Code

Feature/misc fixes#25

Code

Closed

mgehre-amd wants to merge 171 commits intomainXilinx/torch-mlir:mainfrom feature/misc_fixesXilinx/torch-mlir:feature/misc_fixesCopy head branch name to clipboard

+3,872-202

Commits

Commits on Apr 12, 2023

• feat: split pytorch requirements into stable and nightly

  

  Maximilian Bartel

  committed
  1fbf17d

  Copy full SHA for 1fbf17d

Commits on Apr 17, 2023

• fix: add true to tests to see full output

  

  Maximilian Bartel

  committed
  6101852

  Copy full SHA for 6101852

• Merge branch 'main' into add_stable_pytorch_tests

  

  Maximilian Bartel

  committed
  5567c2c

  Copy full SHA for 5567c2c

• refactor: add comments to explain true statement

  

  Maximilian Bartel

  committed
  fbc8496

  Copy full SHA for fbc8496

Commits on May 8, 2023

• Merge remote-tracking branch 'origin/main' into add_stable_pytorch_tests

  

  Maximilian Bartel

  committed
  47c8e29

  Copy full SHA for 47c8e29

• feat: move some tests to experimental mode

  

  Maximilian Bartel

  committed
  c93354a

  Copy full SHA for c93354a

• Merge remote-tracking branch 'origin/main' into add_stable_pytorch_tests

  

  Maximilian Bartel

  committed
  e6906d2

  Copy full SHA for e6906d2

Commits on May 11, 2023

• lib/Conversion/TorchToTosa/TosaLegalizeUtils.cpp: bf16 casts

  mgehre-amd

  committed
  8597919

  Copy full SHA for 8597919

• [TOSA] Add torch.prim.NumToTensor.Scalar float support (llvm#1802)

  

  Chi_Liu

  authored and

  mgehre-amd

  committed
  05efc84

  Copy full SHA for 05efc84

• Fixes for bf16

  mgehre-amd

  committed
  00cba87

  Copy full SHA for 00cba87

• Add support for aten.split.Tensor followed by prim.ListUnpack

  mgehre-amd

  committed
  d091717

  Copy full SHA for d091717

Commits on May 16, 2023

• [tosa] support lowering basic torch binary ops with mixed dtypes

  Show description for b66ab03
  tatwaichong

  committed
  b66ab03

  Copy full SHA for b66ab03

• Merge remote-tracking branch 'xlnx/matthias.bf16_tosa' into mixed_types

  mgehre-amd

  committed
  d6f1d9b

  Copy full SHA for d6f1d9b

• .gitmodules: Move llvm-project to xilinx/llvm-project

  mgehre-amd

  committed
  c9b7bb2

  Copy full SHA for c9b7bb2

• externals/llvm-project: Fix mul fold

  Show description for 6f4d02b
  mgehre-amd

  committed
  6f4d02b

  Copy full SHA for 6f4d02b

• Update workflow

  mgehre-amd

  committed
  47a9745

  Copy full SHA for 47a9745

• gitmodules: use https

  mgehre-amd

  committed
  ca73133

  Copy full SHA for ca73133

• Revert llvm-project changes

  mgehre-amd

  committed
  4fa4154

  Copy full SHA for 4fa4154

• refactor: refactor pipeline into more fine grained difference

  

  Maximilian Bartel

  committed
  5812370

  Copy full SHA for 5812370

• Merge remote-tracking branch 'origin/main' into add_stable_pytorch_tests

  

  Maximilian Bartel

  committed
  f3aa805

  Copy full SHA for f3aa805

• Remove a100

  mgehre-amd

  committed
  80a0a59

  Copy full SHA for 80a0a59

• Don't run macos/windows

  mgehre-amd

  committed
  28206ad

  Copy full SHA for 28206ad

• fix

  mgehre-amd

  committed
  1d6ef09

  Copy full SHA for 1d6ef09

• fix token?

  mgehre-amd

  committed
  7eaf5dc

  Copy full SHA for 7eaf5dc

• Revert "Revert llvm-project changes"

  Show description for e35c96e
  mgehre-amd

  committed
  e35c96e

  Copy full SHA for e35c96e

Commits on May 17, 2023

• Workflow update

  mgehre-amd

  committed
  70c8839

  Copy full SHA for 70c8839

• Elevate permissions

  mgehre-amd

  committed
  ff24581

  Copy full SHA for ff24581

• Fix test failure

  mgehre-amd

  committed
  9919d2c

  Copy full SHA for 9919d2c

• Push to current branch name

  mgehre-amd

  committed
  0f054b5

  Copy full SHA for 0f054b5

• More permissions

  mgehre-amd

  committed
  ed394fb

  Copy full SHA for ed394fb

• Merge remote-tracking branch 'upstream/main' into misc_fixes

  mgehre-amd

  committed
  aaef91b

  Copy full SHA for aaef91b

• More bf16 fixes

  mgehre-amd

  committed
  3c63f53

  Copy full SHA for 3c63f53

• Add torch_mlir.do function

  mgehre-amd

  committed
  3f01668

  Copy full SHA for 3f01668

• .github: Fix release flow

  mgehre-amd

  committed
  31085ad

  Copy full SHA for 31085ad

• python/torch_mlir/__init__.py: Fix wrapper for single tuple return

  mgehre-amd

  committed
  ec44d15

  Copy full SHA for ec44d15

• Tooling to build wheels

  mgehre-amd

  committed
  30eb097

  Copy full SHA for 30eb097

• bf16: fix tests

  mgehre-amd

  committed
  70ef112

  Copy full SHA for 70ef112

• build fixes

  mgehre-amd

  committed
  5ffa57d

  Copy full SHA for 5ffa57d

• Fix workflow

  

  torch-mlir

  authored and

  mgehre-amd

  committed
  1ea3632

  Copy full SHA for 1ea3632

• fix wrapper

  mgehre-amd

  committed
  cd7bc7c

  Copy full SHA for cd7bc7c

• Add all legal checks from tosa spec to the check of to.dtype operator

  

  Ferdinand Lemaire

  committed
  0364ee4

  Copy full SHA for 0364ee4

Commits on May 19, 2023

• Merge remote-tracking branch 'xlnx/misc_fixes' into ferdinand.fix_to_dtype_illegal_casts

  mgehre-amd

  committed
  3f621e7

  Copy full SHA for 3f621e7

• Update tests

  mgehre-amd

  committed
  6e279d2

  Copy full SHA for 6e279d2

• Merge remote-tracking branch 'upstream/main' into mgehre.bump_torch

  mgehre-amd

  committed
  4e40ff1

  Copy full SHA for 4e40ff1

• .github/workflows/buildAndTest.yml: Also build PRs to branch misc_fixes

  mgehre-amd

  committed
  c2f166d

  Copy full SHA for c2f166d

• Merge pull request #10 from Xilinx/ferdinand.fix_to_dtype_illegal_casts

  Show description for e2b9716
  mgehre-amd

  authored
  e2b9716

  Copy full SHA for e2b9716

• Make sure that we have ccache entries for misc_fixes

  mgehre-amd

  committed
  bd87b53

  Copy full SHA for bd87b53

• Merge pull request #12 from Xilinx/mgehre.ci_for_misc_fixes

  Show description for d69263d
  mgehre-amd

  authored
  d69263d

  Copy full SHA for d69263d

• Merge remote-tracking branch 'origin/misc_fixes' into mgehre.bump_torch

  mgehre-amd

  committed
  328231f

  Copy full SHA for 328231f

• Remove deep_copy

  mgehre-amd

  committed
  010e594

  Copy full SHA for 010e594

• Print version

  mgehre-amd

  committed
  a05004f

  Copy full SHA for a05004f

• Merge branch 'add_stable_pytorch_tests' into mgehre.stable_torch

  mgehre-amd

  committed
  23ff2a3

  Copy full SHA for 23ff2a3

• Build wheels from stable torch

  mgehre-amd

  committed
  960492c

  Copy full SHA for 960492c

• Merge pull request #11 from Xilinx/mgehre.bump_torch

  Show description for a1abb1c
  mgehre-amd

  authored
  a1abb1c

  Copy full SHA for a1abb1c

• SliceCopyMax_Module: Fix crash in attached test case

  Show description for c536652
  mgehre-amd

  committed
  c536652

  Copy full SHA for c536652

• Merge pull request #16 from Xilinx/mgehre.version

  Show description for e36e3eb
  mgehre-amd

  authored
  e36e3eb

  Copy full SHA for e36e3eb

• Merge pull request #13 from Xilinx/mgehre.negative_idx

  Show description for 7fdb269
  mgehre-amd

  authored
  7fdb269

  Copy full SHA for 7fdb269

• Merge pull request #15 from Xilinx/mgehre.remove_deepcopy

  Show description for b4ad592
  mgehre-amd

  authored
  b4ad592

  Copy full SHA for b4ad592

• Exclude some tests for stable versions

  mgehre-amd

  committed
  6efa91b

  Copy full SHA for 6efa91b

Commits on May 21, 2023

• Merge pull request #14 from Xilinx/mgehre.stable_torch

  Show description for 915398b
  mgehre-amd

  authored
  915398b

  Copy full SHA for 915398b

Commits on May 22, 2023

• Support aten.pow.scalar

  mgehre-amd

  committed
  ebf7534

  Copy full SHA for ebf7534

• Don't crash when the input to aten.copy is unranked

  Show description for cd3713c
  mgehre-amd

  committed
  cd3713c

  Copy full SHA for cd3713c

• Add torch-to-tosa legalization for torch.aten.sqrt

  ttjost

  committed
  5e066c1

  Copy full SHA for 5e066c1

• Support aten.sign

  mgehre-amd

  committed
  d0fec7c

  Copy full SHA for d0fec7c

• Merge pull request #18 from Xilinx/tiagot.legalize_sqrt_to_tosa

  Show description for c12f76a
  mgehre-amd

  authored
  c12f76a

  Copy full SHA for c12f76a

• TorchToTosa: Support more cast from f64

  mgehre-amd

  committed
  f450421

  Copy full SHA for f450421

• Merge branch 'matthias.aten_split' into mgehre.split_tensor

  mgehre-amd

  committed
  6bf21cc

  Copy full SHA for 6bf21cc

• e2e_testing/xfail_sets.py: TOSA: Add a tests that pass now

  mgehre-amd

  committed
  d996586

  Copy full SHA for d996586

• Mark ElementwiseGe as PASS for tosa

  mgehre-amd

  committed
  6676e7c

  Copy full SHA for 6676e7c

• Print name of the backend when tests fail to help debugging issues in CI

  mgehre-amd

  committed
  e91e2a8

  Copy full SHA for e91e2a8

• Merge pull request #23 from Xilinx/mgehre.help_e2e_reporting

  Show description for b2db85c
  mgehre-amd

  authored
  b2db85c

  Copy full SHA for b2db85c

• Use dyn_cast

  mgehre-amd

  committed
  c2fb24e

  Copy full SHA for c2fb24e

• Merge pull request #17 from Xilinx/mgehre.aten.power.scalar

  Show description for 3984b96
  mgehre-amd

  authored
  3984b96

  Copy full SHA for 3984b96

• split.tensor: Ignore in LTC backend

  mgehre-amd

  committed
  5c401ba

  Copy full SHA for 5c401ba

• Merge pull request #19 from Xilinx/mgehre.fix_crash_decompose_copy_unranked

  Show description for 2b728ec
  mgehre-amd

  authored
  2b728ec

  Copy full SHA for 2b728ec

• Merge remote-tracking branch 'origin/misc_fixes' into matthias.support_aten.sign

  mgehre-amd

  committed
  e981d0f

  Copy full SHA for e981d0f

• Merge pull request #20 from Xilinx/matthias.support_aten.sign

  Show description for 064d817
  mgehre-amd

  authored
  064d817

  Copy full SHA for 064d817

• Mark split as XFAIL for LTC

  mgehre-amd

  committed
  116eb05

  Copy full SHA for 116eb05

• Merge pull request #22 from Xilinx/mgehre.split_tensor

  Show description for a2f46e5
  mgehre-amd

  authored
  a2f46e5

  Copy full SHA for a2f46e5

• Add f64 -> f16

  mgehre-amd

  committed
  331ef78

  Copy full SHA for 331ef78

• Merge pull request #21 from Xilinx/mgehre.f64_todtype

  Show description for feba6b1
  mgehre-amd

  authored
  feba6b1

  Copy full SHA for feba6b1

• Merge remote-tracking branch 'origin/misc_fixes' into mgehre.f64_todtype

  mgehre-amd

  committed
  5cec79a

  Copy full SHA for 5cec79a

• Fix test failures

  mgehre-amd

  committed
  5fd8c58

  Copy full SHA for 5fd8c58

Commits on May 23, 2023

• Merge branch 'feature/misc_fixes' into mgehre.f64_todtype

  mgehre-amd

  committed
  d5f7b92

  Copy full SHA for d5f7b92

• Update CI

  mgehre-amd

  committed
  a89d371

  Copy full SHA for a89d371

• Merge pull request #21 from Xilinx/mgehre.f64_todtype

  Show description for 89b42c5
  mgehre-amd

  authored
  89b42c5

  Copy full SHA for 89b42c5

• Merge remote-tracking branch 'upstream/main' into mgehre.bump_torch

  mgehre-amd

  committed
  6a53517

  Copy full SHA for 6a53517

• do(): Add torch.nograd

  mgehre-amd

  committed
  a026db2

  Copy full SHA for a026db2

• Merge pull request #27 from Xilinx/mgehre.nograd

  Show description for 8e2a0ec
  mgehre-amd

  authored
  8e2a0ec

  Copy full SHA for 8e2a0ec

• Merge pull request #26 from Xilinx/mgehre.bump_torch

  Show description for 03ca799
  mgehre-amd

  authored
  03ca799

  Copy full SHA for 03ca799

• lib/Conversion/TorchToTosa/TosaLegalizeUtils.cpp: Just allow all casts between valid types (#28)

  mgehre-amd

  authored
  a950b6a

  Copy full SHA for a950b6a

Commits on May 24, 2023

• feat: adds a folder for torch.aten.broadcast_to operation. (#29)

  ttjost

  authored
  827a091

  Copy full SHA for 827a091

Commits on May 25, 2023

• Add reproduce() function (#30)

  Show description for f144e55
  

  mgehre-amd

  and

  ljfitz

  authored
  f144e55

  Copy full SHA for f144e55

• Ignore constants in the legality error (#31)

  mgehre-amd

  authored
  f57c080

  Copy full SHA for f57c080

• Merge WIP in llvm#2085 (#34)

  ljfitz

  authored
  26cd473

  Copy full SHA for 26cd473

• .github/workflows/buildRelease.yml: Also build packages for python 3.10 (#33)

  mgehre-amd

  authored
  0c9da71

  Copy full SHA for 0c9da71

• upstream slice fixes (#35)

  Show description for 698307a
  

  mgehre-amd

  and

  ramiro050

  authored
  698307a

  Copy full SHA for 698307a

• lib/Conversion/TorchToTosa/TorchToTosa.cpp: Fix legalization of comparions where the input type is bool (#36)

  mgehre-amd

  authored
  f415bdc

  Copy full SHA for f415bdc

• Some slice fixes (#37)

  Show description for 205a76e
  mgehre-amd

  authored
  205a76e

  Copy full SHA for 205a76e

• Fix test failure

  mgehre-amd

  committed
  a97a393

  Copy full SHA for a97a393

• e2e_testing/xfail_sets.py: Add passing test to TOSA set

  mgehre-amd

  committed
  2914eb9

  Copy full SHA for 2914eb9

Commits on May 30, 2023

• lib/Dialect/Torch/IR/TorchOps.cpp: Canonicalize rank changing broadcast_to into reshape + broadcast_to (#40)

  mgehre-amd

  authored
  262e3b7

  Copy full SHA for 262e3b7

• Add support for aten.asin and aten.acos. (#38)

  Show description for f8dac6a
  

  ttjost

  and

  mgehre-amd

  authored
  f8dac6a

  Copy full SHA for f8dac6a

• Merge upstream main (#41)

  Show description for 30f3902
  

  11 peopleauthored
  30f3902

  Copy full SHA for 30f3902

• Revert cumsum (#46)

  mgehre-amd

  authored
  99825b6

  Copy full SHA for 99825b6

• Move model.to before first invocation (#48)

  mgehre-amd

  authored
  e5215db

  Copy full SHA for e5215db

• Pick llvm#2150 to fix slice + copy_ (#43)

  Show description for 243ad62
  

  11 peopleauthored
  243ad62

  Copy full SHA for 243ad62

• Pick upstream fix for module initializers (#42)

  Show description for 72adf77
  

  10 peopleauthored
  72adf77

  Copy full SHA for 72adf77

• Merge branch 'add_stable_pytorch_tests' into matthias.fix_release (#49)

  Show description for 455c7f2
  

  11 peopleauthored
  455c7f2

  Copy full SHA for 455c7f2

Commits on May 31, 2023

• .github/workflows/buildRelease.yml: Fix release build after previous merge

  mgehre-amd

  committed
  b639885

  Copy full SHA for b639885

• do(): Rewrite dataclasses outputs into tuples (#24)

  mgehre-amd

  authored
  0606078

  Copy full SHA for 0606078

• Use torch.nograd on reproduce() (#52)

  Show description for 3acbaa1
  mgehre-amd

  authored
  3acbaa1

  Copy full SHA for 3acbaa1

• Legalize aten.repeat_interleave.Tensor for torch-mlir (#44)

  Show description for ad17fcd
  flemairen6

  authored
  ad17fcd

  Copy full SHA for ad17fcd

• python/torch_mlir/compiler_utils.py: Add comment why we need to run t… (#51)

  Show description for 25333e9
  

  mgehre-amd

  and

  ttjost

  authored
  25333e9

  Copy full SHA for 25333e9

• Support prims.sum / torch.cumsum (#50)

  Show description for 97503d6
  mgehre-amd

  authored
  97503d6

  Copy full SHA for 97503d6

• [FXML-2172] Torch Div Conversion Fix (#45)

  Show description for e15c359
  

  cmcgirr-amd

  and

  mgehre-amd

  authored
  e15c359

  Copy full SHA for e15c359

• Adds support for floating point types in torch.arange operation

  ttjost

  committed
  1c8ca3c

  Copy full SHA for 1c8ca3c

• Merge pull request #47 from Xilinx/tiagot.arange_fp_support

  Show description for 2e45d5e
  ttjost

  authored
  2e45d5e

  Copy full SHA for 2e45d5e

Commits on Jun 1, 2023

• Merge remote-tracking branch 'upstream/main' into mgehre.merge_upstream (#53)

  Show description for 5aef5b8
  

  12 peopleauthored
  5aef5b8

  Copy full SHA for 5aef5b8

• Support for tosa.custom_op operations. (#58)

  ttjost

  authored
  cbeda16

  Copy full SHA for cbeda16

• Legalize operator torch.aten.empty.memory_format, aten.fill.Scalar & aten.repeat_interleave (#55)

  Show description for 76b0fe6
  flemairen6

  authored
  76b0fe6

  Copy full SHA for 76b0fe6

• [TOSA] Add aten._index_put_impl support (#39)

  Show description for f89c053
  

  mgehre-amd

  and

  AmosLewis

  authored
  f89c053

  Copy full SHA for f89c053

• TOSA: Support le.scalar & le.tensor (#61)

  mgehre-amd

  authored
  8a05987

  Copy full SHA for 8a05987

Commits on Jun 2, 2023

• Support Aten_IndexPutImplOp in 1D (#60)

  Show description for b101938
  mgehre-amd

  authored
  b101938

  Copy full SHA for b101938

• lib/Conversion/TorchToTosa/TosaLegalizeCommon.cpp: Handle type promotion in reduction ops (#64)

  Show description for c4666d8
  mgehre-amd

  authored
  c4666d8

  Copy full SHA for c4666d8

• Adds support for aten.sin and aten.cost through tosa.custom_op. (#62)

  ttjost

  authored
  9dcd1be

  Copy full SHA for 9dcd1be

• Torch: Fold RuntimeAssertOp (#65)

  mgehre-amd

  authored
  b25c53a

  Copy full SHA for b25c53a

• Merge remote-tracking branch 'upstream/main' (#57)

  Show description for e53d054
  

  12 peopleauthored
  e53d054

  Copy full SHA for e53d054

Commits on Jun 5, 2023

• Merge remote-tracking branch 'upstream/main' into matthias.merge_upstream

  mgehre-amd

  committed
  b1350f7

  Copy full SHA for b1350f7

• Add make_fx_tosa variant to end2end tests (#66)

  Show description for 2006c1d
  mgehre-amd

  authored
  2006c1d

  Copy full SHA for 2006c1d

• Fix version comparison against stable (#68)

  mgehre-amd

  authored
  5f6a811

  Copy full SHA for 5f6a811

• Add decomposition for im2col (#63)

  Show description for 0a0c7bc
  flemairen6

  authored
  0a0c7bc

  Copy full SHA for 0a0c7bc

• repro: Consider %arg0 to be an SSA for matching the error message (#71)

  mgehre-amd

  authored
  c012cdd

  Copy full SHA for c012cdd

• Merge upstream (#69)

  Show description for 3130853
  

  12 peopleauthored
  3130853

  Copy full SHA for 3130853

Commits on Jun 6, 2023

• Merge commit 'b1350f7f61d98554f47b6712dd44b1418a42ca88' into matthias.merge_upstream

  mgehre-amd

  committed
  9948a52

  Copy full SHA for 9948a52

• Merge remote-tracking branch 'upstream/main' into matthias.merge_upstream

  mgehre-amd

  committed
  90eaf8c

  Copy full SHA for 90eaf8c

Commits on Jun 7, 2023

• Adds support for 1D convolutions by rewriting them as 2D convolutions. (#72)

  ttjost

  authored
  998c9ff

  Copy full SHA for 998c9ff

• do: Use make_fx flag on compile() (#74)

  mgehre-amd

  authored
  9be83db

  Copy full SHA for 9be83db

• Merge pull request #73 from Xilinx/matthias.merge_upstream

  Show description for 71aee8f
  mgehre-amd

  authored
  71aee8f

  Copy full SHA for 71aee8f

Commits on Jun 9, 2023

• TorchToTosa: aten.embedding: Allow indices with any rank (#77)

  mgehre-amd

  authored
  fb9547d

  Copy full SHA for fb9547d

• [TOSA] Add support for aten.pow.Tensor_Tensor operation. (#81)

  ttjost

  authored
  d91a18a

  Copy full SHA for d91a18a

• [TOSA] Add support for bool in aten.empty.memory_format. (#82)

  ttjost

  authored
  f025991

  Copy full SHA for f025991

• TorchToTosa: Fix type for non-FP32 bias. (#80)

  ttjost

  authored
  2dfc2bf

  Copy full SHA for 2dfc2bf

• TorchToTosa: Support AtenSliceTensorOp for any step (#83)

  mgehre-amd

  authored
  867ff64

  Copy full SHA for 867ff64

• Support for group convolutions in TOSA. (#78)

  ttjost

  authored
  2f240bc

  Copy full SHA for 2f240bc

• [TOSA] Fix result type for Conv1D operations. (#79)

  Show description for 7dfe606
  ttjost

  authored
  7dfe606

  Copy full SHA for 7dfe606

• python/torch_mlir/dynamo.py: Decompose index_select (#85)

  mgehre-amd

  authored
  dfc3c0d

  Copy full SHA for dfc3c0d

• TOSA: slice: Support start < 0, start < end and start + sizeOfDim < 0 (#84)

  mgehre-amd

  authored
  3f8a1cd

  Copy full SHA for 3f8a1cd

Commits on Jun 12, 2023

• python/torch_mlir/dynamo.py: Decompose torch.aten.linalg_vector_norm. (#86)

  ttjost

  authored
  2cb0421

  Copy full SHA for 2cb0421

• Fix SliceOp::fold to check step (introduced by #37) (#87)

  mgehre-amd

  authored
  aae8f14

  Copy full SHA for aae8f14

• TOSA: Concat: promote input types (#88)

  mgehre-amd

  authored
  3f24fe6

  Copy full SHA for 3f24fe6

• Add decomposition for index_select along with a test (#67)

  flemairen6

  authored
  917d5a8

  Copy full SHA for 917d5a8

Commits on Jun 13, 2023

• TorchToTosa: support different variations of aten.clamp operation. (#89)

  ttjost

  authored
  54cad09

  Copy full SHA for 54cad09

• Merge commit 'ddea56a8' into matthias.merge_upstream

  mgehre-amd

  committed
  79bbe93

  Copy full SHA for 79bbe93

• TorchToTosa: Refactor aten.sqrt lowering pass and add support for (#90)

  Show description for 7dfeb6a
  ttjost

  authored
  7dfeb6a

  Copy full SHA for 7dfeb6a

Commits on Jun 14, 2023

• Merge pull request #92 from Xilinx/matthias.merge_upstream

  Show description for 2418e31
  mgehre-amd

  authored
  2418e31

  Copy full SHA for 2418e31

• Add tosa-run and tosa-check to output_types; use IREE for tosa-run (#54)

  mgehre-amd

  authored
  66039d1

  Copy full SHA for 66039d1

• Bump torch-mlir to green commit 6cf7fe4 (Week of 05/29/2023)

  mgehre-amd

  committed
  606b2e2

  Copy full SHA for 606b2e2

• TorchToTosa: attempt to fold tosa::add in AtenBroadcastToOp legalization. (#94)

  ttjost

  authored
  0a19469

  Copy full SHA for 0a19469

• TorchToTosa: legalize aten.erf operation (#97)

  ttjost

  authored
  a5081bb

  Copy full SHA for a5081bb

• TorchToTosa: Legalize aten.repeat_interleave operation. (#96)

  ttjost

  authored
  8ca7252

  Copy full SHA for 8ca7252

• Merge branch 'feature/misc_fixes' into matthias.bump_llvm_green-6cf7fe4a

  mgehre-amd

  authored
  0cafd6f

  Copy full SHA for 0cafd6f

• Recompose index.Tensor

  mgehre-amd

  committed
  e273e67

  Copy full SHA for e273e67

Commits on Jun 16, 2023

• Update to LLVM green commit 2b4807b (Week of 06/05/2023)

  mgehre-amd

  committed
  6b34a9b

  Copy full SHA for 6b34a9b

• Merge branch 'matthias.bump_llvm_green-2b4807' into matthias.recompose_index_tensor

  mgehre-amd

  committed
  c79870c

  Copy full SHA for c79870c

• Merge pull request #98 from Xilinx/matthias.recompose_index_tensor

  Show description for 4e2282d
  mgehre-amd

  authored
  4e2282d

  Copy full SHA for 4e2282d

• Merge pull request #104 from Xilinx/matthias.bump_llvm_green-2b4807

  Show description for 198c510
  mgehre-amd

  authored
  198c510

  Copy full SHA for 198c510

Commits on Jun 19, 2023

• Derive output_size of repeat_interleave when inputs are broadcast(fill(x)) (#109)

  mgehre-amd

  authored
  4b6d89a

  Copy full SHA for 4b6d89a

• python/torch_mlir/repro.py: Reduce inputs (#103)

  mgehre-amd

  authored
  47b3580

  Copy full SHA for 47b3580

Commits on Jun 21, 2023

• RecomposeComplexOps: Don't call erase after replaceOpWithNewOp (#111)

  mgehre-amd

  authored
  42de546

  Copy full SHA for 42de546

• lib/Conversion/TorchToTosa/TorchToTosa.cpp: Empty line (#107)

  mgehre-amd

  authored
  cd9cb51

  Copy full SHA for cd9cb51