Splat operators required the output is 1-D. This was not a required
restriction and was loosene...

This commit provides dummy solution to support elmentwise operations
(mul, add) with scalar arg...

Reduce mean lowerings did not succesfully lower to `linalg` via torched.
There were two separat...

* torch-mlir may not handle `aten.warn`. But it could be handled by
custom users' backend which...

If the broadcast shape is length-1 at a dim while `?` in the input dim
then we need to broadcas...

This folds small version of the tensor-scalar comparison operators as
they are commonly used fo...

uses version checking since attributes exist in both versions, the only
thing that changes is w...

The file was renamed by PR https://github.com/llvm/torch-mlir/pull/2842.

Previous gemm version required a bias vector.
This provides an alternate path to `Torch::AtenM...

The current padding operation was not functional for dynamic shapes.
Updated and enabled tests ...

Spec: https://onnx.ai/onnx/operators/onnx__Clip.html

Set PyTorch and TorchVision version to nightly release 2024-03-07.
This commit also removes the...

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

We can support `onnx.Size` by requesing the size of each dimensions and
taking the product of t...

This mostly copy-pastes the reduce minimum implementation to reduce max
to improve test coverag...

The addition of an e2e test is actually provided in the Shark-Testsuite.
This adds 2 test cases...

Current implementation depends on using `aten.view` which has issues
inferring tensor collapse/...

Add e2d support for `aten.linalg_norm` by decompose it to
`aten.linalg_vector_norm`.

Lowerin...

Convert to individiual slices and tuple together as a list.

---------

Co-authored-by: Scot...

`getRawBuffer` expects a densely packed vector of `i1` values however
`onnx` does not densely p...

* that `dataclass`'s `slots` is supported after python 3.10.

A handful of operations are commonly used in shape calculations (slice,
concat, broadcast). Add...

Alternatives to https://github.com/llvm/torch-mlir/pull/2908

Fix https://github.com/nod-ai/SH...

Existing lowering via aten.view does not work as well for dynamic shapes
as the lowering to ten...

We assumed rank was atleast 1 however it can be rank-0, generating an
illegal pair of flatten /...

If an input is length-0 along the dimension of canonicalization we can
remove the tensor from t...

Collapsing to a rank-0 tensor using `aten.view` was currently bailing
out. Added the special case.

The type returned was changed in
https://github.com/llvm/torch-mlir/pull/2795. This led to erro...

This commit adds the support for scalar conversion to byte.
This commit also fixes the OnnxToL...

Closing https://github.com/nod-ai/SHARK-Turbine/issues/351

This is the lowering of gridsampler from onnx to torch using our prior
implementation of AtenGr...

Finish supporting importing the vast majority of `onnx` operations. This
includes:
- region su...

A bunch of small fixes are interlinked and trigger crashes if not
addressed as a group. This in...

The decomposition only suports a NCHW lowering however the operation can
support arbitrary spat...

Currently lowering uses `tensor.from_elements` which does not allow zero
inputs. In this case w...

The corrective transpose at the end is computed incorrectly. Is it
actually computin the invers...

Torch lowering only supported the most recent version. Refactored the
lowering so more easily h...

Added Support for float dtype in in torch.arange in TOSA Dialect

This resolves the following ...

This required adding a "decompose" pass to the torch lowering, since
torch.mv was not directly ...

We collapsed and broadcasted scatter indices to a single element
version. We should instead upp...

Fixes https://github.com/nod-ai/SHARK-Turbine/issues/461

Signed-Off By: Vivek Khandelwal <viv...

Fixes https://github.com/nod-ai/SHARK-Turbine/issues/451,
https://github.com/nod-ai/SHARK-Turbi...

As titled.

Closes
[nod-ai/SHARK-Turbine#365](https://github.com/nod-ai/SHARK-Turbine/issues/365)

This was found in a downstream that is pegged to an older PyTorch
version.

`mpmath 1.4.0` changes some import locations breaking `torch`. Changing
to `1.3.0` to avoid bre...

Also note that we are in the process of proposing SparseTensorMetadata
to PyTorch FX graph expo...

Lowers `torch.grid_sample` to the equilvalent `linalg` representation.

Set PyTorch and TorchVision version to nightly release 2024-02-20.

Signed-Off By: Vivek Khand...

There is no reason to treat `ConstantOfShape` as a specialized import
any as there exists a onn...

Strided slicing can occur with a negative stride. In these cases we need
to bound end different...

According to the [official TOSA
spec](https://www.mlplatform.org/tosa/tosa_spec.html#_cast), `t...

Crashing tests no longer crash, enable as either passing or xfail tests.

Co-authored-by: Xida...

Some operands are optional so we require a placeholder for missing
operands. We invent an `onnx...

This enables better re-use in downstreams which use different func
implementations and should h...

Useful for `slice` lowerings that depend on tensors made form scalars.

Simple folder for limited size aten tensor operations. This is primarily
useful for shape compu...

Onnx slice lowering used arange needlessly instead of directly
constructing the constant dimens...

For some reason we did not directly use the step size dynamically
despite its constructed using...

https://github.com/nod-ai/SHARK-Turbine/issues/327

Even though the reference compiler is not about performance, inlining
the generated sparse help...

Sample compilation crashes due to sigmoid with integer inputs/outputs.
This fix avoids crashing...

Onnx's gather maps directly to `torch.aten.index_select`. We should just
use that path.

We are crashing due to indexing into a negative shape. Updated the
lowering to avoid the crash.

grammar police

As of https://github.com/pytorch/pytorch/pull/118969, `ExportedProgram`
has the long awaited fi...

We can route the torch tests via `onnx` using the `torch.onnx.export`
tooling. We can then reim...

This commit adds decomposition support into the core aten operators
before importing the module...

Just a minor update to skip the space char if included in the equation
string

---------

C...

this introduces an additional soft dependency on the python ml_dtypes
python packages in order ...

This version of pytorch includes a patch to enable dynamo support on
Windows, so I would like t...

This commit adds the OnnxToTorch lowering for cosh, acosh, asin, asinh,
and atanh op.
This com...

Some operations include a backend matcher for specialized operations. We
map these back to gene...

Fixes https://github.com/llvm/torch-mlir/issues/2866

Some backends / downstream projects expe...

This test exposes issues that need fixing
(1) propagate sparsity into the FX graph (over elt-wi...

The investigation is largely recorded in
https://github.com/llvm/torch-mlir/pull/2881, but this...

This commit adds the OnnxToTorch support for Mean, IsInf, IsNaN, and
PRelu ops. All high priori...

Various improvements on sparsity metadata:

(1) define single data structure for all sparsity ...

This patch makes the Protobuf package mandatory in addition to forcing a
config mode search. T...

This PR introduces a sparse_jit wrapper that can run simple models with
sparse tensor inputs en...

By updating convertScalarToDtype invocation pass original source and
destination datatypes for ...

…dcc7e89.
* to involve `chlo-legalize-to-stablehlo` pass.

Added a folded for a equals operator. This allows an equivalent
comparison folder, primarily fo...

Includes some minor first for `AffineMap::inferFromExprList`

If the input or result is a splat value we can just constant fold the
result. This is common fo...

The lowering decomposes AtenTraceOp into an AtenDiagonalOp followed by
AtenSumOp.

The progre...

There is no lowering support for math::AbsIOp, so if the operand is an
integer type, it will fa...

This adds a few passes that will ensure linalg with sparse tensors are
properly lowered to loop...

This PR contains three commits to update the validation checks in the
ONNX -> Torch conversion ...

This commit modifies the OnnxToTorch lowering of Onnx.Reshape op by
creating the result shape l...