Commit Graph

8 Commits

Author SHA1 Message Date
Tao Xu
495070b388 [Metal] Add the Python binding for optimize_for_mobile (#46456)
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/46456

Add the python binding in CMake. The general workflow is

- Build pytorch -  `USE_PYTORCH_METAL=ON python setup.py install --cmake`
- Run optimize_for_mobile

```
import torch
from torch.utils.mobile_optimizer import optimize_for_mobile

scripted_model = torch.jit.load('./mobilenetv2.pt')
optimized_model = optimize_for_mobile(scripted_model, backend='metal')
torch.jit.export_opnames(optimized_model)
torch.jit.save(optimized_model, './mobilenetv2_metal.bc')
```
The exported ops are

```
['aten::adaptive_avg_pool2d', 'aten::add.Tensor', 'aten::addmm', 'aten::reshape', 'aten::size.int', 'metal::copy_to_host', 'metal_prepack::conv2d_run']
```
ghstack-source-id: 114559878

Test Plan:
- Sandcastle CI
- Circle CI

Reviewed By: kimishpatel

Differential Revision: D24356768

fbshipit-source-id: fb5c4c4b6316347b67edb4132da044a81470ddfd
2020-10-17 10:26:25 -07:00
Ivan Kobzarev
e9941a5dd4 [vulkan][py] torch.utils.optimize_for_vulkan (#44903)
Summary: Pull Request resolved: https://github.com/pytorch/pytorch/pull/44903

Test Plan: Imported from OSS

Reviewed By: kimishpatel

Differential Revision: D23766039

Pulled By: IvanKobzarev

fbshipit-source-id: dbdf484ee7d3a7719aab105efba51b92ebc51568
2020-09-18 18:20:11 -07:00
Xingying Cheng
bcd75bd683 [ModelLints] Refine dropout lint message. (#42046)
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/42046

Refine dropout lint message as we have enabled dropout operator removal in optimize_for_mobile method.
ghstack-source-id: 108607182

Test Plan: buck test ai_infra/ai_mobile_infra/tests:mobile_model_util_tests

Reviewed By: kimishpatel

Differential Revision: D22741132

fbshipit-source-id: 8f87356aae2bd9c89d1cad0d7be7286278bb14ad
2020-07-27 18:15:30 -07:00
Xingying Cheng
04320a47d7 Add optimizer_for_mobile doc into python api root doc (#41211)
Summary: Pull Request resolved: https://github.com/pytorch/pytorch/pull/41211

Test Plan: Imported from OSS

Reviewed By: xta0

Differential Revision: D22543608

fbshipit-source-id: bf522a6c94313bf2696eca3c5bb5812ea98998d0
2020-07-15 09:57:40 -07:00
Kimish Patel
4a174c83ca Add option to preserve certain methods during optimize_for_mobile. (#40629)
Summary:
By default freeze_module pass, invoked from optimize_for_mobile,
preserves only forward method. There is an option to specify a list of
methods that can be preserved during freeze_module. This PR exposes that
to optimize_for_module pass.
Pull Request resolved: https://github.com/pytorch/pytorch/pull/40629

Test Plan: python test/test_mobile_optimizer.py

Reviewed By: dreiss

Differential Revision: D22260972

Pulled By: kimishpatel

fbshipit-source-id: 452c653269da8bb865acfb58da2d28c23c66e326
2020-06-29 09:32:53 -07:00
Xingying Cheng
0b3755b1d0 Add optimization blacklist as second arg to optimizeForMobile method. (#37462)
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/37462

Instead of running all the optimization pass in optimizeForMobile method,
introducing a whitelist optimizer dictionary as second param in the method,
when it is not passed during calling, the method will run all the optimization
passes, otherwise the method will read the dict and only run the pass with
value of True.
ghstack-source-id: 106104503

Test Plan:
python test/test_mobile_optimizer.py

Imported from OSS

Differential Revision: D22096029

fbshipit-source-id: daa9370c0510930f4c032328b225df0bcf97880f
2020-06-17 18:14:45 -07:00
Xingying Cheng
5c9d1e4824 Propagate module lints for mobile scripted module. (#37046)
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/37046
ghstack-source-id: 102669259

Creating a python api entry to generate mobile model lints which takes a scripted module as argument and returns a map of module lints.

The initial version is to create placeholder which included module bundled input as the first lint instance. More lints will be added in the future.

Test Plan: python test/test_optimizer.py

Reviewed By: dreiss

Differential Revision: D21164648

fbshipit-source-id: 9e8f4e19d74b5464a55cc73b9dc18f358c5947d6
2020-04-27 10:20:12 -07:00
Xingying Cheng
86f354c530 Python binding api to optimize for mobile model on script module. (#36357)
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/36357
ghstack-source-id: 101907180

Creating a python api entry to optimize mobile model which takes a scripted module as argument and returns an optimized scripted module. The initial optimization features includes inserting and folding prepack ops.

Test Plan: python test/test_optimizer.py

Differential Revision: D20946076

fbshipit-source-id: 93cb4a5bb2371128f802d738eb26d0a4f3b2fe10
2020-04-17 16:21:27 -07:00