the `torch.lu_unpack` method.
However, while `torch.lu` is a Python wrapper over a native function, so its gradient is implemented via `autograd.Function`,
`torch.lu_solve` is a native function, so it cannot access `torch.lu_unpack` as it is implemented in Python.
Hence this PR presents a native (ATen) `lu_unpack` version. It is also possible to update the gradients for `torch.lu` so that backward+JIT is supported (no JIT for `autograd.Function`) with this function.
~~The interface for this method is different from the original `torch.lu_unpack`, so it is decided to keep it hidden.~~
Pull Request resolved: https://github.com/pytorch/pytorch/pull/46913
Reviewed By: astaff
Differential Revision: D28117714
Pulled By: mruberry
fbshipit-source-id: befd33db12ecc147afacac792418b6f4948fa4a4
Summary:
This PR is focused on the API for `linalg.matrix_norm` and delegates computations to `linalg.norm` for the moment.
The main difference between the norms is when `dim=None`. In this case
- `linalg.norm` will compute a vector norm on the flattened input if `ord=None`, otherwise it requires the input to be either 1D or 2D in order to disambiguate between vector and matrix norm
- `linalg.vector_norm` will flatten the input
- `linalg.matrix_norm` will compute the norm over the last two dimensions, treating the input as batch of matrices
In future PRs, the computations will be moved to `torch.linalg.matrix_norm` and `torch.norm` and `torch.linalg.norm` will delegate computations to either `linalg.vector_norm` or `linalg.matrix_norm` based on the arguments provided.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/57127
Reviewed By: mrshenli
Differential Revision: D28186736
Pulled By: mruberry
fbshipit-source-id: 99ce2da9d1c4df3d9dd82c0a312c9570da5caf25
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/57180
We have now a separate function for computing only the singular values.
`compute_uv` argument is not needed and it was decided in the
offline discussion to remove it. This is a BC-breaking change but our
linalg module is beta, therefore we can do it without a deprecation
notice.
Test Plan: Imported from OSS
Reviewed By: ngimel
Differential Revision: D28142163
Pulled By: mruberry
fbshipit-source-id: 3fac1fcae414307ad5748c9d5ff50e0aa4e1b853
Summary:
As per discussion here https://github.com/pytorch/pytorch/pull/57127#discussion_r624948215
Note that we cannot remove the optional type from the `dim` parameter because the default is to flatten the input tensor which cannot be easily captured by a value other than `None`
### BC Breaking Note
This PR changes the `ord` parameter of `torch.linalg.vector_norm` so that it no longer accepts `None` arguments. The default behavior of `2` is equivalent to the previous default of `None`.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/57662
Reviewed By: albanD, mruberry
Differential Revision: D28228870
Pulled By: heitorschueroff
fbshipit-source-id: 040fd8055bbe013f64d3c8409bbb4b2c87c99d13
Summary:
The new function has the following signature `cholesky_ex(Tensor input, *, bool check_errors=False) -> (Tensor L, Tensor infos)`. When `check_errors=True`, an error is thrown if the decomposition fails; `check_errors=False` - responsibility for checking the decomposition is on the user.
When `check_errors=False`, we don't have host-device memory transfers for checking the values of the `info` tensor.
Rewrote the internal code for `torch.linalg.cholesky`. Added `cholesky_stub` dispatch. `linalg_cholesky` is implemented using calls to `linalg_cholesky_ex` now.
Resolves https://github.com/pytorch/pytorch/issues/57032.
Ref. https://github.com/pytorch/pytorch/issues/34272, https://github.com/pytorch/pytorch/issues/47608, https://github.com/pytorch/pytorch/issues/47953
Pull Request resolved: https://github.com/pytorch/pytorch/pull/56724
Reviewed By: ngimel
Differential Revision: D27960176
Pulled By: mruberry
fbshipit-source-id: f05f3d5d9b4aa444e41c4eec48ad9a9b6fd5dfa5
Summary:
Fixes https://github.com/pytorch/pytorch/issues/53964. cc albanD almson
## Major changes:
- Overhauled the actual loss calculation so that the shapes are now correct (in functional.py)
- added the missing doc in nn.functional.rst
## Minor changes (in functional.py):
- I removed the previous check on whether input and target were the same shape. This is to allow for broadcasting, say when you have 10 predictions that all have the same target.
- I added some comments to explain each shape check in detail. Let me know if these should be shortened/cut.
Screenshots of updated docs attached.
Let me know what you think, thanks!
## Edit: Description of change of behaviour (affecting BC):
The backwards-compatibility is only affected for the `reduction='none'` mode. This was the source of the bug. For tensors with size (N, D), the old returned loss had size (N), as incorrect summation was happening. It will now have size (N, D) as expected.
### Example
Define input tensors, all with size (2, 3).
`input = torch.tensor([[0., 1., 3.], [2., 4., 0.]], requires_grad=True)`
`target = torch.tensor([[1., 4., 2.], [-1., 2., 3.]])`
`var = 2*torch.ones(size=(2, 3), requires_grad=True)`
Initialise loss with reduction mode 'none'. We expect the returned loss to have the same size as the input tensors, (2, 3).
`loss = torch.nn.GaussianNLLLoss(reduction='none')`
Old behaviour:
`print(loss(input, target, var)) `
`# Gives tensor([3.7897, 6.5397], grad_fn=<MulBackward0>. This has size (2).`
New behaviour:
`print(loss(input, target, var)) `
`# Gives tensor([[0.5966, 2.5966, 0.5966], [2.5966, 1.3466, 2.5966]], grad_fn=<MulBackward0>)`
`# This has the expected size, (2, 3).`
To recover the old behaviour, sum along all dimensions except for the 0th:
`print(loss(input, target, var).sum(dim=1))`
`# Gives tensor([3.7897, 6.5397], grad_fn=<SumBackward1>.`
Pull Request resolved: https://github.com/pytorch/pytorch/pull/56469
Reviewed By: jbschlosser, agolynski
Differential Revision: D27894170
Pulled By: albanD
fbshipit-source-id: 197890189c97c22109491c47f469336b5b03a23f
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:
Reference: https://github.com/pytorch/pytorch/issues/50345
Changes:
* Add `i0e`
* Move some kernels from `UnaryOpsKernel.cu` to `UnarySpecialOpsKernel.cu` to decrease compilation time per file.
Time taken by i0e_vs_scipy tests: around 6.33.s
<details>
<summary>Test Run Log</summary>
```
(pytorch-cuda-dev) kshiteej@qgpu1:~/Pytorch/pytorch_module_special$ pytest test/test_unary_ufuncs.py -k _i0e_vs
======================================================================= test session starts ========================================================================
platform linux -- Python 3.8.6, pytest-6.1.2, py-1.9.0, pluggy-0.13.1
rootdir: /home/kshiteej/Pytorch/pytorch_module_special, configfile: pytest.ini
plugins: hypothesis-5.38.1
collected 8843 items / 8833 deselected / 10 selected
test/test_unary_ufuncs.py ...sss.... [100%]
========================================================================= warnings summary =========================================================================
../../.conda/envs/pytorch-cuda-dev/lib/python3.8/site-packages/torch/backends/cudnn/__init__.py:73
test/test_unary_ufuncs.py::TestUnaryUfuncsCUDA::test_special_i0e_vs_scipy_cuda_bfloat16
/home/kshiteej/.conda/envs/pytorch-cuda-dev/lib/python3.8/site-packages/torch/backends/cudnn/__init__.py:73: UserWarning: PyTorch was compiled without cuDNN/MIOpen support. To use cuDNN/MIOpen, rebuild PyTorch making sure the library is visible to the build system.
warnings.warn(
-- Docs: https://docs.pytest.org/en/stable/warnings.html
===================================================================== short test summary info ======================================================================
SKIPPED [3] test/test_unary_ufuncs.py:1182: not implemented: Could not run 'aten::_copy_from' with arguments from the 'Meta' backend. This could be because the operator doesn't exist for this backend, or was omitted during the selective/custom build process (if using custom build). If you are a Facebook employee using PyTorch on mobile, please visit https://fburl.com/ptmfixes for possible resolutions. 'aten::_copy_from' is only available for these backends: [BackendSelect, Named, InplaceOrView, AutogradOther, AutogradCPU, AutogradCUDA, AutogradXLA, UNKNOWN_TENSOR_TYPE_ID, AutogradMLC, AutogradNestedTensor, AutogradPrivateUse1, AutogradPrivateUse2, AutogradPrivateUse3, Tracer, Autocast, Batched, VmapMode].
BackendSelect: fallthrough registered at ../aten/src/ATen/core/BackendSelectFallbackKernel.cpp:3 [backend fallback]
Named: registered at ../aten/src/ATen/core/NamedRegistrations.cpp:7 [backend fallback]
InplaceOrView: fallthrough registered at ../aten/src/ATen/core/VariableFallbackKernel.cpp:56 [backend fallback]
AutogradOther: registered at ../torch/csrc/autograd/generated/VariableType_4.cpp:8761 [autograd kernel]
AutogradCPU: registered at ../torch/csrc/autograd/generated/VariableType_4.cpp:8761 [autograd kernel]
AutogradCUDA: registered at ../torch/csrc/autograd/generated/VariableType_4.cpp:8761 [autograd kernel]
AutogradXLA: registered at ../torch/csrc/autograd/generated/VariableType_4.cpp:8761 [autograd kernel]
UNKNOWN_TENSOR_TYPE_ID: registered at ../torch/csrc/autograd/generated/VariableType_4.cpp:8761 [autograd kernel]
AutogradMLC: registered at ../torch/csrc/autograd/generated/VariableType_4.cpp:8761 [autograd kernel]
AutogradNestedTensor: registered at ../torch/csrc/autograd/generated/VariableType_4.cpp:8761 [autograd kernel]
AutogradPrivateUse1: registered at ../torch/csrc/autograd/generated/VariableType_4.cpp:8761 [autograd kernel]
AutogradPrivateUse2: registered at ../torch/csrc/autograd/generated/VariableType_4.cpp:8761 [autograd kernel]
AutogradPrivateUse3: registered at ../torch/csrc/autograd/generated/VariableType_4.cpp:8761 [autograd kernel]
Tracer: registered at ../torch/csrc/autograd/generated/TraceType_4.cpp:9348 [kernel]
Autocast: fallthrough registered at ../aten/src/ATen/autocast_mode.cpp:250 [backend fallback]
Batched: registered at ../aten/src/ATen/BatchingRegistrations.cpp:1016 [backend fallback]
VmapMode: fallthrough registered at ../aten/src/ATen/VmapModeRegistrations.cpp:33 [backend fallback]
==================================================== 7 passed, 3 skipped, 8833 deselected, 2 warnings in 6.33s =====================================================
```
</details>
TODO:
* [x] Check rendered docs (https://11743402-65600975-gh.circle-artifacts.com/0/docs/special.html)
Pull Request resolved: https://github.com/pytorch/pytorch/pull/54409
Reviewed By: jbschlosser
Differential Revision: D27760472
Pulled By: mruberry
fbshipit-source-id: bdfbcaa798b00c51dc9513c34626246c8fc10548
Summary:
This PR adds a `padding_idx` parameter to `nn.EmbeddingBag` and `nn.functional.embedding_bag`. As with `nn.Embedding`'s `padding_idx` argument, if an embedding's index is equal to `padding_idx` it is ignored, so it is not included in the reduction.
This PR does not add support for `padding_idx` for quantized or ONNX `EmbeddingBag` for opset10/11 (opset9 is supported). In these cases, an error is thrown if `padding_idx` is provided.
Fixes https://github.com/pytorch/pytorch/issues/3194
Pull Request resolved: https://github.com/pytorch/pytorch/pull/49237
Reviewed By: walterddr, VitalyFedyunin
Differential Revision: D26948258
Pulled By: jbschlosser
fbshipit-source-id: 3ca672f7e768941f3261ab405fc7597c97ce3dfc
Summary:
Reference: https://github.com/pytorch/pytorch/issues/50345
Chages:
* Alias for sigmoid and logit
* Adds out variant for C++ API
* Updates docs to link back to `special` documentation
Pull Request resolved: https://github.com/pytorch/pytorch/pull/54759
Reviewed By: mrshenli
Differential Revision: D27615208
Pulled By: mruberry
fbshipit-source-id: 8bba908d1bea246e4aa9dbadb6951339af353556
Summary:
This PR adds `torch.linalg.eig`, and `torch.linalg.eigvals` for NumPy compatibility.
MAGMA uses a hybrid CPU-GPU algorithm and doesn't have a GPU interface for the non-symmetric eigendecomposition. It means that it forces us to transfer inputs living in GPU memory to CPU first before calling MAGMA, and then transfer results from MAGMA to CPU. That is rather slow for smaller matrices and MAGMA is faster than CPU path only for matrices larger than 3000x3000.
Unfortunately, there is no cuSOLVER function for this operation.
Autograd support for `torch.linalg.eig` will be added in a follow-up PR.
Ref https://github.com/pytorch/pytorch/issues/42666
Pull Request resolved: https://github.com/pytorch/pytorch/pull/52491
Reviewed By: anjali411
Differential Revision: D27563616
Pulled By: mruberry
fbshipit-source-id: b42bb98afcd2ed7625d30bdd71cfc74a7ea57bb5
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/52859
This reverts commit 92a4ee1cf6.
Added support for bfloat16 for CUDA 11 and removed fast-path for empty input tensors that was affecting autograd graph.
Test Plan: Imported from OSS
Reviewed By: H-Huang
Differential Revision: D27402390
Pulled By: heitorschueroff
fbshipit-source-id: 73c5ccf54f3da3d29eb63c9ed3601e2fe6951034
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/54702
This fixes subclassing for __iter__ so that it returns an iterator over
subclasses properly instead of Tensor.
Test Plan: Imported from OSS
Reviewed By: H-Huang
Differential Revision: D27352563
Pulled By: ezyang
fbshipit-source-id: 4c195a86c8f2931a6276dc07b1e74ee72002107c
Summary:
Reference: https://github.com/pytorch/pytorch/issues/38349
Wrapper around the existing `torch.gather` with broadcasting logic.
TODO:
* [x] Add Doc entry (see if phrasing can be improved)
* [x] Add OpInfo
* [x] Add test against numpy
* [x] Handle broadcasting behaviour and when dim is not given.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/52833
Reviewed By: malfet
Differential Revision: D27319038
Pulled By: mruberry
fbshipit-source-id: 00f307825f92c679d96e264997aa5509172f5ed1
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/53727
This is first diff to add native support for segment reduction in PyTorch. It provides similar functionality like torch.scatter or "numpy.ufunc.reduceat".
This diff mainly focuses on API layer to make sure future improvements will not cause backward compatibility issues. Once API is settled, here are next steps I am planning:
- Add support for other major reduction types (e.g. min, sum) for 1D tensor
- Add Cuda support
- Backward support
- Documentation for the op
- Perf optimizations and benchmark util
- Support for multi dimensional tensors (on data and lengths) (not high priority)
- Support for 'indices' (not high priority)
Test Plan: Added unit test
Reviewed By: ngimel
Differential Revision: D26952075
fbshipit-source-id: 8040ec96def3013e7240cf675d499ee424437560
Summary:
This PR adds autograd support for `torch.orgqr`.
Since `torch.orgqr` is one of few functions that expose LAPACK's naming and all other linear algebra routines were renamed a long time ago, I also added a new function with a new name and `torch.orgqr` now is an alias for it.
The new proposed name is `householder_product`. For a matrix `input` and a vector `tau` LAPACK's orgqr operation takes columns of `input` (called Householder vectors or elementary reflectors) scalars of `tau` that together represent Householder matrices and then the product of these matrices is computed. See https://www.netlib.org/lapack/lug/node128.html.
Other linear algebra libraries that I'm aware of do not expose this LAPACK function, so there is some freedom in naming it. It is usually used internally only for QR decomposition, but can be useful for deep learning tasks now when it supports differentiation.
Resolves https://github.com/pytorch/pytorch/issues/50104
Pull Request resolved: https://github.com/pytorch/pytorch/pull/52637
Reviewed By: agolynski
Differential Revision: D27114246
Pulled By: mruberry
fbshipit-source-id: 9ab51efe52aec7c137aa018c7bd486297e4111ce
Summary:
Close https://github.com/pytorch/pytorch/issues/51108
Related https://github.com/pytorch/pytorch/issues/38349
This PR implements the `cpu_kernel_multiple_outputs` to support returning multiple values in a CPU kernel.
```c++
auto iter = at::TensorIteratorConfig()
.add_output(out1)
.add_output(out2)
.add_input(in1)
.add_input(in2)
.build();
at::native::cpu_kernel_multiple_outputs(iter,
[=](float a, float b) -> std::tuple<float, float> {
float add = a + b;
float mul = a * b;
return std::tuple<float, float>(add, mul);
}
);
```
The `out1` will equal to `torch.add(in1, in2)`, while the result of `out2` will be `torch.mul(in1, in2)`.
It helps developers implement new torch functions that return two tensors more conveniently, such as NumPy-like functions [divmod](https://numpy.org/doc/1.18/reference/generated/numpy.divmod.html?highlight=divmod#numpy.divmod) and [frexp](https://numpy.org/doc/stable/reference/generated/numpy.frexp.html#numpy.frexp).
This PR adds `torch.frexp` function to exercise the new functionality provided by `cpu_kernel_multiple_outputs`.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/51097
Reviewed By: albanD
Differential Revision: D26982619
Pulled By: heitorschueroff
fbshipit-source-id: cb61c7f2c79873ab72ab5a61cbdb9203531ad469
Summary:
Fixes https://github.com/pytorch/pytorch/issues/44378 by providing a wider range of drivers similar to what SciPy is doing.
The supported CPU drivers are `gels, gelsy, gelsd, gelss`.
The CUDA interface has only `gels` implemented but only for overdetermined systems.
The current state of this PR:
- [x] CPU interface
- [x] CUDA interface
- [x] CPU tests
- [x] CUDA tests
- [x] Memory-efficient batch-wise iteration with broadcasting which fixes https://github.com/pytorch/pytorch/issues/49252
- [x] docs
Pull Request resolved: https://github.com/pytorch/pytorch/pull/49093
Reviewed By: albanD
Differential Revision: D26991788
Pulled By: mruberry
fbshipit-source-id: 8af9ada979240b255402f55210c0af1cba6a0a3c
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:
per title
This PR did
- Migrate `apex.parallel.SyncBatchNorm` channels_last to pytorch `torch.nn.SyncBatchNorm`
- Fix a TODO here by fusing `sum`, `div` kernels into backward elementwise kernel
b167402e2e/torch/nn/modules/_functions.py (L76-L95)
Todo
- [x] Discuss a regression introduced in https://github.com/pytorch/pytorch/pull/37133#discussion_r512530389, which is the synchronized copy here
b167402e2e/torch/nn/modules/_functions.py (L32-L34)
**Comment**: This PR uses apex version for the size check. Test passed and I haven't seen anything wrong so far.
- [x] The restriction to use channels_last kernel will be like this
```
inline bool batch_norm_use_channels_last_kernels(const at::Tensor& self) {
return self.is_contiguous(at::MemoryFormat::ChannelsLast) || self.ndimension() == 2;
}
```
I think we can relax that for channels_last_3d as well?
**Comment**: we don't have benchmark for this now, will check this and add functionality later when needed.
- [x] Add test
- [x] Add benchmark
Detailed benchmark is at https://github.com/xwang233/code-snippet/tree/master/syncbn-channels-last
Close https://github.com/pytorch/pytorch/issues/50781
Pull Request resolved: https://github.com/pytorch/pytorch/pull/46906
Reviewed By: albanD
Differential Revision: D26771437
Pulled By: malfet
fbshipit-source-id: d00387044e9d43ac7e6c0e32a2db22c63d1504de
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/44378 by providing a wider range of drivers similar to what SciPy is doing.
The supported CPU drivers are `gels, gelsy, gelsd, gelss`.
The CUDA interface has only `gels` implemented but only for overdetermined systems.
The current state of this PR:
- [x] CPU interface
- [x] CUDA interface
- [x] CPU tests
- [x] CUDA tests
- [x] Memory-efficient batch-wise iteration with broadcasting which fixes https://github.com/pytorch/pytorch/issues/49252
- [x] docs
Pull Request resolved: https://github.com/pytorch/pytorch/pull/49093
Reviewed By: H-Huang
Differential Revision: D26723384
Pulled By: mruberry
fbshipit-source-id: c9866a95f14091955cf42de22f4ac9e2da009713
Summary:
Apple recently announced ML Compute, a new framework available in macOS Big Sur, which enables users to accelerate the training of neural networks on Mac hardware. This PR is the first on a series of PRs that will enable the integration with ML Compute. Most of the integration code will live on a separate subrepo named `mlc`.
The integration with `mlc` (ML Compute) will be very similar to that of xla. We rely on registering our ops through:
TORCH_LIBRARY_IMPL(aten, PrivateUse1, m) {
m.impl_UNBOXED(<op_schema_name>, &customized_op_kernel)
...
}
Pull Request resolved: https://github.com/pytorch/pytorch/pull/50634
Reviewed By: malfet
Differential Revision: D26614213
Pulled By: smessmer
fbshipit-source-id: 3b492b346c61cc3950ac880ac01a82fbdddbc07b
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/51807
Implemented torch.linalg.multi_dot similar to [numpy.linalg.multi_dot](https://numpy.org/doc/stable/reference/generated/numpy.linalg.multi_dot.html).
This function does not support broadcasting or batched inputs at the moment.
**NOTE**
numpy.linalg.multi_dot allows the first and last tensors to have more than 2 dimensions despite their docs stating these must be either 1D or 2D. This PR diverges from NumPy in that it enforces this restriction.
**TODO**
- [ ] Benchmark against NumPy
- [x] Add OpInfo testing
- [x] Remove unnecessary copy for out= argument
Test Plan: Imported from OSS
Reviewed By: nikithamalgifb
Differential Revision: D26375734
Pulled By: heitorschueroff
fbshipit-source-id: 839642692424c4b1783606c76dd5b34455368f0b
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/51878
`fake_quantize_per_tensor_affine_cachemask` and
`fake_quantize_per_channel_affine_cachemask` are implementation
details of `fake_quantize_per_tensor_affine` and
`fake_quantize_per_channel_affine`, removing the
Python bindings for them since there is no need to
expose them.
Test Plan:
```
python test/test_quantization.py TestFakeQuantize
```
Imported from OSS
Reviewed By: albanD, bugra
Differential Revision: D26314173
fbshipit-source-id: 733c93a3951453e739b6ed46b72fbad2244f6e97
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:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/51706
Pull Request resolved: https://github.com/pytorch/pytorch/pull/50280
As mentioned in gh-43874, this adds a `rounding_mode={'true', 'trunc', 'floor'}`
argument so `torch.div` can be used as a replacement for `floor_divide` during
the transitional period.
I've included dedicated kernels for truncated and floor division which
aren't strictly necessary for float, but do perform significantly better (~2x) than
doing true division followed by a separate rounding kernel.
Note: I introduce new overloads for `aten::div` instead of just adding a default
`rounding_mode` because various JIT passes rely on the exact operator schema.
Test Plan: Imported from OSS
Reviewed By: ngimel
Differential Revision: D26123271
Pulled By: mruberry
fbshipit-source-id: 51a83717602114597ec9c4d946e35a392eb01d46
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/51255
This is the same as #50561, but for per-channel fake_quant.
TODO before land write up better
Memory and performance impact (MobileNetV2): TODO
Performance impact (microbenchmarks): https://gist.github.com/vkuzo/fbe1968d2bbb79b3f6dd776309fbcffc
* forward pass on cpu: 512ms -> 750ms (+46%)
* forward pass on cuda: 99ms -> 128ms (+30%)
* note: the overall performance impact to training jobs should be minimal, because this is used for weights, and relative importance of fq is dominated by fq'ing the activations
* note: we can optimize the perf in a future PR by reading once and writing twice
Test Plan:
```
python test/test_quantization.py TestFakeQuantize.test_forward_per_channel_cachemask_cpu
python test/test_quantization.py TestFakeQuantize.test_forward_per_channel_cachemask_cuda
python test/test_quantization.py TestFakeQuantize.test_backward_per_channel_cachemask_cpu
python test/test_quantization.py TestFakeQuantize.test_backward_per_channel_cachemask_cuda
```
Imported from OSS
Reviewed By: jerryzh168
Differential Revision: D26117721
fbshipit-source-id: 798b59316dff8188a1d0948e69adf9e5509e414c
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/50561
Not for review yet, a bunch of TODOs need finalizing.
tl;dr; add an alternative implementation of `fake_quantize` which saves
a ask during the forward pass and uses it to calculate the backward.
There are two benefits:
1. the backward function no longer needs the input Tensor, and it can be
gc'ed earlier by autograd. On MobileNetV2, this reduces QAT overhead
by ~15% (TODO: link, and absolute numbers). We add an additional mask Tensor
to pass around, but its size is 4x smaller than the input tensor. A
future optimization would be to pack the mask bitwise and unpack in the
backward.
2. the computation of `qval` can be done only once in the forward and
reused in the backward. No perf change observed, TODO verify with better
matrics.
TODO: describe in more detail
Test Plan:
OSS / torchvision / MobileNetV2
```
python references/classification/train_quantization.py
--print-freq 1
--data-path /data/local/packages/ai-group.imagenet-256-smallest-side/prod/
--output-dir ~/nfs/pytorch_vision_tests/
--backend qnnpack
--epochs 5
TODO paste results here
```
TODO more
Imported from OSS
Reviewed By: ngimel
Differential Revision: D25918519
fbshipit-source-id: ec544ca063f984de0f765bf833f205c99d6c18b6
Summary:
Add a new device type 'XPU' ('xpu' for lower case) to PyTorch. Changes are needed for code related to device model and kernel dispatch, e.g. DeviceType, Backend and DispatchKey etc.
https://github.com/pytorch/pytorch/issues/48246
Pull Request resolved: https://github.com/pytorch/pytorch/pull/49786
Reviewed By: mrshenli
Differential Revision: D25893962
Pulled By: ezyang
fbshipit-source-id: 7ff0a316ee34cf0ed6fc7ead08ecdeb7df4b0052
Summary:
This PR adds `torch.linalg.slogdet`.
Changes compared to the original torch.slogdet:
- Complex input now works as in NumPy
- Added out= variant (allocates temporary and makes a copy for now)
- Updated `slogdet_backward` to work with complex input
Ref. https://github.com/pytorch/pytorch/issues/42666
Pull Request resolved: https://github.com/pytorch/pytorch/pull/49194
Reviewed By: VitalyFedyunin
Differential Revision: D25916959
Pulled By: mruberry
fbshipit-source-id: cf9be8c5c044870200dcce38be48cd0d10e61a48
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/49502
It broke the OSS CI the last time I landed it, mostly cuda tests and python bindings.
Similar to permute_out, add the out variant of `aten::narrow` (slice in c2) which does an actual copy. `aten::narrow` creates a view, however, an copy is incurred when we call `input.contiguous` in the ops that follow `aten::narrow`, in `concat_add_mul_replacenan_clip`, `casted_batch_one_hot_lengths`, and `batch_box_cox`.
{F351263599}
Test Plan:
Unit test:
```
buck test //caffe2/aten:math_kernel_test
buck test //caffe2/test:sparse -- test_narrow
```
Benchmark with the adindexer model:
```
bs = 1 is neutral
Before:
I1214 21:32:51.919239 3285258 PyTorchPredictorBenchLib.cpp:209] PyTorch run finished. Milliseconds per iter: 0.0886948. Iters per second: 11274.6
After:
I1214 21:32:52.492352 3285277 PyTorchPredictorBenchLib.cpp:209] PyTorch run finished. Milliseconds per iter: 0.0888019. Iters per second: 11261
bs = 20 shows more gains probably because the tensors are bigger and therefore the cost of copying is higher
Before:
I1214 21:20:19.702445 3227229 PyTorchPredictorBenchLib.cpp:209] PyTorch run finished. Milliseconds per iter: 0.527563. Iters per second: 1895.51
After:
I1214 21:20:20.370173 3227307 PyTorchPredictorBenchLib.cpp:209] PyTorch run finished. Milliseconds per iter: 0.508734. Iters per second: 1965.67
```
Reviewed By: ajyu
Differential Revision: D25596290
fbshipit-source-id: da2f5a78a763895f2518c6298778ccc4d569462c
Summary:
This PR adds `torch.linalg.pinv`.
Changes compared to the original `torch.pinverse`:
* New kwarg "hermitian": with `hermitian=True` eigendecomposition is used instead of singular value decomposition.
* `rcond` argument can now be a `Tensor` of appropriate shape to apply matrix-wise clipping of singular values.
* Added `out=` variant (allocates temporary and makes a copy for now)
Ref. https://github.com/pytorch/pytorch/issues/42666
Pull Request resolved: https://github.com/pytorch/pytorch/pull/48399
Reviewed By: zhangguanheng66
Differential Revision: D25869572
Pulled By: mruberry
fbshipit-source-id: 0f330a91d24ba4e4375f648a448b27594e00dead
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/48965
This PR pulls `__torch_function__` checking entirely into C++, and adds a special `object_has_torch_function` method for ops which only have one arg as this lets us skip tuple construction and unpacking. We can now also do away with the Python side fast bailout for `Tensor` (e.g. `if any(type(t) is not Tensor for t in tensors) and has_torch_function(tensors)`) because they're actually slower than checking with the Python C API.
Test Plan: Existing unit tests. Benchmarks are in #48966
Reviewed By: ezyang
Differential Revision: D25590732
Pulled By: robieta
fbshipit-source-id: 6bd74788f06cdd673f3a2db898143d18c577eb42
Summary:
This PR adds `torch.linalg.inv` for NumPy compatibility.
`linalg_inv_out` uses in-place operations on provided `result` tensor.
I modified `apply_inverse` to accept tensor of Int instead of std::vector, that way we can write a function similar to `linalg_inv_out` but removing the error checks and device memory synchronization.
I fixed `lda` (leading dimension parameter which is max(1, n)) in many places to handle 0x0 matrices correctly.
Zero batch dimensions are also working and tested.
Ref https://github.com/pytorch/pytorch/issues/42666
Pull Request resolved: https://github.com/pytorch/pytorch/pull/48261
Reviewed By: gchanan
Differential Revision: D25849590
Pulled By: mruberry
fbshipit-source-id: cfee6f1daf7daccbe4612ec68f94db328f327651
Summary:
This is related to https://github.com/pytorch/pytorch/issues/42666 .
I am opening this PR to have the opportunity to discuss things.
First, we need to consider the differences between `torch.svd` and `numpy.linalg.svd`:
1. `torch.svd` takes `some=True`, while `numpy.linalg.svd` takes `full_matrices=True`, which is effectively the opposite (and with the opposite default, too!)
2. `torch.svd` returns `(U, S, V)`, while `numpy.linalg.svd` returns `(U, S, VT)` (i.e., V transposed).
3. `torch.svd` always returns a 3-tuple; `numpy.linalg.svd` returns only `S` in case `compute_uv==False`
4. `numpy.linalg.svd` also takes an optional `hermitian=False` argument.
I think that the plan is to eventually deprecate `torch.svd` in favor of `torch.linalg.svd`, so this PR does the following:
1. Rename/adapt the old `svd` C++ functions into `linalg_svd`: in particular, now `linalg_svd` takes `full_matrices` and returns `VT`
2. Re-implement the old C++ interface on top of the new (by negating `full_matrices` and transposing `VT`).
3. The C++ version of `linalg_svd` *always* returns a 3-tuple (we can't do anything else). So, there is a python wrapper which manually calls `torch._C._linalg.linalg_svd` to tweak the return value in case `compute_uv==False`.
Currently, `linalg_svd_backward` is broken because it has not been adapted yet after the `V ==> VT` change, but before continuing and spending more time on it I wanted to make sure that the general approach is fine.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/45562
Reviewed By: H-Huang
Differential Revision: D25803557
Pulled By: mruberry
fbshipit-source-id: 4966f314a0ba2ee391bab5cda4563e16275ce91f
Summary:
I am opening this PR early to have a place to discuss design issues.
The biggest difference between `torch.qr` and `numpy.linalg.qr` is that the former `torch.qr` takes a boolean parameter `some=True`, while the latter takes a string parameter `mode='reduced'` which can be one of the following:
`reduced`
this is completely equivalent to `some=True`, and both are the default.
`complete`
this is completely equivalent to `some=False`.
`r`
this returns only `r` instead of a tuple `(r, q)`. We have already decided that we don't want different return types depending on the parameters, so I propose to return `(r, empty_tensor)` instead. I **think** that in this mode it will be impossible to implement the backward pass, so we should raise an appropriate error in that case.
`raw`
in this mode, it returns `(h, tau)` instead of `(q, r)`. Internally, `h` and `tau` are obtained by calling lapack's `dgeqrf` and are later used to compute the actual values of `(q, r)`. The numpy docs suggest that these might be useful to call other lapack functions, but at the moment none of them is exposed by numpy and I don't know how often it is used in the real world.
I suppose the implementing the backward pass need attention to: the most straightforward solution is to use `(h, tau)` to compute `(q, r)` and then use the normal logic for `qr_backward`, but there might be faster alternatives.
`full`, `f`
alias for `reduced`, deprecated since numpy 1.8.0
`economic`, `e`
similar to `raw but it returns only `h` instead of `(h, tau). Deprecated since numpy 1.8.0
To summarize:
* `reduce`, `complete` and `r` are straightforward to implement.
* `raw` needs a bit of extra care, but I don't know how much high priority it is: since it is used rarely, we might want to not support it right now and maybe implement it in the future?
* I think we should just leave `full` and `economic` out, and possibly add a note to the docs explaining what you need to use instead
/cc mruberry
Pull Request resolved: https://github.com/pytorch/pytorch/pull/47764
Reviewed By: ngimel
Differential Revision: D25708870
Pulled By: mruberry
fbshipit-source-id: c25c70a23a02ec4322430d636542041e766ebe1b
Summary:
This PR adds `torch.linalg.inv` for NumPy compatibility.
`linalg_inv_out` uses in-place operations on provided `result` tensor.
I modified `apply_inverse` to accept tensor of Int instead of std::vector, that way we can write a function similar to `linalg_inv_out` but removing the error checks and device memory synchronization.
I fixed `lda` (leading dimension parameter which is max(1, n)) in many places to handle 0x0 matrices correctly.
Zero batch dimensions are also working and tested.
Ref https://github.com/pytorch/pytorch/issues/42666
Pull Request resolved: https://github.com/pytorch/pytorch/pull/48261
Reviewed By: ngimel
Differential Revision: D25690129
Pulled By: mruberry
fbshipit-source-id: edb2d03721f22168c42ded8458513cb23dfdc712
Summary:
Fixes https://github.com/pytorch/pytorch/issues/49214
**BC-Breaking**
Before this PR, `%=` didn't actually do the operation inplace and returned a new tensor.
After this PR, `%=` operation is actually inplace and the modified input tensor is returned.
Before PR,
```python
>>> import torch
>>> a = torch.tensor([11,12,13])
>>> id(a)
139627966219328
>>> a %= 10
>>> id(a)
139627966219264
```
After PR,
```python
>>> import torch
>>> a = torch.tensor([11,12,13])
>>> id(a)
139804702425280
>>> a %= 10
>>> id(a)
139804702425280
```
Pull Request resolved: https://github.com/pytorch/pytorch/pull/49390
Reviewed By: izdeby
Differential Revision: D25560423
Pulled By: zou3519
fbshipit-source-id: 2b92bfda260582aa4ac22c4025376295e51f854e
Summary:
Related https://github.com/pytorch/pytorch/issues/38349
Implement NumPy-like function `torch.broadcast_to` to broadcast the input tensor to a new shape.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/48997
Reviewed By: anjali411, ngimel
Differential Revision: D25663937
Pulled By: mruberry
fbshipit-source-id: 0415c03f92f02684983f412666d0a44515b99373
Summary:
This PR adds `torch.linalg.solve`.
`linalg_solve_out` uses in-place operations on the provided result tensor.
I modified `apply_solve` to accept tensor of Int instead of std::vector, that way we can write a function similar to `linalg_solve_out` but removing the error checks and device memory synchronization.
In comparison to `torch.solve` this routine accepts 1-dimensional tensors and batches of 1-dim tensors for the right-hand-side term. `torch.solve` requires it to be at least 2-dimensional.
Ref. https://github.com/pytorch/pytorch/issues/42666
Pull Request resolved: https://github.com/pytorch/pytorch/pull/48456
Reviewed By: izdeby
Differential Revision: D25562222
Pulled By: mruberry
fbshipit-source-id: a9355c029e2442c2e448b6309511919631f9e43b
Summary:
This PR is to change the `aten::native_layer_norm` and `aten::native_layer_norm_backward` signature to match `torch.layer_norm` definition. The current definition doesn't provide enough information to the PyTorch JIT to fuse layer_norm during training.
`native_layer_norm(X, gamma, beta, M, N, eps)` =>
`native_layer_norm(input, normalized_shape, weight, bias, eps)`
`native_layer_norm_backward(dY, X, mean, rstd, gamma, M, N, grad_input_mask)` =>
`native_layer_norm_backward(dY, input, normalized_shape, mean, rstd, weight, bias, grad_input_mask)`
Pull Request resolved: https://github.com/pytorch/pytorch/pull/48971
Reviewed By: izdeby
Differential Revision: D25574070
Pulled By: ngimel
fbshipit-source-id: 23e2804295a95bda3f1ca6b41a1e4c5a3d4d31b4
Summary:
Ref https://github.com/pytorch/pytorch/issues/42175
This removes the 4 deprecated spectral functions: `torch.{fft,rfft,ifft,irfft}`. `torch.fft` is also now imported by by default.
The actual `at::native` functions are still used in `torch.stft` so can't be full removed yet. But will once https://github.com/pytorch/pytorch/issues/47601 has been merged.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/48594
Reviewed By: heitorschueroff
Differential Revision: D25298929
Pulled By: mruberry
fbshipit-source-id: e36737fe8192fcd16f7e6310f8b49de478e63bf0
Summary:
Fixes https://github.com/pytorch/pytorch/issues/43837
This adds a `torch.broadcast_shapes()` function similar to Pyro's [broadcast_shape()](7c2c22c10d/pyro/distributions/util.py (L151)) and JAX's [lax.broadcast_shapes()](https://jax.readthedocs.io/en/test-docs/_modules/jax/lax/lax.html). This helper is useful e.g. in multivariate distributions that are parameterized by multiple tensors and we want to `torch.broadcast_tensors()` but the parameter tensors have different "event shape" (e.g. mean vectors and covariance matrices). This helper is already heavily used in Pyro's distribution codebase, and we would like to start using it in `torch.distributions`.
- [x] refactor `MultivariateNormal`'s expansion logic to use `torch.broadcast_shapes()`
- [x] add unit tests for `torch.broadcast_shapes()`
- [x] add docs
cc neerajprad
Pull Request resolved: https://github.com/pytorch/pytorch/pull/43935
Reviewed By: bdhirsh
Differential Revision: D25275213
Pulled By: neerajprad
fbshipit-source-id: 1011fdd597d0a7a4ef744ebc359bbb3c3be2aadc
Summary:
This PR adds `torch.linalg.matrix_rank`.
Changes compared to the original `torch.matrix_rank`:
- input with the complex dtype is supported
- batched input is supported
- "symmetric" kwarg renamed to "hermitian"
Should I update the documentation for `torch.matrix_rank`?
For the input with no elements (for example 0×0 matrix), the current implementation is divergent from NumPy. NumPy stumbles on not defined max for such input, here I chose to return appropriately sized tensor of zeros. I think that's mathematically a correct thing to do.
Ref https://github.com/pytorch/pytorch/issues/42666.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/48206
Reviewed By: albanD
Differential Revision: D25211965
Pulled By: mruberry
fbshipit-source-id: ae87227150ab2cffa07f37b4a3ab228788701837
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:
This PR adds `torch.linalg.eigh`, and `torch.linalg.eigvalsh` for NumPy compatibility.
The current `torch.symeig` uses (on CPU) a different LAPACK routine than NumPy (`syev` vs `syevd`). Even though it shouldn't matter in practice, `torch.linalg.eigh` uses `syevd` (as NumPy does).
Ref https://github.com/pytorch/pytorch/issues/42666
Pull Request resolved: https://github.com/pytorch/pytorch/pull/45526
Reviewed By: gchanan
Differential Revision: D25022659
Pulled By: mruberry
fbshipit-source-id: 3676b77a121c4b5abdb712ad06702ac4944e900a
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:
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:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/47225
Summary
-------
This PR implements Tensor.new_empty_strided. Many of our torch.* factory
functions have a corresponding new_* method (e.g., torch.empty and
torch.new_empty), but there is no corresponding method to
torch.empty_strided. This PR adds one.
Motivation
----------
The real motivation behind this is for vmap to be able to work through
CopySlices. CopySlices shows up a lot in double backwards because a lot
of view functions have backward formulas that perform view+inplace.
e0fd590ec9/torch/csrc/autograd/functions/tensor.cpp (L78-L106)
To support vmap through CopySlices, the approach in this stack is to:
- add `Tensor.new_empty_strided` and replace `empty_strided` in
CopySlices with that so that we can propagate batch information.
- Make some slight modifications to AsStridedBackward (and add
as_strided batching rule)
Please let me know if it would be better if I squashed everything related to
supporting vmap over CopySlices together into a single big PR.
Test Plan
---------
- New tests.
Test Plan: Imported from OSS
Reviewed By: ejguan
Differential Revision: D24741688
Pulled By: zou3519
fbshipit-source-id: b688047d2eb3f92998896373b2e9d87caf2c4c39
Summary:
This PR adds a function for calculating the Kronecker product of tensors.
The implementation is based on `at::tensordot` with permutations and reshape.
Tests pass.
TODO:
- [x] Add more test cases
- [x] Write documentation
- [x] Add entry `common_methods_invokations.py`
Ref. https://github.com/pytorch/pytorch/issues/42666
Pull Request resolved: https://github.com/pytorch/pytorch/pull/45358
Reviewed By: mrshenli
Differential Revision: D24680755
Pulled By: mruberry
fbshipit-source-id: b1f8694589349986c3abfda3dc1971584932b3fa
Summary:
Fixes https://github.com/pytorch/pytorch/issues/46373
As noted in https://github.com/pytorch/pytorch/issues/46373, there needs to be a flag passed into the engine that indicates whether it was executed through the backward api or grad api. Tentatively named the flag `accumulate_grad` since functionally, backward api accumulates grad into .grad while grad api captures the grad and returns it.
Moving changes not necessary to the python api (cpp, torchscript) to a new PR.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/46855
Reviewed By: ngimel
Differential Revision: D24649054
Pulled By: soulitzer
fbshipit-source-id: 6925d5a67d583eeb781fc7cfaec807c410e1fc65
Summary:
Related https://github.com/pytorch/pytorch/issues/38349
This PR implements `column_stack` as the composite ops of `torch.reshape` and `torch.hstack`, and makes `row_stack` as the alias of `torch.vstack`.
Todo
- [x] docs
- [x] alias pattern for `row_stack`
Pull Request resolved: https://github.com/pytorch/pytorch/pull/46313
Reviewed By: ngimel
Differential Revision: D24585471
Pulled By: mruberry
fbshipit-source-id: 62fc0ffd43d051dc3ecf386a3e9c0b89086c1d1c
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/45847
Original PR here https://github.com/pytorch/pytorch/pull/45084. Created this one because I was having problems with ghstack.
Test Plan: Imported from OSS
Reviewed By: mruberry
Differential Revision: D24136629
Pulled By: heitorschueroff
fbshipit-source-id: dd7c7540a33f6a19e1ad70ba2479d5de44abbdf9
Summary: Pull Request resolved: https://github.com/pytorch/pytorch/pull/45586
Test Plan: The unit test has been softened to be less platform sensitive.
Reviewed By: mruberry
Differential Revision: D24025415
Pulled By: robieta
fbshipit-source-id: ee986933b984e736cf1525e1297de6b21ac1f0cf
Summary:
This PR allows Timer to collect deterministic instruction counts for (some) snippets. Because of the intrusive nature of Valgrind (effectively replacing the CPU with an emulated one) we have to perform our measurements in a separate process. This PR writes a `.py` file containing the Timer's `setup` and `stmt`, and executes it within a `valgrind` subprocess along with a plethora of checks and error handling. There is still a bit of jitter around the edges due to the Python glue that I'm using, but the PyTorch signal is quite good and thus this provides a low friction way of getting signal. I considered using JIT as an alternative, but:
A) Python specific overheads (e.g. parsing) are important
B) JIT might do rewrites which would complicate measurement.
Consider the following bit of code, related to https://github.com/pytorch/pytorch/issues/44484:
```
from torch.utils._benchmark import Timer
counts = Timer(
"x.backward()",
setup="x = torch.ones((1,)) + torch.ones((1,), requires_grad=True)"
).collect_callgrind()
for c, fn in counts[:20]:
print(f"{c:>12} {fn}")
```
```
812800 ???:_dl_update_slotinfo
355600 ???:update_get_addr
308300 work/Python/ceval.c:_PyEval_EvalFrameDefault'2
304800 ???:__tls_get_addr
196059 ???:_int_free
152400 ???:__tls_get_addr_slow
138400 build/../c10/core/ScalarType.h:c10::typeMetaToScalarType(caffe2::TypeMeta)
126526 work/Objects/dictobject.c:_PyDict_LoadGlobal
114268 ???:malloc
101400 work/Objects/unicodeobject.c:PyUnicode_FromFormatV
85900 work/Python/ceval.c:_PyEval_EvalFrameDefault
79946 work/Objects/typeobject.c:_PyType_Lookup
72000 build/../c10/core/Device.h:c10::Device::validate()
70000 /usr/include/c++/8/bits/stl_vector.h:std::vector<at::Tensor, std::allocator<at::Tensor> >::~vector()
66400 work/Objects/object.c:_PyObject_GenericGetAttrWithDict
63000 ???:pthread_mutex_lock
61200 work/Objects/dictobject.c:PyDict_GetItem
59800 ???:free
58400 work/Objects/tupleobject.c:tupledealloc
56707 work/Objects/dictobject.c:lookdict_unicode_nodummy
```
Moreover, if we backport this PR to 1.6 (just copy the `_benchmarks` folder) and load those counts as `counts_1_6`, then we can easily diff them:
```
print(f"Head instructions: {sum(c for c, _ in counts)}")
print(f"1.6 instructions: {sum(c for c, _ in counts_1_6)}")
count_dict = {fn: c for c, fn in counts}
for c, fn in counts_1_6:
_ = count_dict.setdefault(fn, 0)
count_dict[fn] -= c
count_diffs = sorted([(c, fn) for fn, c in count_dict.items()], reverse=True)
for c, fn in count_diffs[:15] + [["", "..."]] + count_diffs[-15:]:
print(f"{c:>8} {fn}")
```
```
Head instructions: 7609547
1.6 instructions: 6059648
169600 ???:_dl_update_slotinfo
101400 work/Objects/unicodeobject.c:PyUnicode_FromFormatV
74200 ???:update_get_addr
63600 ???:__tls_get_addr
46800 work/Python/ceval.c:_PyEval_EvalFrameDefault
33512 work/Objects/dictobject.c:_PyDict_LoadGlobal
31800 ???:__tls_get_addr_slow
31700 build/../aten/src/ATen/record_function.cpp:at::RecordFunction::RecordFunction(at::RecordScope)
28300 build/../torch/csrc/utils/python_arg_parser.cpp:torch::FunctionSignature::parse(_object*, _object*, _object*, _object**, bool)
27800 work/Objects/object.c:_PyObject_GenericGetAttrWithDict
27401 work/Objects/dictobject.c:lookdict_unicode_nodummy
24115 work/Objects/typeobject.c:_PyType_Lookup
24080 ???:_int_free
21700 work/Objects/dictobject.c:PyDict_GetItemWithError
20700 work/Objects/dictobject.c:PyDict_GetItem
...
-3200 build/../c10/util/SmallVector.h:at::TensorIterator::binary_op(at::Tensor&, at::Tensor const&, at::Tensor const&, bool)
-3400 build/../aten/src/ATen/native/TensorIterator.cpp:at::TensorIterator::resize_outputs(at::TensorIteratorConfig const&)
-3500 /usr/include/c++/8/x86_64-redhat-linux/bits/gthr-default.h:std::unique_lock<std::mutex>::unlock()
-3700 build/../torch/csrc/utils/python_arg_parser.cpp:torch::PythonArgParser::raw_parse(_object*, _object*, _object**)
-4207 work/Objects/obmalloc.c:PyMem_Calloc
-4500 /usr/include/c++/8/bits/stl_vector.h:std::vector<at::Tensor, std::allocator<at::Tensor> >::~vector()
-4800 build/../torch/csrc/autograd/generated/VariableType_2.cpp:torch::autograd::VariableType::add__Tensor(at::Tensor&, at::Tensor const&, c10::Scalar)
-5000 build/../c10/core/impl/LocalDispatchKeySet.cpp:c10::impl::ExcludeDispatchKeyGuard::ExcludeDispatchKeyGuard(c10::DispatchKey)
-5300 work/Objects/listobject.c:PyList_New
-5400 build/../torch/csrc/utils/python_arg_parser.cpp:torch::FunctionParameter::check(_object*, std::vector<pybind11::handle, std::allocator<pybind11::handle> >&)
-5600 /usr/include/c++/8/bits/std_mutex.h:std::unique_lock<std::mutex>::unlock()
-6231 work/Objects/obmalloc.c:PyMem_Free
-6300 work/Objects/listobject.c:list_repeat
-11200 work/Objects/listobject.c:list_dealloc
-28900 build/../torch/csrc/utils/python_arg_parser.cpp:torch::FunctionSignature::parse(_object*, _object*, _object**, bool)
```
Remaining TODOs:
* Include a timer in the generated script for cuda sync.
* Add valgrind to CircleCI machines and add a unit test.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/44717
Reviewed By: soumith
Differential Revision: D24010742
Pulled By: robieta
fbshipit-source-id: df6bc765f8efce7193893edba186cd62b4b23623
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/44433
Not entirely sure why, but changing the type of beta from `float` to `double in autocast_mode.cpp and FunctionsManual.h fixes my compiler errors, failing instead at link time
fixing some type errors, updated fn signature in a few more files
removing my usage of Scalar, making beta a double everywhere instead
Test Plan: Imported from OSS
Reviewed By: mrshenli
Differential Revision: D23636720
Pulled By: bdhirsh
fbshipit-source-id: caea2a1f8dd72b3b5fd1d72dd886b2fcd690af6d
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/45149
The choose_qparams_optimized calculates the the optimized qparams.
It uses a greedy approach to nudge the min and max and calculate the l2 norm
and tries to minimize the quant error by doing `torch.norm(x-fake_quant(x,s,z))`
Test Plan: Imported from OSS
Reviewed By: raghuramank100
Differential Revision: D23848060
fbshipit-source-id: c6c57c9bb07664c3f1c87dd7664543e09f634aee
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/43680
As discussed [here](https://github.com/pytorch/pytorch/issues/43342),
adding in a Python-only implementation of the triplet-margin loss that takes a
custom distance function. Still discussing whether this is necessary to add to
PyTorch Core.
Test Plan:
python test/run_tests.py
Imported from OSS
Reviewed By: albanD
Differential Revision: D23363898
fbshipit-source-id: 1cafc05abecdbe7812b41deaa1e50ea11239d0cb
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/39955
resolves https://github.com/pytorch/pytorch/issues/36323 by adding `torch.sgn` for complex tensors.
`torch.sgn` returns `x/abs(x)` for `x != 0` and returns `0 + 0j` for `x==0`
This PR doesn't test the correctness of the gradients. It will be done as a part of auditing all the ops in future once we decide the autograd behavior (JAX vs TF) and add gradchek.
Test Plan: Imported from OSS
Reviewed By: mruberry
Differential Revision: D23460526
Pulled By: anjali411
fbshipit-source-id: 70fc4e14e4d66196e27cf188e0422a335fc42f92
Summary:
These alias are consistent with NumPy. Note that C++'s naming would be different (std::multiplies and std::divides), and that PyTorch's existing names (mul and div) are consistent with Python's dunders.
This also improves the instructions for adding an alias to clarify that dispatch keys should be removed when copying native_function.yaml entries to create the alias entries.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/44463
Reviewed By: ngimel
Differential Revision: D23670782
Pulled By: mruberry
fbshipit-source-id: 9f1bdf8ff447abc624ff9e9be7ac600f98340ac4
Summary:
Ref https://github.com/pytorch/pytorch/issues/42175, fixes https://github.com/pytorch/pytorch/issues/34797
This adds complex support to `torch.stft` and `torch.istft`. Note that there are really two issues with complex here: complex signals, and returning complex tensors.
## Complex signals and windows
`stft` currently assumes all signals are real and uses `rfft` with `onesided=True` by default. Similarly, `istft` always takes a complex fourier series and uses `irfft` to return real signals.
For `stft`, I now allow complex inputs and windows by calling the full `fft` if either are complex. If the user gives `onesided=True` and the signal is complex, then this doesn't work and raises an error instead. For `istft`, there's no way to automatically know what to do when `onesided=False` because that could either be a redundant representation of a real signal or a complex signal. So there, the user needs to pass the argument `return_complex=True` in order to use `ifft` and get a complex result back.
## stft returning complex tensors
The other issue is that `stft` returns a complex result, represented as a `(... X 2)` real tensor. I think ideally we want this to return proper complex tensors but to preserver BC I've had to add a `return_complex` argument to manage this transition. `return_complex` defaults to false for real inputs to preserve BC but defaults to True for complex inputs where there is no BC to consider.
In order to `return_complex` by default everywhere without a sudden BC-breaking change, a simple transition plan could be:
1. introduce `return_complex`, defaulted to false when BC is an issue but giving a warning. (this PR)
2. raise an error in cases where `return_complex` defaults to false, making it a required argument.
3. change `return_complex` default to true in all cases.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/43886
Reviewed By: glaringlee
Differential Revision: D23760174
Pulled By: mruberry
fbshipit-source-id: 2fec4404f5d980ddd6bdd941a63852a555eb9147
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/44393
torch.quantile now correctly propagates nan and implemented torch.nanquantile similar to numpy.nanquantile.
Test Plan: Imported from OSS
Reviewed By: albanD
Differential Revision: D23649613
Pulled By: heitorschueroff
fbshipit-source-id: 5201d076745ae1237cedc7631c28cf446be99936
Summary:
This PR adds the following aliaes:
- not_equal for torch.ne
- greater for torch.gt
- greater_equal for torch.ge
- less for torch.lt
- less_equal for torch.le
This aliases are consistent with NumPy's naming for these functions.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/43870
Reviewed By: zou3519
Differential Revision: D23498975
Pulled By: mruberry
fbshipit-source-id: 78560df98c9f7747e804a420c1e53fd1dd225002
Summary:
Adds two more "missing" NumPy aliases: arctanh and arcsinh, and simplifies the dispatch of other arc* aliases.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/43762
Reviewed By: ngimel
Differential Revision: D23396370
Pulled By: mruberry
fbshipit-source-id: 43eb0c62536615fed221d460c1dec289526fb23c
Summary:
Add a max/min operator that only return values.
## Some important decision to discuss
| **Question** | **Current State** |
|---------------------------------------|-------------------|
| Expose torch.max_values to python? | No |
| Remove max_values and only keep amax? | Yes |
| Should amax support named tensors? | Not in this PR |
## Numpy compatibility
Reference: https://numpy.org/doc/stable/reference/generated/numpy.amax.html
| Parameter | PyTorch Behavior |
|--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|-----------------------------------------------------------------------------------|
| `axis`: None or int or tuple of ints, optional. Axis or axes along which to operate. By default, flattened input is used. If this is a tuple of ints, the maximum is selected over multiple axes, instead of a single axis or all the axes as before. | Named `dim`, behavior same as `torch.sum` (https://github.com/pytorch/pytorch/issues/29137) |
| `out`: ndarray, optional. Alternative output array in which to place the result. Must be of the same shape and buffer length as the expected output. | Same |
| `keepdims`: bool, optional. If this is set to True, the axes which are reduced are left in the result as dimensions with size one. With this option, the result will broadcast correctly against the input array. | implemented as `keepdim` |
| `initial`: scalar, optional. The minimum value of an output element. Must be present to allow computation on empty slice. | Not implemented in this PR. Better to implement for all reductions in the future. |
| `where`: array_like of bool, optional. Elements to compare for the maximum. | Not implemented in this PR. Better to implement for all reductions in the future. |
**Note from numpy:**
> NaN values are propagated, that is if at least one item is NaN, the corresponding max value will be NaN as well. To ignore NaN values (MATLAB behavior), please use nanmax.
PyTorch has the same behavior
Pull Request resolved: https://github.com/pytorch/pytorch/pull/43092
Reviewed By: ngimel
Differential Revision: D23360705
Pulled By: mruberry
fbshipit-source-id: 5bdeb08a2465836764a5a6fc1a6cc370ae1ec09d
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/42708
Add rowwise prune pytorch op.
This operator introduces sparsity to the 'weights' matrix with the help
of the importance indicator 'mask'.
A row is considered important and not pruned if the mask value for that
particular row is 1(True) and not important otherwise.
Test Plan:
buck test caffe2/torch/fb/sparsenn:test -- rowwise_prune
buck test caffe2/test:pruning
Reviewed By: supriyar
Differential Revision: D22849432
fbshipit-source-id: 456f4f77c04158cdc3830b2e69de541c7272a46d
Summary:
Related to https://github.com/pytorch/pytorch/issues/38349
Implement NumPy-like functions `maximum` and `minimum`.
The `maximum` and `minimum` functions compute input tensors element-wise, returning a new array with the element-wise maxima/minima.
If one of the elements being compared is a NaN, then that element is returned, both `maximum` and `minimum` functions do not support complex inputs.
This PR also promotes the overloaded versions of torch.max and torch.min, by re-dispatching binary `torch.max` and `torch.min` to `torch.maximum` and `torch.minimum`.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/42579
Reviewed By: mrshenli
Differential Revision: D23153081
Pulled By: mruberry
fbshipit-source-id: 803506c912440326d06faa1b71964ec06775eac1
Summary:
This adds the torch.arccosh alias and updates alias testing to validate the consistency of the aliased and original operations. The alias testing is also updated to run on CPU and CUDA, which revealed a memory leak when tracing (see https://github.com/pytorch/pytorch/issues/43119).
Pull Request resolved: https://github.com/pytorch/pytorch/pull/43107
Reviewed By: ngimel
Differential Revision: D23156472
Pulled By: mruberry
fbshipit-source-id: 6155fac7954fcc49b95e7c72ed917c85e0eabfcd
Summary:
This PR:
- updates test_op_normalization.py, which verifies that aliases are correctly translated in the JIT
- adds torch.linalg.det as an alias for torch.det
- moves the torch.linalg.outer alias to torch.outer (to be consistent with NumPy)
The torch.linalg.outer alias was put the linalg namespace erroneously as a placeholder since it's a "linear algebra op" according to NumPy but is actually still in the main NumPy namespace.
The updates to test_op_normalization are necessary. Previously it was using method_tests to generate tests, and method_tests assumes test suites using it also use the device generic framework, which test_op_normalization did not. For example, some ops require decorators like `skipCPUIfNoLapack`, which only works in device generic test classes. Moving test_op_normalization to the device generic framework also lets these tests run on CPU and CUDA.
Continued reliance on method_tests() is excessive since the test suite is only interested in testing aliasing, and a simpler and more readable `AliasInfo` class is used for the required information. An example impedance mismatch between method_tests and the new tests, for example, was how to handle ops in namespaces like torch.linalg.det. In the future this information will likely be folded into a common 'OpInfo' registry in the test suite.
The actual tests performed are similar to what they were previously: a scripted and traced version of the op is run and the test verifies that both graphs do not contain the alias name and do contain the aliased name.
The guidance for adding an alias has been updated accordingly.
cc mattip
Note:
ngimel suggests:
- deprecating and then removing the `torch.ger` name
- reviewing the implementation of `torch.outer`
Pull Request resolved: https://github.com/pytorch/pytorch/pull/42802
Reviewed By: zou3519
Differential Revision: D23059883
Pulled By: mruberry
fbshipit-source-id: 11321c2a7fb283a6e7c0d8899849ad7476be42d1
Summary:
Per title. Also updates our guidance for adding aliases to clarify interned_string and method_test requirements. The alias is tested by extending test_clamp to also test clip.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/42770
Reviewed By: ngimel
Differential Revision: D23020655
Pulled By: mruberry
fbshipit-source-id: f1d8e751de9ac5f21a4f95d241b193730f07b5dc
Summary:
According to pytorch/rfcs#3
From the goals in the RFC:
1. Support subclassing `torch.Tensor` in Python (done here)
2. Preserve `torch.Tensor` subclasses when calling `torch` functions on them (done here)
3. Use the PyTorch API with `torch.Tensor`-like objects that are _not_ `torch.Tensor`
subclasses (done in https://github.com/pytorch/pytorch/issues/30730)
4. Preserve `torch.Tensor` subclasses when calling `torch.Tensor` methods. (done here)
5. Propagating subclass instances correctly also with operators, using
views/slices/indexing/etc. (done here)
6. Preserve subclass attributes when using methods or views/slices/indexing. (done here)
7. A way to insert code that operates on both functions and methods uniformly
(so we can write a single function that overrides all operators). (done here)
8. The ability to give external libraries a way to also define
functions/methods that follow the `__torch_function__` protocol. (will be addressed in a separate PR)
This PR makes the following changes:
1. Adds the `self` argument to the arg parser.
2. Dispatches on `self` as well if `self` is not `nullptr`.
3. Adds a `torch._C.DisableTorchFunction` context manager to disable `__torch_function__`.
4. Adds a `torch::torch_function_enabled()` and `torch._C._torch_function_enabled()` to check the state of `__torch_function__`.
5. Dispatches all `torch._C.TensorBase` and `torch.Tensor` methods via `__torch_function__`.
TODO:
- [x] Sequence Methods
- [x] Docs
- [x] Tests
Closes https://github.com/pytorch/pytorch/issues/28361
Benchmarks in https://github.com/pytorch/pytorch/pull/37091#issuecomment-633657778
Pull Request resolved: https://github.com/pytorch/pytorch/pull/37091
Reviewed By: ngimel
Differential Revision: D22765678
Pulled By: ezyang
fbshipit-source-id: 53f8aa17ddb8b1108c0997f6a7aa13cb5be73de0
