Ecosyste.ms: OpenCollective
An open API service for software projects hosted on Open Collective.
github.com/llvm/torch-mlir
The Torch-MLIR project aims to provide first class support from the PyTorch ecosystem to the MLIR ecosystem.
https://github.com/llvm/torch-mlir
99fb4c8637272cd2714f795c0bbf1972becba707 authored about 2 years ago by 武家伟 <[email protected]>
* Fix OOT LTC CI build failure
* Disable LTC during macOS package gen
* Add more details a...
ba17a4d6c09b4bbb4ef21b1d8d4a93cb056be109 authored about 2 years ago by Henry Tu <[email protected]>65d811e26779f48bdd5e0c790aab18ff8221be93 authored about 2 years ago by Vivek Khandelwal <[email protected]>
Lint all Bazel files (BUILD, WORKSPACE, .bzl, or .sky).
Steps followed:
```shell
./utils/ba...
[MHLO] Eliminate explicit dynamic output shape generating in converting AtenSliceTensorOp
7bd173a1c46046be99e8bd042fd18513f0215d8c authored about 2 years ago by 武家伟 <[email protected]>0d1aa437641d2fc2d15ce299a61c940da3f3e7bc authored about 2 years ago by powderluv <[email protected]>
This makes it easier to read in dark mode browsers.
f601435fdf2fa8df2dd5518564d6e0f4f2f959cc authored about 2 years ago by Sean Silva <[email protected]>
One of the simplifications made by the pass `RefinePublicReturn`
currently only happens if the ...
This attempts to get out of my head most of the critical layering and
project structure decision...
This is a first step towards formalizing the set of ops in our backend
contract. The goal is to ...
I recently fixed the handling of the `dim` argument in
`sum_mean_dim` (https://github.com/pytor...
Pass is introduced in [this commit](https://github.com/llvm/torch-mlir/commit/57681f794764a34c34...
7d4a0d0e2b65c7ce8de19993f3b10ad5344fe32b authored about 2 years ago by Sambhav Jain <[email protected]>This fixes a seeding issue with the [previous PR](https://github.com/llvm/torch-mlir/pull/1240) ...
114f48e96c578ee76a6f83b3aa4aa229a8d5b76e authored about 2 years ago by Sambhav Jain <[email protected]>
This introduces a new pass LowerToBackendContract (better name very
welcome) which performs the ...
This PR adds:
- A minimal docker wrapper to the bazel GHA workflow to make it reproducible lo...
9c8b96272057f4f8210de5842b6952228434cfa2 authored about 2 years ago by Sambhav Jain <[email protected]>This reverts commit c9357950864d84ab9a2adf10b2aaa15fec6196e9.
9be89975367176615e8b1d1799777757864ecd78 authored about 2 years ago by Yan Xu <[email protected]>
Bumps the shape library:
- Updates the function signature for aten.arange.start_step
- upstr...
* [MHLO] Add AtenRSubScalarOp conversion pattern
Co-authored-by: Bairen Yi <yibairen.byron@byte...
fde390c7669e29362b18388448ef2b188713383f authored about 2 years ago by nithinsubbiah <[email protected]>
* Propagate device data names
* Address PR comment
* Add example usage
* Add test for d...
0af55781aede053f32d2b618f0f6d36d875fd908 authored about 2 years ago by Jae Hoon (Antonio) Kim <[email protected]>
Summary of changes:
- Tensor dialect now sets `emitAccessorPrefix` to prefixed, thus
requr...
Generalize canonicalization pattern for more sub/div/mul/add op, but for AtenDivTensorModeOp in ...
3b3cb99ef88d0e6430f21764c35afd82902acbd5 authored about 2 years ago by 武家伟 <[email protected]>
According to the documentation for
`torch.embedding_bag` (https://pytorch.org/docs/stable/gener...
c9357950864d84ab9a2adf10b2aaa15fec6196e9 authored about 2 years ago by Yan Xu <[email protected]>
Fix remaining instances of `external/llvm-project`.
41aa562fb48aa11f1d30badf278cc9fb00ff80bd authored about 2 years ago by Sambhav Jain <[email protected]>606f4d2c0ed2843e6dd2584dc287833968877527 authored about 2 years ago by Ashay Rane <[email protected]>
We're seeing large CI queue times ([example](https://discord.com/channels/636084430946959380/742...
34478ab1c702bc7e229708eb5fc6b0b1c6aada70 authored about 2 years ago by Sambhav Jain <[email protected]>
When we renamed the directory containing submodules from `external` to
`externals`, we accident...
My earlier[ PR](https://github.com/llvm/torch-mlir/pull/1213) had (among other things) decoupled...
aed0ec3a2c3d32589cbb048fc340dfb57042dbe7 authored about 2 years ago by Sambhav Jain <[email protected]>b8bd0a46cc4b3208b683445ecd216024e90bfafc authored about 2 years ago by Sambhav Jain <[email protected]>
Addresses https://github.com/llvm/torch-mlir/issues/1207.
#### Provisioned jobs:
```
# ubu...
Building on a fresh environment + virtualenv + in-tree build errors out
becayse PyYaml isn't in...
`aten.masked.tensor` op has been decomposed to `aten.masked.scalar` op.
b1a506624ce64a8f3ae2f878d5d35cd3f0dfae1d authored about 2 years ago by Prashant Kumar <[email protected]>d96ec64be17fb123bb1c37c41d6a22776591a4bc authored about 2 years ago by Yan Xu <[email protected]>
dd2da5a038e67cdd16b8ec4eef8fbdcaf661deb0 authored about 2 years ago by Vidush Singhal <[email protected]>
79b9cf9468e44c2a46c537ef14018f0554b788cc authored about 2 years ago by gpetters94 <[email protected]>
The CI didn't catch the missing rename of TorchToArith until the
merge had happened. This is fo...
Adding an example on how to extract MLIR output from the compilation
process in various differe...
All the converters in this pass appear to create ops from the
arith dialect. Hence the full ren...
* mac m1 cross compile
Add support for M1 cross compile
* Remove redundant ExecutionEngine...
2342456356e0bb3e1b63547764062335f4ddd6ea authored about 2 years ago by powderluv <[email protected]>
Co-authored-by: Bairen Yi <[email protected]>
Co-authored-by: Jiawei Wu <xremold@gma...
9cf0b6e8ff4ed2400175475c7ca7694688cf17f1 authored about 2 years ago by powderluv <[email protected]>
The torch-mlir bazel build is [failing](https://github.com/llvm/torch-mlir/runs/7737425906?check...
072c2b5aafdd5858bb19c5b296453a9cf0b65113 authored about 2 years ago by Sambhav Jain <[email protected]>At the moment we don't gate torch-mlir PRs with bazel builds. This means bazel builds don't get ...
d41c7becf5c759721b829155e0de1bf44f1c7499 authored about 2 years ago by Sambhav Jain <[email protected]>b696362b7d2a5ab1a6c2ea8d93ff452c9ea473e3 authored about 2 years ago by Sambhav Jain <[email protected]>
LLVM now uses C++17.
e75c7c5292423913624756be6ddea71025d42aca authored about 2 years ago by Jacques Pienaar <[email protected]>
With llvm/llvm-project@112499f landed, `MLIR_TABLEGEN_EXE` is given as a
cache variable in the ...
The Torch dialect has an include to `mlir/Dialect/Func/IR/FuncOps.h` and
should therefore have ...
fixed lowering failed on torch::max.dim while shape type is i32
f83a90585682c25367565fe8d612dd600e27ee04 authored about 2 years ago by Yan Xu <[email protected]>This reverts commit 34e207eeb502109957bebee6654709bb89f2d026.
e55fc4deb501bc2f52d3b7ab001dc9dc19a817a9 authored about 2 years ago by powderluv <[email protected]>
Summary of changes:
- Switch to C++17 (similar to https://reviews.llvm.org/D131348)
- Update...
This reverts commit f85ae9c685d6c9f2d43fe60fdad2ad5d3183ec52.
3e97a33c80a85315122c56ac9dc6183d1e41998f authored about 2 years ago by Henry Tu <[email protected]>
Co-authored-by: Bairen Yi <[email protected]>
Co-authored-by: Jiawei Wu <xremold@gma...
Rather than a per-global-slot initializer region, we now have one for
the whole module. For exam...
* E2E support for AtenRemainderScalarOp
34e207eeb502109957bebee6654709bb89f2d026 authored about 2 years ago by Vidush Singhal <[email protected]>* Add decomposition and E2E support for Aten_EmbeddingBag
b70548edff7f55d5fa8335bf5af67f3d1ba0ba8f authored about 2 years ago by Vidush Singhal <[email protected]>f85ae9c685d6c9f2d43fe60fdad2ad5d3183ec52 authored about 2 years ago by Henry Tu <[email protected]>
follow up #761:
This patch updates the `torch_mlir::convertTensorToMlirElementsAttr()`
...
* [MHLO] fix tensor mode aten.div op pattern
See RFC #999
Co-authored-by: Bairen Yi <yibaire...
5618890ca05db7336022f02224022db250a8e420 authored about 2 years ago by Sean Silva <[email protected]>
1fdaf2faa0de0e177eb36da742e66686cc25ebac authored about 2 years ago by Sean Silva <[email protected]>
* Update CMakeLists.txt
* Update CMakeLists.txt
* Update CMakeLists.txt
* Update CMakeL...
e322f6a8784009b37aa354abfa9a40a80f30877d authored about 2 years ago by Henry Tu <[email protected]>
This helps to triage issues related to backends that don't support all
dtypes.
PyTorch recently added support for `dim=None` in the `torch.var`
(https://github.com/pytorch/pyt...
In some cases, users know that a traced graph is valid for a wider set
of shapes than they origi...
To test the summation stability of `torch.aten.var`, add a large
constant to it, which increases...
Co-authored-by: Bairen Yi <[email protected]>
Co-authored-by: Jiawei Wu <xremold@gma...
08fc2d89bbda3ca160bc7e7fe73ac97f1d18be94 authored about 2 years ago by gpetters94 <[email protected]>
* [MHLO] Init MHLO pooling-like op conversion and remove 'op' suffix in filenames
Co-authored...
d030591df99ce19f14e3ec0d982b530a42e1d8bb authored about 2 years ago by 武家伟 <[email protected]>See RFC https://github.com/llvm/torch-mlir/issues/999
Co-authored-by: Bairen Yi yibairen.byro...
f0a24f59f6354df5a254779caaaaaf6918a4f038 authored about 2 years ago by Tanyo Kwok <[email protected]>48ec3005865cb55b6e70a85ccf67f5e63d3c5b83 authored about 2 years ago by Ahmed S. Taei <[email protected]>
37a229cffc86be9ff50f40252aedf22fcbb2085d authored about 2 years ago by powderluv <[email protected]>
This should enabling better caching of builds.
0d25b6f10efe47b567d1a2c6fe884f77947bc51e authored about 2 years ago by powderluv <[email protected]>
This commit fixes the CI failure by temporarily adding the failing
test to xfail set.
Signed-Of...
f2a0e321276f5acc5f9fe91a4c6ca20f1365d82e authored about 2 years ago by Vivek Khandelwal <[email protected]>
* [MHLO] init reduce-like op conversion from Torch to MHLO
Co-authored-by: Bairen Yi <yibairen....
See RFC https://github.com/llvm/torch-mlir/issues/999
Co-authored-by: Bairen Yi yibairen.byro...
0b23af27d338fa005e2bd1c93a61e88b40edca23 authored about 2 years ago by Tanyo Kwok <[email protected]>82af44da2fbc701eddbcde4116259460bf483dd6 authored about 2 years ago by Ahmed S. Taei <[email protected]>
PyTorch recently added support for `dim=None` in the `torch.mean`
op (https://github.com/pytorc...
- Includes math-to-libm pass in refbackend for math::atan2 support
38d8498b21d0892c2c7196aeaab64aeec6852a62 authored about 2 years ago by Quinn Dawkins <[email protected]>add aten::gelu op pattern, and moved some unit tests from basic.mlir to elementwise.mlir
704efdc259a19a162b7ea0dbeb7399b72f0bebda authored about 2 years ago by Yan Xu <[email protected]>
* [MHLO] Support for dynamic shape in basic op conversion by introducing CHLO dialect
Co-author...
See RFC: https://github.com/llvm/torch-mlir/issues/999
Co-authored-by: Bairen Yi yibairen.byr...
3772e0bd91aa84da4acad73a1de065a62e8f8e57 authored about 2 years ago by Tanyo Kwok <[email protected]>fe3c9f5208ee1f77bdd40807c77ae64ed344dcc9 authored about 2 years ago by Vidush Singhal <[email protected]>
ed13ebfd8dae0d311208eea89826ee77e29271cc authored about 2 years ago by Vidush Singhal <[email protected]>
554570f3abddbfb6dde9e546ea247de237b282d4 authored about 2 years ago by Alec <[email protected]>
30017af5b54b6eb628786892727aa14e190eaff7 authored about 2 years ago by powderluv <[email protected]>
As part of this update, this patch also updates MHLO tag to ad54b43.
55bbbec144334f77bece3595c7cbc9875b0be016 authored about 2 years ago by Ashay Rane <[email protected]>* Fix autogen build dir issue
* Got functionalization pass to compile
* Add slice/diagonal...
fb21c9e6cbf4788c121b76de08a4085381b71dc9 authored about 2 years ago by Jae Hoon (Antonio) Kim <[email protected]>* Blacklist _convolution op in LTC
* Removed duplicate Torch_AtenSelectScatterOp instance fro...
cec74b8d375f6db8a4a201b86fff23a0475e3039 authored about 2 years ago by Henry Tu <[email protected]>* Changed Example MLIR backend to Reference MLIR backend
* Moved reference_ltc_backend into csr...
47bb38d180187019e54d262a093ecdadd24414c2 authored about 2 years ago by Henry Tu <[email protected]>
- Pruned number of xfailed e2e LTC tests from 305 to 134
- Reviewed every failure to ensure th...
9de06f3ebd05dd8e4105bba07192faec1a3c6e55 authored about 2 years ago by Henry Tu <[email protected]>
* Update README.md
* Update README.md
Co-authored-by: Jae Hoon (Antonio) Kim <17433012+ant...
a3244b6a444bc8db622344312306bbcb1aa175b7 authored about 2 years ago by Henry Tu <[email protected]>* Removed compute_shape_native_batch_norm
* Removed compute_shape_native_batch_norm_backward
1510eae75d51cdf2a770875ce03a0cd1a217fd07 authored about 2 years ago by Henry Tu <[email protected]>a62d60829c13076b5ff04d7d6402b69a886ed38f authored about 2 years ago by Jae Hoon (Antonio) Kim <[email protected]>
* Added e2e LTC Torch MLIR tests
* Fix seed for reproducability
* Check if computation is None...
dfcc26556ad30cd02a8ec2f9ad04483cf93721a3 authored about 2 years ago by Henry Tu <[email protected]>8312fa535bd0877ac4a8c8e2ab876366fbaf00ed authored about 2 years ago by Jae Hoon (Antonio) Kim <[email protected]>
* Assume zero rank tensors are scalar
* Run RefineTypes pass on JIT Graph
* Rollback assum...
0c35e607b36e3087e95d50c71c6cef3eab5b6c45 authored about 2 years ago by Henry Tu <[email protected]>de5b380143ddce097e235a56acc941063aeeacdb authored about 2 years ago by Henry Tu <[email protected]>
* Update LazyShapeInference header
* Use JIT GraphExecutor for execution in example backend
406d1e75385dcae64944384a4395f952ff8d3099 authored about 2 years ago by Henry Tu <[email protected]>