Summary:
Retake on https://github.com/pytorch/pytorch/issues/40493 after all the feedback from albanD
This PR implements the generic Lazy mechanism and a sample `LazyLinear` layer with the `UninitializedParameter`.
The main differences with the previous PR are two;
Now `torch.nn.Module` remains untouched.
We don't require an explicit initialization or a dummy forward pass before starting the training or inference of the actual module. Making this much simpler to use from the user side.
As we discussed offline, there was the suggestion of not using a mixin, but changing the `__class__` attribute of `LazyLinear` to become `Linear` once it's completely initialized. While this can be useful, by the time being we need `LazyLinear` to be a `torch.nn.Module` subclass since there are many checks that rely on the modules being instances of `torch.nn.Module`.
This can cause problems when we create complex modules such as
```
class MyNetwork(torch.nn.Module):
def __init__(self):
super(MyNetwork, self).__init__()
self.conv = torch.nn.Conv2d(20, 4, 2)
self.linear = torch.nn.LazyLinear(10)
def forward(self, x):
y = self.conv(x).clamp(min=0)
return self.linear(y)
```
Here, when the __setattr__ function is called at the time LazyLinear is registered, it won't be added to the child modules of `MyNetwork`, so we have to manually do it later, but currently there is no way to do such thing as we can't access the parent module from LazyLinear once it becomes the Linear module. (We can add a workaround to this if needed).
TODO:
Add convolutions once the design is OK
Fix docstrings
Pull Request resolved: https://github.com/pytorch/pytorch/pull/44538
Reviewed By: ngimel
Differential Revision: D24162854
Pulled By: albanD
fbshipit-source-id: 6d58dfe5d43bfb05b6ee506e266db3cf4b885f0c
Summary:
Apologize if this seems trivial, but i'd like to fix them on my way of reading some of the source code. Thanks!
Pull Request resolved: https://github.com/pytorch/pytorch/pull/40692
Differential Revision: D22284651
Pulled By: mrshenli
fbshipit-source-id: 4259d1808aa4d15a02cfd486cfb44dd75fdc58f8
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/38211
Just because the annotations are inline doesn't mean the files type
check; most of the newly annotated files have type errors and I
added exclusions for them in mypy.ini. The payoff of moving
all of these modules inline is I can delete the relevant code
generation logic for the pyi files (which was added ignore
annotations that weren't actually relevant anymore.)
For the most part the translation was completely mechanical, but there
were two hairy issues. First, I needed to work around a Python 3.6 and
earlier bug where Generic has a nontrivial metaclass. This fix is in
torch/jit/__init__.py. Second, module.py, we need to apply the same
fix for avoiding contravariance checks that the pyi file used to have;
this is done by declaring forward as a variable (rather than a
function), which appears to be sufficient enough to get mypy to not
contravariantly check input arguments.
Because we aren't actually typechecking these modules in most
cases, it is inevitable that some of these type annotations are wrong.
I slavishly copied the old annotations from the pyi files unless there
was an obvious correction I could make. These annotations will probably
need fixing up later.
Signed-off-by: Edward Z. Yang <ezyang@fb.com>
Test Plan: Imported from OSS
Differential Revision: D21497397
Pulled By: ezyang
fbshipit-source-id: 2b08bacc152c48f074e7edc4ee5dce1b77d83702
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/32745
Some parameters (like `bias` in conv) are optional. To achieve this
previously, you had to add `bias` as a constant, which would invoke some
pretty weird behavior in the frontend, summarized as:
```
if bias is not None:
add it as a parameter normally
else: # bias is None
add it as a constant with the value None
```
There are several things bad about this:
1. Bias is not a constant. Marking it `__constants__` is confusing.
2. It basically relies on an implementation detail (the frontend
processes parameters before constants) to work.
Okay, whatever. I don't even know why we did this originally, but
getting rid of it doesn't break anything, so I assume improved NoneType
refinement has made this a non-issue.
Note on perf: this will make no difference; if bias was `None` it's still
folded out today, if bias is a Tensor it would be added as a parameter
both before and after this change
Test Plan: Imported from OSS
Differential Revision: D19628634
Pulled By: suo
fbshipit-source-id: d9128a09c5d096b938fcf567b8c23b09ac9ab37f
Summary:
* Deletes all weak script decorators / associated data structures / methods
* In order to keep supporting the standard library in script, this enables recursive script on any function defined in `torch.nn`
* Most changes in `torch/nn` are the result of `ag -Q "weak" torch/nn/ -l | xargs sed -i '/weak/d'`, only `rnn.py` needed manual editing to use the `ignore` and `export` to continue supporting the overloaded `forward` methods
* `Sequential`/`ModuleList` no longer need to be added to constants since they are compiled on demand
This should also fix https://github.com/pytorch/pytorch/issues/22212
Pull Request resolved: https://github.com/pytorch/pytorch/pull/22212
Differential Revision: D15988346
Pulled By: driazati
fbshipit-source-id: af223e3ad0580be895377312949997a70e988e4f
Summary:
* Constructs a new type at runtime so that `isinstance` checks work for
weak modules assigned to `ScriptModule`s
* Fix some extraneous names in `__constants__`
* Add `in_features` and `out_features` to `nn.Linear` `__constants__`
Fixes#19363
Pull Request resolved: https://github.com/pytorch/pytorch/pull/20190
Pulled By: driazati
Differential Revision: D15302350
fbshipit-source-id: 1d4d21ed44ab9578a4bc2a72396a82e9bbcd387c
Summary:
This is a simple yet useful addition to the torch.nn modules: an identity module. This is a first draft - please let me know what you think and I will edit my PR.
There is no identity module - nn.Sequential() can be used, however it is argument sensitive so can't be used interchangably with any other module. This adds nn.Identity(...) which can be swapped with any module because it has dummy arguments. It's also more understandable than seeing an empty Sequential inside a model.
See discussion on #9160. The current solution is to use nn.Sequential(). However this won't work as follows:
```python
batch_norm = nn.BatchNorm2d
if dont_use_batch_norm:
batch_norm = Identity
```
Then in your network, you have:
```python
nn.Sequential(
...
batch_norm(N, momentum=0.05),
...
)
```
If you try to simply set `Identity = nn.Sequential`, this will fail since `nn.Sequential` expects modules as arguments. Of course there are many ways to get around this, including:
- Conditionally adding modules to an existing Sequential module
- Not using Sequential but writing the usual `forward` function with an if statement
- ...
**However, I think that an identity module is the most pythonic strategy,** assuming you want to use nn.Sequential.
Using the very simple class (this isn't the same as the one in my commit):
```python
class Identity(nn.Module):
def __init__(self, *args, **kwargs):
super().__init__()
def forward(self, x):
return x
```
we can get around using nn.Sequential, and `batch_norm(N, momentum=0.05)` will work. There are of course other situations this would be useful.
Thank you.
Best,
Miles
Pull Request resolved: https://github.com/pytorch/pytorch/pull/19249
Differential Revision: D15012969
Pulled By: ezyang
fbshipit-source-id: 9f47e252137a1679e306fd4c169dca832eb82c0c
Summary:
PR to update the shape notation for all of the torch.nn modules to take a unified form. The goal is to make these definitions machine-readable and those checkable by unifying the style across all of the different modules.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/15741
Differential Revision: D13709601
Pulled By: ezyang
fbshipit-source-id: fb89a03903fdf0cd0dcf76f3e469b8582b2f3634
Summary:
This PR:
1. Handle None value attr in the WeakScriptModuleProxy
2. add back module tests that now passing
Pull Request resolved: https://github.com/pytorch/pytorch/pull/14715
Differential Revision: D13313573
Pulled By: wanchaol
fbshipit-source-id: a6b7892707350290a6d69b6f6270ad089bfc954b
Summary:
This will resolve the code and comments mis-match issue.
Closes https://github.com/pytorch/pytorch/pull/9070
Differential Revision: D8712261
Pulled By: ezyang
fbshipit-source-id: a8a7d8af890a41ec246e11c2a62b0bde297be9c1
This PR enables users to print extra information of their subclassed nn.Module.
Now I simply insert the user-defined string at the ending of module name, which should be discussed in this PR.
Before this PR, users should redefine the __repr__ and copy&paste the source code from Module.
* Add support for extra information on Module
* Rewrite the repr method of Module
* Fix flake8
* Change the __repr__ to get_extra_repr in Linear
* Fix extra new-line for empty line
* Add test for __repr__ method
* Fix bug of block string indent
* Add indent for multi-line repr test.
* Address review comments
* Update tutorial for creating nn.Module
* Fix flake8, add extra_repr of bilinear
* Refactor DropoutNd
* Change to extra_repr in some Modules
* Fix flake8
* Refactor padding modules
* Refactor pooling module
* Fix typo
* Change to extra_repr
* Fix bug for GroupNorm
* Fix bug for LayerNorm
* support n-d inputs in bilinear and move to aten
* support n-d inputs in bilinear and move to aten
* add asserts to bilinear inputs
* address comments
* cast int64_t in asserts
* Fix some minor errors in existing docs.
* Fix Convolution and Pooling docs in torch.nn.functional
* Cleaned up torch.nn.functional docs
* Address @SsnL 's comments
* Add multiplication sign missing in docs
* Fix more typos, and clear some warnings
* Change infinity symbol in LPPool2d
* Revert some changes in torch.nn.functional
* Few more minor changes
* Improvize documentation
1. Add formula for erf, erfinv
2. Make exp, expm1 similar to log, log1p
3. Symbol change in ge, le, ne, isnan
* Fix minor nit in the docstring
* More doc improvements
1. Added some formulae
2. Complete scanning till "Other Operations" in Tensor docs
* Add more changes
1. Modify all torch.Tensor wherever required
* Fix Conv docs
1. Fix minor nits in the references for LAPACK routines
* Improve Pooling docs
1. Fix lint error
* Improve docs for RNN, Normalization and Padding
1. Fix flake8 error for pooling
* Final fixes for torch.nn.* docs.
1. Improve Loss Function documentation
2. Improve Vision Layers documentation
* Fix lint error
* Improve docstrings in torch.nn.init
* Fix lint error
* Fix minor error in torch.nn.init.sparse
* Fix Activation and Utils Docs
1. Fix Math Errors
2. Add explicit clean to Makefile in docs to prevent running graph generation script
while cleaning
3. Fix utils docs
* Make PYCMD a Makefile argument, clear up prints in the build_activation_images.py
* Fix batch norm doc error
The nn.* counterpart of #5443 . Mostly removed Variable wrapper. Also added doc for nn.RReLU.
Notice that torch.randn(*, requires_grad=True) isn't documented until #5462 is done.
Adds a missing bias term to the __repr__ functions of the
Linear and Bilinear modules. Fixes the spacing in the Conv2d
__repr__ to make it consistent with other modules.
This takes advantage of the broadcasting behavior of torch.matmul to
support inputs with more than two dimensions. The extra dimensions are
treated like part of the batch dimension, much like nn.Bottle in Lua
Torch.
There are a few related small performance changes:
* Addmm computes the gradient in column-major for inputs in
column-major format
* Variable.mm calls Addmm in-place with the desired output buffer
Here's the command I used to invoke autopep8 (in parallel!):
git ls-files | grep '\.py$' | xargs -n1 -P`nproc` autopep8 -i
Several rules are ignored in setup.cfg. The goal is to let autopep8
handle everything which it can handle safely, and to disable any rules
which are tricky or controversial to address. We may want to come back
and re-enable some of these rules later, but I'm trying to make this
patch as safe as possible.
Also configures flake8 to match pep8's behavior.
Also configures TravisCI to check the whole project for lint.
Uses the assignment syntax to get deterministic ordering of parameters.
The ordering of parameters using the constructor syntax is
non-deterministic because kwargs use dict() in Python 3.5 and earlier.
* _forward is renamed forward since users should override it
* some __call__ overrides are changed to forward
* function which return a single variable are changed to return that
variable instead of a one-element tuple
