Commit Graph

  • 4b1e241e1e
    revert ext_modules saienduri 2024-09-12 23:51:46 -0700
  • 153b78126d add conditional logic for TorchMLIRPythonTorchExtensionsSources saienduri 2024-09-13 01:49:50 -0500
  • c27bd37d22 update xfail #3709 Yuanqiang Liu 2024-09-13 12:57:56 +0800
  • e47839f6a0 [Stablehlo] fix aten compare ops' promote rules Yuanqiang Liu 2024-09-13 11:57:33 +0800
  • a13f65026a [Torch] add AtenAsStridedOp in torch dialect (#3706) yyp0 2024-09-12 19:07:11 +0800
  • 0604fd93b3 [Torch] enhance fold of aten.alias (#3705) Yuanqiang Liu 2024-09-12 17:04:57 +0800
  • 08776854f8
    Update setup.py saienduri 2024-09-12 18:08:57 -0700
  • 09e345f46c
    Update setup.py saienduri 2024-09-12 18:07:33 -0700
  • 4bb8b4a674
    Update setup.py saienduri 2024-09-12 15:29:53 -0700
  • d61986cfcf
    Add Decompostion for `Aten_SafeSoftmaxOp` (#3708) zjgarvey 2024-09-12 14:58:10 -0700
  • 52c8ebdc7d more xfail updates #3708 zjgarvey 2024-09-12 14:15:36 -0500
  • 928f02606f remove some xfails zjgarvey 2024-09-12 13:22:23 -0500
  • 42612d88b2 Finish Decomposition for _safe_softmax zjgarvey 2024-09-12 13:01:57 -0500
  • 0dccc97754
    disable pytorch extensions saienduri 2024-09-12 10:35:41 -0700
  • 0616af5ba1
    Update setup.py saienduri 2024-09-12 10:28:00 -0700
  • e0c18f815f removed mention of AtenFmodTensorOp from a few more places #3707 Srinath Avadhanula 2024-09-12 09:33:09 -0700
  • fc242aa35c [WIP] Add aten._safe_softmax op Vivek Khandelwal 2024-09-12 16:28:01 +0000
  • b0c625d903 Delete unnecessary linalg conversion for aten.fmod Srinath Avadhanula 2024-09-12 06:59:45 -0700
  • 3507121764 use fixed torch and torchvision Yuanqiang Liu 2024-09-12 19:36:52 +0800
  • edf725ef42
    [Torch] add AtenAsStridedOp in torch dialect (#3706) yyp0 2024-09-12 19:07:11 +0800
  • 3b6fcdebd9 update #3706 yyp0 2024-09-12 17:15:08 +0800
  • 05c78aa506 update yyp0 2024-09-12 17:07:08 +0800
  • 3f07077ff9
    [Torch] enhance fold of aten.alias (#3705) Yuanqiang Liu 2024-09-12 17:04:57 +0800
  • c502a7732b update yyp0 2024-09-12 17:03:54 +0800
  • 40cc4c062d recover scaled_dot_product_attention yyp0 2024-09-12 16:56:24 +0800
  • 9db168feb0 [Torch] enhance fold of aten.alias #3705 Yuanqiang Liu 2024-09-12 15:16:25 +0800
  • f2b5a33574 add canonicalization pattern for maxpool3d with indices op #3704 lingzhiz1998 2024-09-12 14:00:56 +0800
  • 98fc724bee support AtenAsStridedOp geenerate yyp0 2024-09-12 14:05:26 +0800
  • 997d39d2a9
    Update setup.py hengittää 2024-09-11 23:09:14 -0400
  • 00ebb004d5 clean code #3702 penguin-wwy 2024-09-12 09:53:56 +0800
  • bb69014a96
    bump llvm/llvm-project@d418a03e01 (#3700) Rob Suderman 2024-09-11 08:42:31 -0700
  • b723f0160b
    Merge 317488490d into 1c4b9d6a0e #3669 yyp0 2024-09-11 23:20:47 +0800
  • 43b5b72247 [stablehlo] fix type for compare scalar op penguin-wwy 2024-09-11 23:17:08 +0800
  • 1c4b9d6a0e
    Implement lowering of torch.aten.hstack (#3563) Branko Trifkovic 2024-09-11 04:11:47 -0700
  • 75d88076c3 [TorchMLIRModule] Drop get_int64_max/min functions #3699 Vimal Patel 2024-09-11 16:05:44 +0530
  • 04740824ae
    [ci] enable fx_importer2stablehlo ci test (#3698) penguin_wwy 2024-09-11 09:53:23 +0800
  • 7fb9359d2c bump llvm/llvm-project@d418a03e01 #3700 Rob Suderman 2024-09-10 15:42:24 -0700
  • 8a4ec3be44 fix crash #3698 penguin-wwy 2024-09-10 22:34:14 +0800
  • 6934ab81b0
    Bump llvm/llvm-project@b6603e1bf1 (#3697) Rob Suderman 2024-09-10 08:57:15 -0700
  • b35675a78e
    [onnx] Add support for `auto_pad` in `onnx.Conv` (#3670) giacs-epic 2024-09-10 15:01:53 +0000
  • 198e6f0bec [ci] enable fx_importer2stablehlo ci test penguin-wwy 2024-09-10 22:15:29 +0800
  • ef9a8dbf86
    Merge 9edb08f557 into b5d95ff399 #3686 Branko Trifkovic 2024-09-10 17:34:31 +0530
  • b5d95ff399
    build: manually update PyTorch version (#3692) Vivek Khandelwal 2024-09-10 16:02:28 +0530
  • fbb0db17dc
    Disable TORCH_MLIR_ENABLE_JIT_IR_IMPORTER and TORCH_MLIR_ENABLE_PYTORCH_EXTENSIONS by default (#3693) Chi_Liu 2024-09-09 22:58:27 -0700
  • 125934ce9e Disable TORCH_MLIR_ENABLE_PYTORCH_EXTENSIONS instead #3693 AmosLewis 2024-09-09 17:12:41 -0700
  • 5f1989d7be fix up linear tests #3697 Rob Suderman 2024-09-09 20:57:17 -0700
  • 17c6443fa2 Bump llvm/llvm-project@b6603e1bf1 Rob Suderman 2024-09-09 18:03:34 -0700
  • 3648259bf0 Disable TORCH_MLIR_ENABLE_JIT_IR_IMPORTER by default - Only enable it in CI for update_abstract_interp_lib.sh and update_torch_ods.sh usage. AmosLewis 2024-09-09 09:44:05 -0700
  • 12fa3518dd remove test files #3694 Rob Suderman 2024-09-09 16:03:11 -0700
  • 1acaaa3d68
    Merge eb9fc70db4 into e86f56bc76 #3654 Chi_Liu 2024-09-09 15:54:56 -0700
  • 2dbb169957 wipeout Rob Suderman 2024-09-09 15:53:54 -0700
  • e86f56bc76
    [Torch] [TMTensor] Added mask and is_causal support for torch.aten.scaled_dot_product_attention (#3690) rohan-tan-bhowmik 2024-09-09 15:51:41 -0700
  • bc5829e52e WIP Rob Suderman 2024-09-09 15:49:50 -0700
  • 421b74ef5d XFAIL Changes #3690 rohan-tan-bhowmik 2024-09-09 15:16:37 -0700
  • 1a9466e663 XFAIL Changes rohan-tan-bhowmik 2024-09-09 14:58:15 -0700
  • bf88ec560a XFAIL Changes rohan-tan-bhowmik 2024-09-09 14:45:36 -0700
  • 04f9e4df62 XFAIL Changes rohan-tan-bhowmik 2024-09-09 14:23:49 -0700
  • fc41aa3d02 XFAIL Changes rohan-tan-bhowmik 2024-09-09 13:34:37 -0700
  • f78f878740 XFAIL Changes rohan-tan-bhowmik 2024-09-09 13:22:31 -0700
  • d01f743222 Review changes rohan-tan-bhowmik 2024-09-09 12:38:32 -0700
  • 75f124030a Review changes rohan-tan-bhowmik 2024-09-09 11:55:39 -0700
  • fd9491181a Review changes rohan-tan-bhowmik 2024-09-09 11:35:50 -0700
  • 2cab3c030c Merge branch 'sdpa_mask' of https://github.com/rohan-tan-bhowmik/torch-mlir into sdpa_mask rohan-tan-bhowmik 2024-09-09 11:18:10 -0700
  • a743b2fb41 XFAIL on stable version, PASS on nightly version rohan-tan-bhowmik 2024-09-09 11:17:58 -0700
  • 67b318f410
    Merge branch 'llvm:main' into sdpa_mask rohan-tan-bhowmik 2024-09-09 10:24:49 -0700
  • 0ea7cbaa86 Merge branch 'sdpa_mask' of https://github.com/rohan-tan-bhowmik/torch-mlir into sdpa_mask rohan-tan-bhowmik 2024-09-09 10:17:08 -0700
  • 3c269d46fc XFAIL on stable version, PASS on nightly version rohan-tan-bhowmik 2024-09-09 10:16:08 -0700
  • 0a788e0467
    Decompose aten.fmod into aten.mul,sub,div etc. (#3689) Srinath Avadhanula 2024-09-09 12:00:11 -0400
  • d112f6e200 build: manually update PyTorch version #3692 Vivek Khandelwal 2024-09-09 15:28:18 +0000
  • daeff4a166
    Merge branch 'llvm:main' into sdpa_mask rohan-tan-bhowmik 2024-09-08 16:28:42 -0700
  • ebd4c0c82b XFAIL on stable version, PASS on nightly version rohan-tan-bhowmik 2024-09-08 16:23:58 -0700
  • df6098e43d
    [TorchToLinalg] Use `linalg.transpose` instead of `generic` when lowering `aten.T` (#3660) Felix Schneider 2024-09-07 08:09:10 +0200
  • fd981c4049 XFAIL on stable version, PASS on nightly version rohan-tan-bhowmik 2024-09-06 17:00:22 -0700
  • 1c0f1389e2 XFAIL on stable version, PASS on nightly version rohan-tan-bhowmik 2024-09-06 16:49:56 -0700
  • 84cc5cb970 XFAIL on stable version, PASS on nightly version rohan-tan-bhowmik 2024-09-06 16:17:45 -0700
  • f70166df61 XFAIL on stable version, PASS on nightly version rohan-tan-bhowmik 2024-09-06 16:09:27 -0700
  • f067dd0c35 Restored llvm-project rohan-tan-bhowmik 2024-09-06 15:40:00 -0700
  • 04ba8d4722 add a test #3660 Felix Schneider 2024-09-07 00:25:07 +0200
  • 0e41ca3840 Merge branch 'sdpa_mask' of https://github.com/rohan-tan-bhowmik/torch-mlir into sdpa_mask rohan-tan-bhowmik 2024-09-06 15:22:45 -0700
  • 9d6f70d254 Restored llvm-project rohan-tan-bhowmik 2024-09-06 15:22:11 -0700
  • f342f001a4 opt out of decomposing aten.fmod.Tensor for stablehlo #3689 Srinath Avadhanula 2024-09-06 13:58:20 -0700
  • f6e873ea3d
    Merge branch 'llvm:main' into sdpa_mask rohan-tan-bhowmik 2024-09-06 12:17:07 -0700
  • 2a112ec71f Formatting rohan-tan-bhowmik 2024-09-06 12:12:32 -0700
  • 898321016a Clean up and robustness work rohan-tan-bhowmik 2024-09-06 12:12:09 -0700
  • eb9fc70db4 Bring the onnx/onnx_tosa test back #3654 AmosLewis 2024-09-04 09:04:53 -0700
  • 525e5cd4b5 Add comment on TORCH_MLIR_ENABLE_PYTORCH_EXTENSIONS and TORCH_MLIR_ENABLE_JIT_IR_IMPORTER AmosLewis 2024-08-24 18:58:40 -0700
  • 657768a42a Remove stablehlo/linalg/tosa CI tests for default torch - also delete the onnx and onnx_tosa flag in config AmosLewis 2024-08-24 18:51:51 -0700
  • efa65c840e Change the e2e test default from linalg to fx_importer AmosLewis 2024-08-21 16:01:40 -0700
  • 65d6cf9097 Delete the onnx/onnx_tosa/torchdynamo e2e test whose TestConfig using torchscript - No module named 'torch_mlir.jit_ir_importer' AmosLewis 2024-08-21 15:29:51 -0700
  • 1cf7b78f64 Comment out tosa/stablehlo/linalg default test with torchscript - To fix error when e2etest from torch_mlir import torchscript then output bug: No module named 'torch_mlir.jit_ir_importer' AmosLewis 2024-08-21 14:46:39 -0700
  • 3c87d9d675 Delete custom_op_shape_dtype_fn.py test relate to torchscript AmosLewis 2024-08-20 19:36:20 -0700
  • 1420f97ca5 Only Disable the TORCH_MLIR_ENABLE_JIT_IR_IMPORTER - Delete torchscript related test AmosLewis 2024-08-20 19:19:02 -0700
  • 14c2ac8d98 Enable torch_mlir_e2e_test when TORCH_MLIR_ENABLE_JIT_IR_IMPORTER OFF - To fix the bug No module named 'torch_mlir_e2e_test' when test torch-stable AmosLewis 2024-08-20 17:01:58 -0700
  • 56f6c4e02f Delete the compile.py for torchscript.compile test Ideas from this discussion https://discourse.llvm.org/t/drastically-reducing-documented-scope-of-project/80484 AmosLewis 2024-08-20 16:49:52 -0700
  • 93873416b9 Disable TORCH_MLIR_ENABLE_PYTORCH_EXTENSIONS To fix the no module named 'torch_mlir._mlir_libs._jit_ir_importer' bug. AmosLewis 2024-08-20 16:10:31 -0700
  • 70d5730c87
    [LINALG] Implement lowering of torch.aten.rot90 (#3551) Branko Trifkovic 2024-09-05 22:06:17 -0700
  • 0fbe8cb0b5 fixes Srinath Avadhanula 2024-09-05 15:59:35 -0700
  • 263b380b7f (WIP) Added checks and tests rohan-tan-bhowmik 2024-09-04 12:26:59 -0700
  • a29fce4947 (WIP) Added checks and tests rohan-tan-bhowmik 2024-09-04 12:26:45 -0700
  • dd290bf06b (WIP) Added causal and boolean masking rohan-tan-bhowmik 2024-09-03 16:00:22 -0700