31 Commits

Author SHA1 Message Date
Vivek Khandelwal
03601ccdd6 Add v_diffusion_pytorch model in shark/tank (#271) 2022-08-17 22:53:31 +05:30
powderluv
f7f24dc4d9 Revert "Add Debug log of torch_model_blacklist.txt (#242)" (#249)
This reverts commit 7023d556b5.
2022-08-09 10:23:14 -07:00
Chi_Liu
7023d556b5 Add Debug log of torch_model_blacklist.txt (#242)
* Add debug log of torch_model_blacklist.txt

* Add make_fx for torch model

* Update torch_model_blacklists.txt

* Add some Xfails
2022-08-09 17:54:02 +05:30
Prashant Kumar
d934765b1d Add mobilenet_v3_small torch model to the test_suite.
-- The model doesn't validate with the correct results on the GPU.
-- The model passes on CPU and levelzero.
-- The static version of the model gets stuck for vulkan.
2022-08-05 14:10:43 +05:30
Chi_Liu
90fddc6cb0 Add more torch hg model tests (#238) 2022-08-03 18:00:04 -07:00
Chi_Liu
abce0b1c91 Move torch tests up to /tank (#234) 2022-08-03 10:50:53 -07:00
Phaneesh Barwaria
189466bbe4 Mark XFail for M1 Vulkan Failures (#235) 2022-08-02 19:56:02 -07:00
Ean Garvey
c1a14a8db1 Update wide_resnet50_2_test.py (#229)
Removed old xfails.
2022-08-01 20:38:53 -05:00
Ean Garvey
0ee515a7be Fix GPU benchmarks on PyTorch model tests. (#218)
* Add CUDA_BENCHMARKS option to setup.venv to enable PyTorch benchmarks on CUDA.

* Fix PyTorch GPU benchmarks for tank models.
2022-07-29 15:17:55 -07:00
Vivek Khandelwal
cc4fa96831 Remove op decomposition from the v_diffusion.py (#210)
The PyTorch decomposition for the op `aten.upsample_bilinear2d.vec`
is merged in the upstream repo and hence removed from this file.
2022-07-25 17:36:26 +05:30
Prashant Kumar
e62c531170 Update the torch_model for dynamic cases. 2022-07-22 00:21:35 +05:30
Vivek Khandelwal
0e2485d85b Add v-diffusion model 2022-07-20 21:36:42 +05:30
Ean Garvey
c5ae01311d Make PyTorch model tests importer-independent and add benchmarks. 2022-07-15 21:54:41 -05:00
monorimet
08448a013f Refactor TF tests for importer split, update pytorch tests. 2022-07-14 12:59:21 -05:00
Prashant Kumar
0dcf387089 Add shark_importer for torch_models. (#183)
All the torch_models are imported to gs::shark_tank.
Scripts have been updated.
2022-07-12 20:38:19 -07:00
Prashant Kumar
2e22d0b690 Bug fix: Pass the device attribute appropriately.
Previously the device attribute was not passed and device was
hardcoded to "cpu". So every tests were running on cpu.
2022-07-09 01:56:17 +05:30
Chi_Liu
b98c9ee484 torch fix in gen_shark_tank (#180) 2022-07-08 10:13:58 -07:00
Prashant Kumar
fa7ee7e099 Update pytorch tests to support vulkan and cuda.
All the model validation pass except distilbert which is failing
in torch-mlir lowering. Also, added the mobilebert-uncased model to
the torch test suite.
2022-07-08 14:40:13 +05:30
Chi_Liu
d496c66d11 Torch support of gen_shark_tank (#175) 2022-07-06 22:42:15 -07:00
Anush Elangovan
a7435973d9 Fix black formatting 2022-06-30 20:42:02 +00:00
Prashant Kumar
b07377cbfd Refactor the shark_runner shark_inference to only support mlir_modules.
1. The shark_inference is divided into shark_importer and
   shark_inference.
2. All the tank/pytorch tests have been updated.
2022-06-28 18:46:18 +05:30
Prashant Kumar
e8aa105b2a Divide iree_utils and do module imports on function calls. 2022-06-22 14:17:33 +05:30
Ean Garvey
08eda2ce35 Equip TF tests with save_mlir, save_vmfb, and benchmark to .csv options. (#148) 2022-06-21 21:54:44 -07:00
Anush Elangovan
2b3fac7519 Vulkan xfails
https://github.com/google/iree/issues/9554
2022-06-18 09:39:09 +00:00
Prashant Kumar
b43d627b20 Add distilbert-base-uncased to tank pytorch test.
distilbert-base-uncased model validation test is added.
2022-06-15 21:14:56 +05:30
Ean Garvey
dc8c221467 Add option to save .vmfb in pytest + move pytorch model tests. (#131) 2022-06-14 20:39:10 -07:00
Ean Garvey
01f900a9ff Move conftest.py to fix nightly release. (#126)
* Move conftest.py into frontend directory.

* xfail albert static on vulkan backend.
2022-06-14 07:07:33 -07:00
Ean Garvey
e872296498 Enable passing of --save-mlir cmd-line option through pytest. 2022-06-14 00:27:34 +00:00
Ean Garvey
acdd1d1785 Update model testing utilities. (#119) 2022-06-13 02:16:30 -07:00
Ean Garvey
cbd131d588 Refactor testing framework. (#117) 2022-06-12 20:15:30 -07:00
Maksim Levental
1eb493dae9 Demo for currently not MLIR lowerable v-diffusion model (including both demo path and traceable path). (#86) 2022-06-03 14:17:50 -07:00