Fixes#144196
Extends #144106 and #144110
## Open Problems:
- [ ] Annotating with `numbers.Number` is a bad idea, should consider using `float`, `SupportsFloat` or some `Procotol`. https://github.com/pytorch/pytorch/pull/144197#discussion_r1903324769
# Notes
- `beta.py`: needed to add `type: ignore` since `broadcast_all` is untyped.
- `categorical.py`: converted `else` branches of mutually exclusive arguments to `if` branch[^2].
- ~~`dirichlet.py`: replaced `axis` with `dim` arguments.~~ #144402
- `gemoetric.py`: converted `else` branches of mutually exclusive arguments to `if` branch[^2].
- ~~`independent.py`: fixed bug in `Independent.__init__` where `tuple[int, ...]` could be passed to `Distribution.__init__` instead of `torch.Size`.~~ **EDIT:** turns out the bug is related to typing of `torch.Size`. #144218
- `independent.py`: made `Independent` a generic class of its base distribution.
- `multivariate_normal.py`: converted `else` branches of mutually exclusive arguments to `if` branch[^2].
- `relaxed_bernoulli.py`: added class-level type hint for `base_dist`.
- `relaxed_categorical.py`: added class-level type hint for `base_dist`.
- ~~`transforms.py`: Added missing argument to docstring of `ReshapeTransform`~~ #144401
- ~~`transforms.py`: Fixed bug in `AffineTransform.sign` (could return `Tensor` instead of `int`).~~ #144400
- `transforms.py`: Added `type: ignore` comments to `AffineTransform.log_abs_det_jacobian`[^1]; replaced `torch.abs(scale)` with `scale.abs()`.
- `transforms.py`: Added `type: ignore` comments to `AffineTransform.__eq__`[^1].
- `transforms.py`: Fixed type hint on `CumulativeDistributionTransform.domain`. Note that this is still an LSP violation, because `Transform.domain` is defined as `Constraint`, but `Distribution.domain` is defined as `Optional[Constraint]`.
- skipped: `constraints.py`, `constraints_registry.py`, `kl.py`, `utils.py`, `exp_family.py`, `__init__.py`.
## Remark
`TransformedDistribution`: `__init__` uses the check `if reinterpreted_batch_ndims > 0:`, which can lead to the creation of `Independent` distributions with only 1 component. This results in awkward code like `base_dist.base_dist` in `LogisticNormal`.
```python
import torch
from torch.distributions import *
b1 = Normal(torch.tensor([0.0]), torch.tensor([1.0]))
b2 = MultivariateNormal(torch.tensor([0.0]), torch.eye(1))
t = StickBreakingTransform()
d1 = TransformedDistribution(b1, t)
d2 = TransformedDistribution(b2, t)
print(d1.base_dist) # Independent with 1 dimension
print(d2.base_dist) # MultivariateNormal
```
One could consider changing this to `if reinterpreted_batch_ndims > 1:`.
[^1]: Usage of `isinstance(value, numbers.Real)` leads to problems with static typing, as the `numbers` module is not supported by `mypy` (see <https://github.com/python/mypy/issues/3186>). This results in us having to add type-ignore comments in several places
[^2]: Otherwise, we would have to add a bunch of `type: ignore` comments to make `mypy` happy, as it isn't able to perform the type narrowing. Ideally, such code should be replaced with structural pattern matching once support for Python 3.9 is dropped.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/144197
Approved by: https://github.com/malfet
Co-authored-by: Aaron Gokaslan <aaronGokaslan@gmail.com>
Fixes#76772, #144196
Extends #144106
- added type annotations to `lazy_property`.
- added type annotation to all `@property` and `@lazy_property` inside `torch.distributions` module.
- added simply type-check unit test to ensure type inference is working.
- replaced deprecated annotations like `typing.List` with the corresponding counterpart.
- simplified `torch.Tensor` hints with plain `Tensor`, otherwise signatures can become very verbose.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/144110
Approved by: https://github.com/Skylion007
This is a new version of #15648 based on the latest master branch.
Unlike the previous PR where I fixed a lot of the doctests in addition to integrating xdoctest, I'm going to reduce the scope here. I'm simply going to integrate xdoctest, and then I'm going to mark all of the failing tests as "SKIP". This will let xdoctest run on the dashboards, provide some value, and still let the dashboards pass. I'll leave fixing the doctests themselves to another PR.
In my initial commit, I do the bare minimum to get something running with failing dashboards. The few tests that I marked as skip are causing segfaults. Running xdoctest results in 293 failed, 201 passed tests. The next commits will be to disable those tests. (unfortunately I don't have a tool that will insert the `#xdoctest: +SKIP` directive over every failing test, so I'm going to do this mostly manually.)
Fixes https://github.com/pytorch/pytorch/issues/71105
@ezyang
Pull Request resolved: https://github.com/pytorch/pytorch/pull/82797
Approved by: https://github.com/ezyang
Summary:
A number of derived distributions use base distributions in their
implementation.
We add what we hope is a comprehensive test whether all distributions
actually honor skipping validation of arguments in log_prob and then
fix the bugs we found. These bugs are particularly cumbersome in
PyTorch 1.8 and master when validate_args is turned on by default
In addition one might argue that validate_args is not performing
as well as it should when the default is not to validate but the
validation is turned on in instantiation.
Arguably, there is another set of bugs or at least inconsistencies
when validation of inputs does not prevent invalid indices in
sample validation (when with validation an IndexError is raised
in the test). We would encourage the implementors to be more
ambitious when validation is turned on and amend sample validation
to throw a ValueError for consistency.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/53600
Reviewed By: mrshenli
Differential Revision: D26928088
Pulled By: neerajprad
fbshipit-source-id: 52784a754da2faee1a922976e2142957c6c02e28
Summary:
Fixes https://github.com/pytorch/pytorch/issues/50496
Fixes https://github.com/pytorch/pytorch/issues/34859
Fixes https://github.com/pytorch/pytorch/issues/21596
This fixes many bugs involving `TransformedDistribution` and `ComposeTransform` when the component transforms changed their event shapes. Part of the fix is to introduce an `IndependentTransform` analogous to `distributions.Independent` and `constraints.independent`, and to introduce methods `Transform.forward_shape()` and `.inverse_shape()`. I have followed fehiepsi's suggestion and replaced `.input_event_dim` -> `.domain.event_dim` and `.output_event_dim` -> `.codomain.event_dim`. This allows us to deprecate `.event_dim` as an attribute.
## Summary of changes
- Fixes `TransformDistribution` and `ComposeTransform` shape errors.
- Fixes a behavior bug in `LogisticNormal`.
- Fixes `kl_divergence(TransformedDistribution, TransformedDistribution)`
- Adds methods `Transform.forward_shape()`, `.inverse_shape()` which are required for correct shape computations in `TransformedDistribution` and `ComposeTransform`.
- Adds an `IndependentTransform`.
- Adds a `ReshapeTransform` which is invaluable in testing shape logic in `ComposeTransform` and `TransformedDistribution` and which will be used by stefanwebb flowtorch.
- Fixes incorrect default values in `constraints.dependent.event_dim`.
- Documents the `.event_dim` and `.is_discrete` attributes.
## Changes planned for follow-up PRs
- Memoize `constraints.dependent_property` as we do with `lazy_property`, since we now consult those properties much more often.
## Tested
- [x] added a test for `Dist.support` vs `Dist(**params).support` to ensure static and dynamic attributes agree.
- [x] refactoring is covered by existing tests
- [x] add test cases for `ReshapedTransform`
- [x] add a test for `TransformedDistribution` on a wide grid of input shapes
- [x] added a regression test for https://github.com/pytorch/pytorch/issues/34859
cc fehiepsi feynmanliang stefanwebb
Pull Request resolved: https://github.com/pytorch/pytorch/pull/50581
Reviewed By: ezyang, glaringlee, jpchen
Differential Revision: D26024247
Pulled By: neerajprad
fbshipit-source-id: f0b9a296f780ff49659b132409e11a29985dde9b
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/18598
ghimport-source-id: c74597e5e7437e94a43c163cee0639b20d0d0c6a
Stack from [ghstack](https://github.com/ezyang/ghstack):
* **#18598 Turn on F401: Unused import warning.**
This was requested by someone at Facebook; this lint is turned
on for Facebook by default. "Sure, why not."
I had to noqa a number of imports in __init__. Hypothetically
we're supposed to use __all__ in this case, but I was too lazy
to fix it. Left for future work.
Be careful! flake8-2 and flake8-3 behave differently with
respect to import resolution for # type: comments. flake8-3 will
report an import unused; flake8-2 will not. For now, I just
noqa'd all these sites.
All the changes were done by hand.
Signed-off-by: Edward Z. Yang <ezyang@fb.com>
Differential Revision: D14687478
fbshipit-source-id: 30d532381e914091aadfa0d2a5a89404819663e3
Summary:
This PR:
- adds a `.expand` method for `TransformedDistribution` along the lines of #11341.
- uses this method to simplify `.expand` in distribution classes that subclass off of `TransformedDistribution`.
- restores testing of `TransformedDistribution` fixtures.
- fixes some bugs wherein we were not setting certain attributes in the expanded instances, and adds tests for `.mean` and `.variance` which use these attributes.
There are many cases where users directly use `TransformedDistribution` rather than subclassing off it. In such cases, it seems rather inconvenient to have to write a separate class just to define a `.expand` method. The default implementation should suffice in these cases.
cc. fritzo, vishwakftw, alicanb
Pull Request resolved: https://github.com/pytorch/pytorch/pull/11607
Differential Revision: D9818225
Pulled By: soumith
fbshipit-source-id: 2c4b3812b9a03e6985278cfce0f9a127ce536f23
Summary:
This adds a `.expand` method for distributions that is akin to the `torch.Tensor.expand` method for tensors. It returns a new distribution instance with batch dimensions expanded to the desired `batch_shape`. Since this calls `torch.Tensor.expand` on the distribution's parameters, it does not allocate new memory for the expanded distribution instance's parameters.
e.g.
```python
>>> d = dist.Normal(torch.zeros(100, 1), torch.ones(100, 1))
>>> d.sample().shape
torch.Size([100, 1])
>>> d.expand([100, 10]).sample().shape
torch.Size([100, 10])
```
We have already been using the `.expand` method in Pyro in our [patch](https://github.com/uber/pyro/blob/dev/pyro/distributions/torch.py#L10) of `torch.distributions`. We use this in our models to enable dynamic broadcasting. This has also been requested by a few users on the distributions slack, and we believe will be useful to the larger community.
Note that currently, there is no convenient and efficient way to expand distribution instances:
- Many distributions use `TransformedDistribution` (or wrap over another distribution instance. e.g. `OneHotCategorical` uses a `Categorical` instance) under the hood, or have lazy parameters. This makes it difficult to collect all the relevant parameters, broadcast them and construct new instances.
- In the few cases where this is even possible, the resulting implementation would be inefficient since we will go through a lot of broadcasting and args validation logic in `__init__.py` that can be avoided.
The `.expand` method allows for a safe and efficient way to expand distribution instances. Additionally, this bypasses `__init__.py` (using `__new__` and populating relevant attributes) since we do not need to do any broadcasting or args validation (which was already done when the instance was first created). This can result in significant savings as compared to constructing new instances via `__init__` (that said, the `sample` and `log_prob` methods will probably be the rate determining steps in many applications).
e.g.
```python
>>> a = dist.Bernoulli(torch.ones([10000, 1]), validate_args=True)
>>> %timeit a.expand([10000, 100])
15.2 µs ± 224 ns per loop (mean ± std. dev. of 7 runs, 100000 loops each)
>>> %timeit dist.Bernoulli(torch.ones([10000, 100]), validate_args=True)
11.8 ms ± 153 µs per loop (mean ± std. dev. of 7 runs, 100 loops each)
```
cc. fritzo, apaszke, vishwakftw, alicanb
Pull Request resolved: https://github.com/pytorch/pytorch/pull/11341
Differential Revision: D9728485
Pulled By: soumith
fbshipit-source-id: 3b94c23bc6a43ee704389e6287aa83d1e278d52f
