### Introduction
<!-- What did you change and why was it needed? -->
Removing unnecessary weight gradient calculation is very important for applications that need high-order derivatives during training. However, this is not supported by the current Autograd engine.
For more detail: The backward function of a `matmul` operator (e.g., `linear` `addmm` `mm`), has two matmuls, one for `input gradient` and another for `weight gradient`. For a typical neural network (nn) with a few linear layers and activation functions, if the user calls `torch.autograd.grad()` to calculate the derivative of the nn output `y` w.r.t the nn input `x`, only the `input gradient` of the `matmul` operator is needed, and the `weight gradient` is discarded. However, the current PyTorch autograd engine will always calculate the `weight gradient` if `weight` requires gradient (the calculation of the high-order derivative is performed during training).
The figure attached shows the autograd graph of the following code snippet:
```py
y = torch.nn.functional.linear(x, weight, bias)
y = y.pow(2)
# first order derivative
y__x, = torch.autograd.grad(y, x, grad_outputs=grad_outputs, create_graph=True)
# first order derivative
y__x__x, = torch.autograd.grad(y__x, x, grad_outputs=grad_outputs, create_graph=True)
```
The path with ❌ is not needed when calculating derivatives.
<img width="50%" alt="image" src="https://user-images.githubusercontent.com/9999318/182018117-719c5a23-bcc6-4a63-8e8d-1bca3ebda2e3.png">
### Issue
<!-- Link to Issue ticket or RFP -->
Related issue: https://github.com/pytorch/pytorch/issues/56500
### Method
When calling `torch.autograd.grad`, `exec_info_` is created for each GraphTask, which allows filtering paths on the graph that are not needed. However, when the GraphTask calls into the node, the node still does not know whether the edges are needed or not. In the case of matmul, `weight.requires_grad is True` so the weight gradient is always calculated.
Following https://github.com/pytorch/pytorch/issues/56500#issuecomment-825694656, this PR passes the graph task's thread_local `exec_info_` into the node, so it could trim unnecessary edges during `torch.autograd.grad` calls.
### Benchmark
Benchmark script: https://gist.github.com/yueyericardo/24158433a2021c51eeef9c3e2722df99
Benchmark result:
6 hidden layers, batch size 10000, on A100
FP32 result
| hessian benchmark | FP32 (before) | FP32 (After) | FP32 (Functorch v0.1.1) |
| ----------------------------- | ------------- | ----------------- | ----------------------- |
| Linear + ReLU (no backward) | 55.658 ms | 29.392 ms (1.90X) | 29.547 ms (1.90X) |
| Linear + ReLU (with backward) | 81.173 ms | 54.917 ms (1.47X) | 68.988 ms (1.18X) |
TF32 result
| hessian benchmark | TF32 (before) | TF32 (after) | TF32 (Functorch v0.1.1) |
| ----------------------------- | ------------- | ----------------- | ----------------------- |
| Linear + ReLU (no backward) | 19.801 ms | 11.259 ms (1.76X) | 10.754 ms (1.84X) |
| Linear + ReLU (with backward) | 29.167 ms | 20.466 ms (1.42X) | 22.784 ms (1.28X) |
For FP32 result, we could get 1.9X speed up for hessian calculation, and 1.47X speed up during training, which is even faster than functorch `vmap(jacfwd(jacrev` implementation. (functorch has performance regression on v0.2.0, https://github.com/pytorch/functorch/issues/989, so we are using v0.1.1 for benchmark)
@zou3519 does functorch also includes similar optimizations during hessian calculation? If not, what do we need to do so the functorch could also benefit from this PR?
### Testing
<!-- How did you test your change? -->
- [x] we need to figure out a way for unittest
### Thanks
Thanks for the great blog: [How Computational Graphs are Executed in PyTorch | PyTorch](https://pytorch.org/blog/how-computational-graphs-are-executed-in-pytorch/)
cc @zasdfgbnm @albanD
Pull Request resolved: https://github.com/pytorch/pytorch/pull/82544
Approved by: https://github.com/soulitzer
And use it throughout the CMakeLists and rectify `IF(APPLE)`/`IF(GNU_CXX_VERSION VERSION_GREATER A.B)` and so on
Also, add `target_compile_options_if_supported` and use it in `Dependencies.cmake` as well as in test's `CMakeListst.txt`
Delete `-Wno-unknown-warning-option` to test that conditions indeed working as expected
Pull Request resolved: https://github.com/pytorch/pytorch/pull/82883
Approved by: https://github.com/seemethere
And use it throughout the CMakeLists and rectify `IF(APPLE)`/`IF(GNU_CXX_VERSION VERSION_GREATER A.B)` and so on
Also, add `target_compile_options_if_supported` and use it in `Dependencies.cmake` as well as in test's `CMakeListst.txt`
Delete `-Wno-unknown-warning-option` to test that conditions indeed working as expected
Pull Request resolved: https://github.com/pytorch/pytorch/pull/82883
Approved by: https://github.com/seemethere
See this doc: https://docs.google.com/document/d/1KiRdnoj6B4cI3yl017hTbCqcOGO1gWIpUf20sldipHM/edit#
Two issues (1) regarding hooks in general and (2) regarding retains grad hooks are fixed, Python hooks, which rely on a different mechanism are not discussed here:
- Hooks in cpp in general
- (fixed) new hooks to registered to a newer version of the tensor no longer get applied to grad_fn
associated with older version of the tensor when the first hook was ever registered
- (unchanged) hooks registered to the older version of the tensor remain active on
- Retains grad hooks
- (fixed) now get moved to the latest grad_fn. NB: To the user, retains_grad is not considered hooks
or expected to behave like hooks (which we consider properties of the grad_fn) vs retains_gradness
which is a property of the tensor.
- (not in this PR) Python hooks
- (will fix) same issue as hooks in cpp where new hooks are being applied to grad_fn associated
with the older version of the tensor
Pull Request resolved: https://github.com/pytorch/pytorch/pull/79996
Approved by: https://github.com/albanD
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/75244
Original commit changeset: d653a5af662a
Original Phabricator Diff: D35060736 (d9d34922a0)
Test Plan: Model loading test, verified that D35060736 (d9d34922a0) will cause the torch::save => torch::load failure.
Reviewed By: yinghai, jianyuh
Differential Revision: D35387009
fbshipit-source-id: 9d176992d402d57779e2af3d905b3c1538335298
(cherry picked from commit 6c8cc0d3b8a88b15e35702d70e18bbae8aa4628a)
Summary:
There are various possible approaches, but the approach chosen minimizes disruption to source control blame.
Addresses:
```
error: Function _ZN23FunctionalTest_Pad_Test8TestBodyEv is too big to optimize [-Werror,-Wignored-optimization-argument]
```
Test Plan: buck2 build mode/opt caffe2/test/cpp/api:functional
Reviewed By: jamesr66a
Differential Revision: D34027291
fbshipit-source-id: 9dfd771ad56d3d4bc0d41b38b04654c8dae7c006
(cherry picked from commit d43b5a7ed6)
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/69394
Modified loops in files under fbsource/fbcode/caffe2/ from the format
```
for(TYPE var=x0;var<x_max;x++)
```
to the format
```
for(const auto var: irange(xmax))
```
This was achieved by running r-barnes's loop upgrader script (D28874212) with some modification to exclude all files under /torch/jit and a number of reversions or unused variable suppression warnings added by hand.
Test Plan: Sandcastle
Reviewed By: malfet
Differential Revision: D32837991
fbshipit-source-id: fc7c4f76d2f32a17a0faf329294b3fe7cb81df32
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/66743
Modified loops in files under fbsource/fbcode/caffe2/ from the format
`for(TYPE var=x0;var<x_max;x++)`
to the format
`for(const auto var: irange(xmax))`
This was achieved by running r-barnes's loop upgrader script (D28874212) with some modification to exclude all files under /torch/jit and a number of reversions or unused variable suppression warnings added by hand.
Test Plan: Sandcastle
Reviewed By: malfet
Differential Revision: D31705359
fbshipit-source-id: c9ea2fbc0f9cd29e97a52dcb203addc5f2abb09b
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/67195
Now that `is_nonzero` is part of `at::native` refer https://github.com/pytorch/pytorch/pull/66663, replacing `TensorCompare::is_nonzero` to `at::native::is_nonzero`
ghstack-source-id: 141514416
Test Plan: CI
Reviewed By: larryliu0820
Differential Revision: D31704041
fbshipit-source-id: 36813e5411d0aa2eb2d0442e2a195bbed417b33d
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/66744
Modified loops in files under fbsource/fbcode/caffe2/ from the format
`for(TYPE var=x0;var<x_max;x++)`
to the format
`for(const auto var: irange(xmax))`
This was achieved by running r-barnes's loop upgrader script (D28874212) with some modification to exclude all files under /torch/jit and a number of reversions or unused variable suppression warnings added by hand.
Test Plan: Sandcastle
Reviewed By: ngimel
Differential Revision: D31705358
fbshipit-source-id: d6ea350cbaa8f452fc78f238160e5374be637a48
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/66234
Modified loops in files under fbsource/fbcode/caffe2/ from the format
`for(TYPE var=x0;var<x_max;x++)`
to the format
`for(const auto var: irange(xmax))`
This was achieved by running r-barnes's loop upgrader script (D28874212) with some modification to exclude all files under /torch/jit and a number of reversions or unused variable suppression warnings added by hand.
bypass_size_limit
allow-large-files
Test Plan: Sandcastle
Reviewed By: ngimel
Differential Revision: D30652629
fbshipit-source-id: 0ae6c4bbbb554bad42e372792a6430e1acf15e3e
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/63878
See https://github.com/pytorch/pytorch/issues/64407, https://github.com/pytorch/pytorch/issues/62032 for context:
In this PR:
- Add boxed kernel by replicating `gen_inplace_or_view`'s logic that is ONLY for use with the Autograd not-implemented kernel
- Unlike `gen_inplace_or_view` we always pass a view_func to as_view in order to ensure that an "derivative is not implemented" error is raised even if an in-place update is performed on the view. Without the `view_func`, the CopySlice + AsStridedBackward nodes would replace the NotImplemented node.
- This limitation makes it impossible to use this node for general use
- view relationship must be between first input (must be tensor) and first output (may be tensor or vec of tensor)
- do not support non-differentiable views (_values, _indices, view.dtype) - view relationship is always fw and bw differentiable
- Adds the macro `#define REGISTER_AUTOGRAD_NOT_IMPLEMENTED_FALLBACK(ns, op)` to be the interface for this feature:
- static initialization can be slowed down(? not measured) if there are many registrations, because each line translates to 2 library calls but the workaround is just to manually use the two functions `AutogradNotImplementedFallback` and `ADInplaceOrViewFallback` and call `m.impl`.
- Adds testing:
- for views: view relationship created
- performing in-place operation on the view, raises properly
- trying to create two view relationships is not allowed,
- single view relationship but not first input/first output should error
- view relation created properly for tensor vector output
- for in-place:
- version count bump
- triggers rebase_history
- multiple mutations is okay and also updates version counter
- TODO (follow up): Update tutorials for adding third-party operators (and document the above limitations)
- TODO (follow up): Look at torch-audio/torch-vision and identify places where this can simplify existing code
EDIT: Made it more clear what is introduced in this PR and moved some more contextual stuff into the issue itself
Test Plan: Imported from OSS
Reviewed By: albanD
Differential Revision: D30901714
Pulled By: soulitzer
fbshipit-source-id: 48de14c28be023ff4bd31b7ea5e7cba88aeee04c
Summary:
Delete `-Wno-unused-variable` from top level `CMakeLists.txt`
Still suppress those warnings for tests and `torch_python`
Delete number of unused variables from caffe2 code
Use `(void)var;` to suppress unused variable in range loops
Use `C10_UNUSED` for global constructors and use `constexpr` instead of `static` for global constants
Do not delete `caffe2::OperatorBase::Output` calls as they have side effects
Pull Request resolved: https://github.com/pytorch/pytorch/pull/66041
Reviewed By: ngimel
Differential Revision: D31360142
Pulled By: malfet
fbshipit-source-id: 6fdfb9f91efdc49ca984a2f2a17ee377d28210c8
Summary:
Delete `-Wno-unused-variable` from top level `CMakeLists.txt`
Still suppress those warnings for tests and `torch_python`
Delete number of unused variables from caffe2 code
Use `(void)var;` to suppress unused variable in range loops
Use `C10_UNUSED` for global constructors and use `constexpr` instead of `static` for global constants
Pull Request resolved: https://github.com/pytorch/pytorch/pull/65954
Reviewed By: ngimel
Differential Revision: D31326599
Pulled By: malfet
fbshipit-source-id: 924155f1257a2ba1896c50512f615e45ca1f61f3
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/65861
First in a series. This PR changes the code in deploy.h/cpp and
interpreter_impl.h/cpp to be camel case instead of snake case. Starting
with this as it has the most impact on downstream users.
Test Plan: Imported from OSS
Reviewed By: shannonzhu
Differential Revision: D31291183
Pulled By: suo
fbshipit-source-id: ba6f74042947c9a08fb9cb3ad7276d8dbb5b2934
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/63612
This makes Tensor inherit from a new class TensorBase, that provides a subset of Tensor that doesn't
directly depend on native_functions.yaml. Code that only includes TensorBase.h with thus not need to
be rebuilt every time someone changes an operator signature.
Making `Tensor` inherit from this class means that `const TensorBase&` parameters will be callable
with an ordinary `Tensor`. I've also made `Tensor` constructible and assignable from `TensorBase` to
minimize friction in code mixing the two types.
To help enforce that `Tensor.h` and `Functions.h` aren't accidentally included, I've added an error
into `Operators.h` if `TORCH_ASSERT_NO_OPERATORS` is defined. We can either set this in the build
system for certain folders, or just define it at the top of any file.
I've also included an example of manually special-casing the commonly used `contiguous` operator.
The inline function's slow path defers to `TensorBase::__dispatch_contiguous` which is defined in
`Tensor.cpp`. I've made it so `OptionalTensorRef` is constructible from `TensorBase`, so I can
materialize a `Tensor` for use in dispatch without actually increasing its refcount.
Test Plan: Imported from OSS
Reviewed By: gchanan
Differential Revision: D30728580
Pulled By: ezyang
fbshipit-source-id: 2cbc8eee08043382ee6904ea8e743b1286921c03
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/61801
resubmitting because the last one was unrecoverable due to making changes incorrectly in the stack
Test Plan: Imported from OSS
Reviewed By: desertfire
Differential Revision: D29812510
Pulled By: makslevental
fbshipit-source-id: ba9685dc81b6699724104d5ff3211db5852370a6
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/63995
JIT methods already have name() in their interface, and Py methods have names in their implementation. I'm adding this for a particular case where someone tried to use name() on a JIT method that we're replacing with an IMethod.
Test Plan: add case to imethod API test
Reviewed By: suo
Differential Revision: D30559401
fbshipit-source-id: 76236721f5cd9a9d9d488ddba12bfdd01d679a2c
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/63345
This diff did the following few things to enable the tests:
1. Exposed IMethod as TORCH_API.
2. Linked torch_deploy to test_api if USE_DEPLOY == 1.
3. Generated torch::deploy examples when building torch_deploy library.
Test Plan: ./build/bin/test_api --gtest_filter=IMethodTest.*
Reviewed By: ngimel
Differential Revision: D30346257
Pulled By: alanwaketan
fbshipit-source-id: 932ae7d45790dfb6e00c51893933a054a0fad86d
