This is the first step to making the e2e framework convenient to use
by downstream backends.

This commit adds support for the cases of view op where the rank and
the shapes of the input and...

This commit adds decomposition of `aten.zero_` op.

Signed-Off-by: Gaurav Shukla <gaurav@nod-lab...

`albert-base-v2` for sequence classification is added in e2e_heavy_test.

In order to make sure that the TorchToLinalg conversions leave the
graph in a valid state, the ...

* fix type inference
* fix Torch2Linalg conversion
* add test cases

Effectively, this mode works by compiling op by op as the NN is eagerly executed by PyTorch. Ent...

Torch FX was never really a different path, since all FX modules are
actually valid TorchScript ...

- This commit adds decomposition of `aten.dropout` op. It also covers the
training mode of the...

The `assemblyFormat` stuff (which generates unrolled, per-op C++ code)
was taking up a lot of co...

This commit adds the op `ValsemVariantAtenCopyOp` that represents
`AtenCopy_Op` without the unde...

This commit decomposes `aten._to_copy` op into
`valsem.aten.copy` op.

Signed-Off By: Vivek Khan...

This commit decomposes `aten.expand_as` op into `aten.broadcast_to` op.

Signed-Off By: Vivek Kh...

This commit adds support for type refinement when
`torch.tensor_static_info_cast`s are involved...

This commit fixes the 2nd and 3rd return types of the `aten.native_layer_norm`.
Previously the m...

The ODS-generated code included via the `TorchOps.cpp.inc` file takes a
very long time to compil...

This commit decomposes `aten.index_put` op into
`valsem.aten.index_put_impl` op.

Signed-Off By:...

This commit adds the op `ValsemVariantAtenIndexPutImplOp` that represents
`Aten_IndexPutImpl_Op`...

This commit adds value semantics support for ops such as
`aten.view_as` and `aten.expand_as` th...

The term "pseudo" is very vague and was getting confusing (I felt I had
to explain it in every c...

This is code that we always want to treat as "foreign" and not get too
comfortable using in many...

This leads to much more succinct types in many cases:

```
!torch.list<!torch.int>
!torch.list<i...

This was an aspirational goal at an earlier stage in the project where
the focus was heavily on ...

See the documentation in `docs/shape_lib.md` and
`docs/adding_a_shape_function.md` for an overvi...

This helps keep things organized and also exposes more parallelism to
the build system. It seems...

The location of packed_params.h is changed in aten src.

This commit replaces the two rewrite patterns of
maximize-value-semantics with a single pattern...

Signed-Off-By: Prateek Gupta <[email protected]>

This commit moves the helper function which are common across
different torch-mlir conversion pa...

This pass is added to lower ops, which can not be lowered
via the TorchToLinalg pass, such as `t...

This commit adds support for integer type inputs for
`AtenMaxOp`, `AtenSumOp`, `AtenSumDimIntLis...

- This commit adds E2E support for `aten.rand_like` and
`aten.bernoulli_.Tensor` ops.
- The `a...

This commit decomposes `aten.full_like` op into `aten.empty_like`
and `aten.fill` op.

Signed-Of...

This commit decomposes `aten.full` op into `aten.empty` and
`aten.fill` op.

Signed-Off By: Vive...

This commit adds static type information support to `aten.bmm`. This
is needed for the forward ...

The pass is mostly borrowed from the BufferizeAnyLinalgOp pass in mlir
upstream with some minor ...

This commit adds a lowering for `TensorStaicInfoCastOp` that simply
replaces the op with the `t...

The view op allows for the new shape argument to have a -1 value for
one of the dimensions, and...

Decomposition of aten.silu.op is added as silu(x) = x * sigmoid(x).

This commit adds lowering of `aten.new_zeros` and `aten.new_ones` op

Signed-Off By: Vivek Khande...

This commit adds support for the contiguous memory format for the ops
`AtenCloneOp` and `AtenEm...

This commit does a couple of things. First, it fixes a bug in the
`linalg.generic` body of the ...

Its unclear to me what the right layering is here: Are you expecting
torch-mlir-dialects to alwa...

This should be set elsewhere depending on the build configuration.
In particular, we need to be...

`aten.hardswish` op is decomposed into (x/6) * Relu6(x+3).

This reverts commit 442ff4605c54f836afa17d2780bd9bddcc2a5ed8.

This commit adds the invariant to the op `torch.overwrite.tensor.contents` that
both of its ope...

There are cases where the op may update the same indices multiple
times. In this context, we can...

* Fix setup.py backwards compatibiity

* Remove version check

This commit adds handling to the `maximize-value-semantics` pass for
the case where a view-like...

This commit adds static type information support to `aten.mm`. This is
needed for the forward p...

New str argument approximation is added.

The check for `self.hasSizes` was missing before performing `.size()`
operation.

Major changes: opTrait changed to Trait, selectOp moved to arith dialect
assertOp moved to cf dia...

- This commit decomposes the `aten.batch_norm` op into the
`aten.native_batch_norm` op, instea...

Signed-off-by: Anup Gangwar <[email protected]>

Co-authored-by: Anup Gangwar <anup.gangwar...

The external projects torch-mlir and torch-mlir-dialects should be
placed inside double quotes.

`aten.log_softmax` is decomposed to be more numerically stable.

This is intended to explore support for non-structured ops that can't
be modeled by Linalg diale...

This commit adds lowering of `aten.Hardsigmoid` op.

Signed-Off-by: Gaurav Shukla <gaurav@nod-la...

This commit adds the op `PseudoAtenFillScalarOp` that represents
`AtenFill_ScalarOp` without th...

- This commit adds lowering of `aten.le.Scalar` and `aten.ge.Scalar` ops
as a part of `convert...

This commit adds the op `PseudoAtenBernoulliFloatOp` that represents
`AtenBernoulli_FloatOp` wi...

- This commit adds lowering of `aten.eq.int` op as a part of
`convert-torch-to-std` pass.
- It...

- This commit adds lowering of `aten.Bool.Tensor` and
`aten.Float.Tensor` op as a part of `con...

- The `self` name is being used as a keyword argument to the
`torch.ops.aten.nll_loss_backward...

Some of the lowerings use the result type obtained from the op itself
to tell the `linalg::Gene...

Prior to this commit, importing a `prim::Constant` node with list type would result in an error ...

* [tosa] Support for AtenNe[Tensor|Scalar]Op, AtenLog2Op,
AtenBitwiseAndTensorOp, AtenSquareOp ...

This commit fixes an error in the refine types pass of constant
allocation ops. The function us...

aten.bernoulli is decomposed to aten.gtTensor(aten.uniform(x), x).

Decompose _log_softmax into log(softmax(x)).

This commit adds decomposition of `aten._unsafe_view` op into
`aten.view` op.

Signed-Off-By: Pr...

This commit adds support for the aten.clone op.

- This commit adds `aten.assert` op in the Torch dialect.
- The `aten.assert` op is lowered to `...

Also change the 3rd dimension to be smaller so that CI can pass without
killing the process.

- This commit adds support for `aten.native_batch_norm` operation.
- The current implementation ...

Conversion of torch.bool tensor type to float and integer type is
handled.

* [tosa] Support for AtenPermuteOp

Signed-off-by: Anup Gangwar <[email protected]>

The lowering of aten::nll_loss_backward op has been added
from torch to linalg dialect. The chan...

The softmax decomposition is modified according to https://github.com/pytorch/functorch/blob/mai...

This commit decomposes `aten.reshape` into `aten.view` op in the case of
value tensor type opera...

Handles Linear, Adaptive_AvgPool2D and FlattenUsintInts
Adds ResNet18 static model for TOSA

...

This PR include the following pieces:
- Add torch `Generator` type. `Generator` type is converte...

Signed-off-by: Suraj Sudhir <[email protected]>

Lowering of `aten.max.dim` op has been added.