Signed-Off By: Vivek Khandelwal <[email protected]>

Set PyTorch and TorchVision version to nightly release 2023-01-08.

Signed-Off By: Vivek Khandelw...

* feat(TorchToTOSA): LeakyReLU legalization

* test(LeakyReLU): Add LIT test and enable e2e te...

Credit to @vivekkhandelwal1 for finding the necessary changes.

Summary of changes:

- Swit...

* Added support for aten::norm.ScalarOpt_dim

* Disable NormalizeModule_basic for linalg

- Use v3 of actions/checkout, since the version we use (v2) uses
Node.js 12, which is deprec...

* Fix OptionalCType class name

* Rmove LTC xfail tests

After a recent LLVM tag update, the Windows Release build failed even
though pre- and post-merge...

There are several decompositions that assume the operands of the op
have dtypes available; howe...

This commit replaces the `tanh` dtype function, which was being used
to test the implementation...

Co-authored-by: srirammaswamy <[email protected]>

This reverts commit fa356cce50b418269de5a3fa52e95a337cfc78fd.

This reverts commit eaab9be207c6ada9944f63a71b7686f8f5af6bd1, since it
is causing the post-merg...

Prelu lowering to linalg

This commit adds support for passing to `torch_mlir.compile` the
result of running `torch.jit.t...

Signed-Off By: Vivek Khandelwal<[email protected]>

This reverts commit 50b524546f389b7f6243780f4332e077388dd512.

* Unrevert #1724

* Update pytorch requirements.txt

-- The dtype of the result of `aten.embedding` should match that of
the `weight` operand's (o...

Summary of changes:

- LLVM now includes <optional> instead of "llvm/ADT/Optional.h" in most
...

Now that the RollPyTorch tracker issue exists, we can automate the job
of notifying folks of fa...

pytorch/pytorch@140a3139 reverted a change from yesterday, causing the
RollPyTorch action to br...

Add Complex number dtype support to torch tensors. Add
aten.fft_fft op to test complex numbers.

* Replace asIntArrayRefSlow with macro

* Update pytorch requirements.txt

Functions like `getTypeForScalarType` that do a mapping from one set
of types to another should...

In order to verify if a given IR satisfies the backend contract, the
verifier needs to know if ...

This will require https://reviews.llvm.org/D139121 patch to go through.

As [@ezyang suggested](https://github.com/pytorch/pytorch/issues/90276#issuecomment-1339791275),...

There appear to be two problems with the caching layer in our CI runs:
(a) the sizes of some of...

- Mention the rotation doc
- Fix minor typos / broken link

Summary of changes:

- Replace `llvm::None` with `std::nullopt`, since the former is deprecat...

The RollPyTorch action often takes more than 1.5 hours to finish.
During this time, if another ...

See https://github.com/llvm/torch-mlir/issues/1709

* [custom op] Generalize shape library logic to work with dtypes

This commit generalizes the ...

Currently `getTensorRank` returns -1 if it was unable to get the rank
of the tensor. However, n...

This is the more predominant style in the codebase. I'm sure there are
more in other parts of th...

Signed-Off By: Vivek Khandelwal<[email protected]>

Set PyTorch and TorchVision version to nightly release 2022-12-11.

Signed-Off By: Vivek Khandel...

This reverts commit 55c7e66aa7e3df407652d796280d02144b8ea899.

This was an experimental attempt at rolling out own op-by-op executor
with `__torch_dispatch__`,...

Ensures the buildifier (linter for bazel build files) is run and changes are pushed.

* Extended TorchMLIRLoweringContext with virtual CreateComputation method

* Fix device_data_c...

A circular dependency was introduced in https://github.com/llvm/torch-mlir/commit/e7edcc62fde300...

The current implementation of `DecomposeComplexOps` fails if an op
expected to be decomposed do...

This should run faster and is overall clearer.

This commit removes lit tests from the `decompose-complex-ops` that
are essentially testing a m...

This gives some decent improvements to memory consumption and latency of
testing. I would have e...

This more accurately reflects what it is. The previous name was
conflating the use of RefBackend...

- Support for non-prefixed accessors has been removed. See:
https://reviews.llvm.org/D136727
...

This is a minor variation on our other resnet18 examples swapping in
TorchDynamo.

We replicate ...

This test has been disabled a long time, and since RefBackend is so slow
we don't want to add th...

Signed-Off By: Vivek Khandelwal<[email protected]>