Summary:
Meant to make tasks like #53728 easier. The `-n` flag enables line numbers, and the `-o` flag reduces noise by only showing the part of the line that matched (which in this case is just the trailing whitespace).

Pull Request resolved: #53733

Test Plan:
```
$ git checkout e937db5
```

Before:
```
$ (! git grep -I -l ' $' -- . ':(exclude)**/contrib/**' ':(exclude)third_party' || (echo "The above files have trailing spaces; please remove them"; false))
aten/src/ATen/native/cuda/BatchLinearAlgebra.cu
The above files have trailing spaces; please remove them
```

After:

```
$ (! git grep -I -no ' $' -- . ':(exclude)**/contrib/**' ':(exclude)third_party' || (echo "The above files have trailing spaces; please remove them"; false))
aten/src/ATen/native/cuda/BatchLinearAlgebra.cu:1972:
The above files have trailing spaces; please remove them
```

Reviewed By: mruberry

Differential Revision: D26953538

Pulled By: samestep

fbshipit-source-id: 5f7d48b79f1a02e5e5a09fe00316ec350cfc340e

Summary:
Pull Request resolved: #51674

See
https://docs.python.org/3/library/importlib.html#importlib.abc.ResourceReader

Test Plan: Imported from OSS

Reviewed By: zdevito

Differential Revision: D26237034

Pulled By: suo

fbshipit-source-id: 4c19f6172d16b710737528d3de48372873b9368d

Summary:
Pull Request resolved: #51676

We offer the ability to access the importer from within packaged modules by doing
`import resources`. This behavior is nice (and more powerful than the
importlib resources API), but I think `resources` is too common a name
(pip has a package for it)

Change to `import torch_package_importer` but open to bikeshedding

Test Plan: Imported from OSS

Reviewed By: jamesr66a

Differential Revision: D26620314

Pulled By: suo

fbshipit-source-id: 0942c99f02c0f55f5f3a1b2566961018b796bdd4

Summary:
ref: #16574

Pull Request resolved: #52408

Reviewed By: anjali411

Differential Revision: D26654687

Pulled By: malfet

fbshipit-source-id: 6feb603d8fb03c2ba2a01468bfde1a9901e889fd

…ultiprocessing to torch/distributed/elastic (#53574)

Summary:
Pull Request resolved: #53574

Upstreams `torchelastic/timer|multiprocessing` to `torch/distributed/elastic/timer|multiprocessing`

Test Plan:
```
buck test mode/dev-nosan //caffe2/torch/distributed/elastic/...
buck test mode/dev-nosan //caffe2/test/distributed/elastic/...
buck test mode/dev-nosan //pytorch/elastic/torchelastic/...
buck test mode/dev-nosan //hpc/...
buck test mode/dev-nosan //caffe2/torch/fb/training_toolkit/...
```

Reviewed By: borovsky-d, wilson100hong

Differential Revision: D26899809

fbshipit-source-id: e6dbc2a78282eac296c262b3206a979e3ef1ff53

)

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: #53600

Reviewed By: mrshenli

Differential Revision: D26928088

Pulled By: neerajprad

fbshipit-source-id: 52784a754da2faee1a922976e2142957c6c02e28

Summary:
#52875 introduced this bug, as `supports_tensor_out` was replaced with `supports_out` in #53259, so CI checks are failing.

Pull Request resolved: #53745

Reviewed By: gmagogsfm

Differential Revision: D26958151

Pulled By: malfet

fbshipit-source-id: 7cfe5d1c1a33f06cb8be94281ca98c635df76838

Summary:
Pull Request resolved: #53665

ngimel pointed out to me where we already test the behavior of the `Upsample` ops in `test_nn.py`. This PR deleting my bespoke tests in `test_torch.py` and updates those in `test_nn.py` to test memory format properly.

There were two reasons the original test didn't pick up on a memory format regression:
- They didn't test the memory format of the output tensor explicitly, i.e. `output.is_contiguous(memory_format=...)`
- Even with that change, the test tensors were to simple to fail the tests. From some trial and error, it looks like one of the first two dimensions in the inputs needs to be > 1 in order for the `channels_last` memory format to actually re-order the strides.

Test Plan: Imported from OSS

Reviewed By: ngimel

Differential Revision: D26929683

Pulled By: bdhirsh

fbshipit-source-id: d17bc660ff031e9b3e2c93c60a9e9308e56ea612