mmdeploy/csrc/core/mpl/static_any.h
lvhan028 9306bcec80
Dev v0.4.0 (#301)
* bump version to v0.4.0

* [Enhancement] Make rewriter more powerful (#150)

* Finish function tests

* lint

* resolve comments

* Fix tests

* docstring & fix

* Complement informations

* lint

* Add example

* Fix version

* Remove todo

Co-authored-by: RunningLeon <mnsheng@yeah.net>

* Torchscript support (#159)

* support torchscript

* add nms

* add torchscript configs and update deploy process and dump-info

* typescript -> torchscript

* add torchscript custom extension support

* add ts custom ops again

* support mmseg unet

* [WIP] add optimizer for torchscript (#119)

* add passes

* add python api

* Torchscript optimizer python api (#121)

* add passes

* add python api

* use python api instead of executable

* Merge Master, update optimizer (#151)

* [Feature] add yolox ncnn (#29)

* add yolox ncnn

* add ncnn android performance of yolox

* add ut

* fix lint

* fix None bugs for ncnn

* test codecov

* test codecov

* add device

* fix yapf

* remove if-else for img shape

* use channelshuffle optimize

* change benchmark after channelshuffle

* fix yapf

* fix yapf

* fuse continuous reshape

* fix static shape deploy

* fix code

* drop pad

* only static shape

* fix static

* fix docstring

* Added mask overlay to output image, changed fprintf info messages to … (#55)

* Added mask overlay to output image, changed fprintf info messages to stdout

* Improved box filtering (filter area/score), make sure roi coordinates stay within bounds

* clang-format

* Support UNet in mmseg (#77)

* Repeatdataset in train has no CLASSES & PALETTE

* update result for unet

* update docstring for mmdet

* remove ppl for unet in docs

* fix ort wrap about input type (#81)

* Fix memleak (#86)

* delete []

* fix build error when enble MMDEPLOY_ACTIVE_LEVEL

* fix lint

* [Doc] Nano benchmark and tutorial (#71)

* add cls benchmark

* add nano zh-cn benchmark and en tutorial

* add device row

* add doc path to index.rst

* fix typo

* [Fix] fix missing deploy_core (#80)

* fix missing deploy_core

* mv flag to demo

* target link

* [Docs] Fix links in Chinese doc (#84)

* Fix docs in Chinese link

* Fix links

* Delete symbolic link and add links to html

* delete files

* Fix link

* [Feature] Add docker files (#67)

* add gpu and cpu dockerfile

* fix lint

* fix cpu docker and remove redundant

* use pip instead

* add build arg and readme

* fix grammar

* update readme

* add chinese doc for dockerfile and add docker build to build.md

* grammar

* refine dockerfiles

* add FAQs

* update Dpplcv_DIR for SDK building

* remove mmcls

* add sdk demos

* fix typo and lint

* update FAQs

* [Fix]fix check_env (#101)

* fix check_env

* update

* Replace convert_syncbatchnorm in mmseg (#93)

* replace convert_syncbatchnorm with revert_sync_batchnorm from mmcv

* change logger

* [Doc] Update FAQ for TensorRT (#96)

* update FAQ

* comment

* [Docs]: Update doc for openvino installation (#102)

* fix docs

* fix docs

* fix docs

* fix mmcv version

* fix docs

* rm blank line

* simplify non batch nms (#99)

* [Enhacement] Allow test.py to save evaluation results (#108)

* Add log file

* Delete debug code

* Rename logger

* resolve comments

* [Enhancement] Support mmocr v0.4+ (#115)

* support mmocr v0.4+

* 0.4.0 -> 0.4.1

* fix onnxruntime wrapper for gpu inference (#123)

* fix ncnn wrapper for ort-gpu

* resolve comment

* fix lint

* Fix typo (#132)

* lock mmcls version (#131)

* [Enhancement] upgrade isort in pre-commit config (#141)

* [Enhancement] upgrade isort in pre-commit config by refering to mmflow pr #87

* fix lint

* remove .isort.cfg and put its known_third_party to setup.cfg

* Fix ci for mmocr (#144)

* fix mmocr unittests

* remove useless

* lock mmdet maximum version to 2.20

* pip install -U numpy

* Fix capture_output (#125)

Co-authored-by: hanrui1sensetime <83800577+hanrui1sensetime@users.noreply.github.com>
Co-authored-by: Johannes L <tehkillerbee@users.noreply.github.com>
Co-authored-by: RunningLeon <mnsheng@yeah.net>
Co-authored-by: VVsssssk <88368822+VVsssssk@users.noreply.github.com>
Co-authored-by: lvhan028 <lvhan_028@163.com>
Co-authored-by: AllentDan <41138331+AllentDan@users.noreply.github.com>
Co-authored-by: Yifan Zhou <singlezombie@163.com>
Co-authored-by: 杨培文 (Yang Peiwen) <915505626@qq.com>
Co-authored-by: Semyon Bevzyuk <semen.bevzuk@gmail.com>

* configs for all tasks

* use torchvision roi align

* remote unnecessary code

* fix ut

* fix ut

* export

* det dynamic

* det dynamic

* add ut

* fix ut

* add ut and docs

* fix ut

* skip torchscript ut if no ops available

* add torchscript option to build.md

* update benchmark and resolve comments

* resolve conflicts

* rename configs

* fix mrcnn cuda test

* remove useless

* add version requirements to docs and comments to codes

* enable empty image exporting for torchscript and accelerate ORT inference for MRCNN

* rebase

* update example for torchscript.md

* update FAQs for torchscript.md

* resolve comments

* only use torchvision roi_align for torchscript

* fix ut

* use torchvision roi align when pool model is avg

* resolve comments

Co-authored-by: grimoire <streetyao@live.com>
Co-authored-by: grimoire <yaoqian@sensetime.com>
Co-authored-by: hanrui1sensetime <83800577+hanrui1sensetime@users.noreply.github.com>
Co-authored-by: Johannes L <tehkillerbee@users.noreply.github.com>
Co-authored-by: RunningLeon <mnsheng@yeah.net>
Co-authored-by: VVsssssk <88368822+VVsssssk@users.noreply.github.com>
Co-authored-by: lvhan028 <lvhan_028@163.com>
Co-authored-by: Yifan Zhou <singlezombie@163.com>
Co-authored-by: 杨培文 (Yang Peiwen) <915505626@qq.com>
Co-authored-by: Semyon Bevzyuk <semen.bevzuk@gmail.com>

* Update supported mmseg models (#181)

* fix ocrnet cascade decoder

* update mmseg support models

* update mmseg configs

* support emanet and icnet

* set max K of TopK for tensorrt

* update supported models for mmseg in docs

* add test for emamodule

* add configs and update docs

* Update docs

* update benchmark

* [Features]Support mmdet3d (#103)

* add mmdet3d code

* add code

* update code

* [log]This commit finish pointpillar export and evaluate on onnxruntime.The model is sample with nvidia repo model

* add tensorrt config

* fix config

* update

* support for tensorrt

* add config

* fix config`

* fix apis about torch2onnx

* update

* mmdet3d deploy version1.0

* map is ok

* fix code

* version1.0

* fix code

* fix visual

* fix bug

* tensorrt support success

* add docstring

* add docs

* fix docs

* fix comments

* fix comment

* fix comment

* fix openvino wrapper

* add unit test

* fix device about cpu

* fix comment

* fix show_result

* fix lint

* fix requirments

* remove ci about det3d

* fix ut

* add ut data

* support for new version pointpillars

* fix comment

* fix support_list

* fix comments

* fix config name

* [Enhancement] Update pad logic in detection heads (#168)

* pad with register

* fix lint

Co-authored-by: AllentDan <dongchunyu@sensetime.com>

* [Enhancement] Additional arguments support for OpenVINO Model Optimizer (#178)

* Add mo args.

* [Docs]: update docs and argument descriptions (#196)

* bump version to v0.4.0

* update docs and argument descriptions

* revert version change

* fix unnecessary change of config for dynamic exportation (#199)

* fix mmcls get classes (#215)

* fix mmcls get classes

* resolve comment

* resolve comment

* Add ModelOptimizerOptions.

* Fix merge bugs.

* Update mmpose.md (#224)

* [Dostring]add example in apis docstring (#214)

* add example in apis docstring

* add backend example in docstring

* rm blank line

* Fixed get_mo_options_from_cfg args

* fix l2norm test

Co-authored-by: RunningLeon <mnsheng@yeah.net>
Co-authored-by: Haofan Wang <frankmiracle@outlook.com>
Co-authored-by: VVsssssk <88368822+VVsssssk@users.noreply.github.com>
Co-authored-by: grimoire <yaoqian@sensetime.com>

* [Enhancement] Switch to statically typed Value::Any (#209)

* replace std::any with StaticAny

* fix __compare_typeid

* remove fallback id support

* constraint on traits::TypeId<T>::value

* fix includes

* [Enhancement] TensorRT DCN support (#205)

* add tensorrt dcn support

* fix lint

* remove roi_align plugin for ORT (#258)

* remove roi_align plugin

* remove ut

* skip single_roi_extractor UT for ORT in CI

* move align to symbolic and update docs

* recover UT

* resolve comments

* [Enhancement]: Support fcn_unet deployment with dynamic shape (#251)

* support mmseg fcn+unet dynamic shape

* add test

* fix ci

* fix units

* resolve comments

* [Enhancement] fix-cmake-relocatable (#223)

* require user to specify xxx_dir

* fix line ending

* fix end-of-file-fixer

* try to fix ld cudart cublas

* add ENV var search

* fix CMAKE_CUDA_COMPILER

* cpu, cuda should all work well

* remove commented code

* fix ncnn example find ncnn package (#282)

* table format is wrong (#283)

* update pre-commit (#284)

* update pre-commit

* fix clang-format

* fix mmseg config (#281)

* fix mmseg config

* fix mmpose evaluate outputs

* fix lint

* update pre-commit config

* fix lint

* Revert "update pre-commit config"

This reverts commit c3fd71611f0b79dfa9ad73fc0f4555c1b3563665.

* miss code symbol (#296)

* refactor cmake build (#295)

* add-mmpose-sdk (#259)

* Torchscript support (#159)

* support torchscript

* add nms

* add torchscript configs and update deploy process and dump-info

* typescript -> torchscript

* add torchscript custom extension support

* add ts custom ops again

* support mmseg unet

* [WIP] add optimizer for torchscript (#119)

* add passes

* add python api

* Torchscript optimizer python api (#121)

* add passes

* add python api

* use python api instead of executable

* Merge Master, update optimizer (#151)

* [Feature] add yolox ncnn (#29)

* add yolox ncnn

* add ncnn android performance of yolox

* add ut

* fix lint

* fix None bugs for ncnn

* test codecov

* test codecov

* add device

* fix yapf

* remove if-else for img shape

* use channelshuffle optimize

* change benchmark after channelshuffle

* fix yapf

* fix yapf

* fuse continuous reshape

* fix static shape deploy

* fix code

* drop pad

* only static shape

* fix static

* fix docstring

* Added mask overlay to output image, changed fprintf info messages to … (#55)

* Added mask overlay to output image, changed fprintf info messages to stdout

* Improved box filtering (filter area/score), make sure roi coordinates stay within bounds

* clang-format

* Support UNet in mmseg (#77)

* Repeatdataset in train has no CLASSES & PALETTE

* update result for unet

* update docstring for mmdet

* remove ppl for unet in docs

* fix ort wrap about input type (#81)

* Fix memleak (#86)

* delete []

* fix build error when enble MMDEPLOY_ACTIVE_LEVEL

* fix lint

* [Doc] Nano benchmark and tutorial (#71)

* add cls benchmark

* add nano zh-cn benchmark and en tutorial

* add device row

* add doc path to index.rst

* fix typo

* [Fix] fix missing deploy_core (#80)

* fix missing deploy_core

* mv flag to demo

* target link

* [Docs] Fix links in Chinese doc (#84)

* Fix docs in Chinese link

* Fix links

* Delete symbolic link and add links to html

* delete files

* Fix link

* [Feature] Add docker files (#67)

* add gpu and cpu dockerfile

* fix lint

* fix cpu docker and remove redundant

* use pip instead

* add build arg and readme

* fix grammar

* update readme

* add chinese doc for dockerfile and add docker build to build.md

* grammar

* refine dockerfiles

* add FAQs

* update Dpplcv_DIR for SDK building

* remove mmcls

* add sdk demos

* fix typo and lint

* update FAQs

* [Fix]fix check_env (#101)

* fix check_env

* update

* Replace convert_syncbatchnorm in mmseg (#93)

* replace convert_syncbatchnorm with revert_sync_batchnorm from mmcv

* change logger

* [Doc] Update FAQ for TensorRT (#96)

* update FAQ

* comment

* [Docs]: Update doc for openvino installation (#102)

* fix docs

* fix docs

* fix docs

* fix mmcv version

* fix docs

* rm blank line

* simplify non batch nms (#99)

* [Enhacement] Allow test.py to save evaluation results (#108)

* Add log file

* Delete debug code

* Rename logger

* resolve comments

* [Enhancement] Support mmocr v0.4+ (#115)

* support mmocr v0.4+

* 0.4.0 -> 0.4.1

* fix onnxruntime wrapper for gpu inference (#123)

* fix ncnn wrapper for ort-gpu

* resolve comment

* fix lint

* Fix typo (#132)

* lock mmcls version (#131)

* [Enhancement] upgrade isort in pre-commit config (#141)

* [Enhancement] upgrade isort in pre-commit config by refering to mmflow pr #87

* fix lint

* remove .isort.cfg and put its known_third_party to setup.cfg

* Fix ci for mmocr (#144)

* fix mmocr unittests

* remove useless

* lock mmdet maximum version to 2.20

* pip install -U numpy

* Fix capture_output (#125)

Co-authored-by: hanrui1sensetime <83800577+hanrui1sensetime@users.noreply.github.com>
Co-authored-by: Johannes L <tehkillerbee@users.noreply.github.com>
Co-authored-by: RunningLeon <mnsheng@yeah.net>
Co-authored-by: VVsssssk <88368822+VVsssssk@users.noreply.github.com>
Co-authored-by: lvhan028 <lvhan_028@163.com>
Co-authored-by: AllentDan <41138331+AllentDan@users.noreply.github.com>
Co-authored-by: Yifan Zhou <singlezombie@163.com>
Co-authored-by: 杨培文 (Yang Peiwen) <915505626@qq.com>
Co-authored-by: Semyon Bevzyuk <semen.bevzuk@gmail.com>

* configs for all tasks

* use torchvision roi align

* remote unnecessary code

* fix ut

* fix ut

* export

* det dynamic

* det dynamic

* add ut

* fix ut

* add ut and docs

* fix ut

* skip torchscript ut if no ops available

* add torchscript option to build.md

* update benchmark and resolve comments

* resolve conflicts

* rename configs

* fix mrcnn cuda test

* remove useless

* add version requirements to docs and comments to codes

* enable empty image exporting for torchscript and accelerate ORT inference for MRCNN

* rebase

* update example for torchscript.md

* update FAQs for torchscript.md

* resolve comments

* only use torchvision roi_align for torchscript

* fix ut

* use torchvision roi align when pool model is avg

* resolve comments

Co-authored-by: grimoire <streetyao@live.com>
Co-authored-by: grimoire <yaoqian@sensetime.com>
Co-authored-by: hanrui1sensetime <83800577+hanrui1sensetime@users.noreply.github.com>
Co-authored-by: Johannes L <tehkillerbee@users.noreply.github.com>
Co-authored-by: RunningLeon <mnsheng@yeah.net>
Co-authored-by: VVsssssk <88368822+VVsssssk@users.noreply.github.com>
Co-authored-by: lvhan028 <lvhan_028@163.com>
Co-authored-by: Yifan Zhou <singlezombie@163.com>
Co-authored-by: 杨培文 (Yang Peiwen) <915505626@qq.com>
Co-authored-by: Semyon Bevzyuk <semen.bevzuk@gmail.com>

* Update supported mmseg models (#181)

* fix ocrnet cascade decoder

* update mmseg support models

* update mmseg configs

* support emanet and icnet

* set max K of TopK for tensorrt

* update supported models for mmseg in docs

* add test for emamodule

* add configs and update docs

* Update docs

* update benchmark

* [Features]Support mmdet3d (#103)

* add mmdet3d code

* add code

* update code

* [log]This commit finish pointpillar export and evaluate on onnxruntime.The model is sample with nvidia repo model

* add tensorrt config

* fix config

* update

* support for tensorrt

* add config

* fix config`

* fix apis about torch2onnx

* update

* mmdet3d deploy version1.0

* map is ok

* fix code

* version1.0

* fix code

* fix visual

* fix bug

* tensorrt support success

* add docstring

* add docs

* fix docs

* fix comments

* fix comment

* fix comment

* fix openvino wrapper

* add unit test

* fix device about cpu

* fix comment

* fix show_result

* fix lint

* fix requirments

* remove ci about det3d

* fix ut

* add ut data

* support for new version pointpillars

* fix comment

* fix support_list

* fix comments

* fix config name

* [Enhancement] Additional arguments support for OpenVINO Model Optimizer (#178)

* Add mo args.

* [Docs]: update docs and argument descriptions (#196)

* bump version to v0.4.0

* update docs and argument descriptions

* revert version change

* fix unnecessary change of config for dynamic exportation (#199)

* fix mmcls get classes (#215)

* fix mmcls get classes

* resolve comment

* resolve comment

* Add ModelOptimizerOptions.

* Fix merge bugs.

* Update mmpose.md (#224)

* [Dostring]add example in apis docstring (#214)

* add example in apis docstring

* add backend example in docstring

* rm blank line

* Fixed get_mo_options_from_cfg args

* fix l2norm test

Co-authored-by: RunningLeon <mnsheng@yeah.net>
Co-authored-by: Haofan Wang <frankmiracle@outlook.com>
Co-authored-by: VVsssssk <88368822+VVsssssk@users.noreply.github.com>
Co-authored-by: grimoire <yaoqian@sensetime.com>

* add-mmpose-codebase

* fix ci

* fix img_shape after TopDownAffine

* rename TopDown module -> XheadDecode & implement regression decode

* align keypoints_from_heatmap

* remove hardcode keypoint_head, need refactor, current only support topdown config

* add mmpose python api

* update mmpose-python code

* can't clip fake box

* fix rebase error

* fix rebase error

* link mspn decoder to base decoder

* fix ci

* compile with gcc7.5

* remove no use code

* fix

* fix prompt

* remove unnecessary cv::parallel_for_

* rewrite TopdownHeatmapMultiStageHead.inference_model

* add comment

* add more detail docstring why use _cs2xyxy in sdk backend

* fix Registry name

* remove no use param & add comment of output result

Co-authored-by: AllentDan <41138331+AllentDan@users.noreply.github.com>
Co-authored-by: grimoire <streetyao@live.com>
Co-authored-by: grimoire <yaoqian@sensetime.com>
Co-authored-by: hanrui1sensetime <83800577+hanrui1sensetime@users.noreply.github.com>
Co-authored-by: Johannes L <tehkillerbee@users.noreply.github.com>
Co-authored-by: RunningLeon <mnsheng@yeah.net>
Co-authored-by: VVsssssk <88368822+VVsssssk@users.noreply.github.com>
Co-authored-by: lvhan028 <lvhan_028@163.com>
Co-authored-by: Yifan Zhou <singlezombie@163.com>
Co-authored-by: 杨培文 (Yang Peiwen) <915505626@qq.com>
Co-authored-by: Semyon Bevzyuk <semen.bevzuk@gmail.com>
Co-authored-by: Haofan Wang <frankmiracle@outlook.com>

* update faq about WinError 1455 (#297)

* update faq about WinError 1455

* Update faq.md

* Update faq.md

* fix ci

Co-authored-by: chenxin2 <chenxin2@sensetime.com>

* [Feature]Support centerpoint (#252)

* bump version to v0.4.0

* [Enhancement] Make rewriter more powerful (#150)

* Finish function tests

* lint

* resolve comments

* Fix tests

* docstring & fix

* Complement informations

* lint

* Add example

* Fix version

* Remove todo

Co-authored-by: RunningLeon <mnsheng@yeah.net>

* Torchscript support (#159)

* support torchscript

* add nms

* add torchscript configs and update deploy process and dump-info

* typescript -> torchscript

* add torchscript custom extension support

* add ts custom ops again

* support mmseg unet

* [WIP] add optimizer for torchscript (#119)

* add passes

* add python api

* Torchscript optimizer python api (#121)

* add passes

* add python api

* use python api instead of executable

* Merge Master, update optimizer (#151)

* [Feature] add yolox ncnn (#29)

* add yolox ncnn

* add ncnn android performance of yolox

* add ut

* fix lint

* fix None bugs for ncnn

* test codecov

* test codecov

* add device

* fix yapf

* remove if-else for img shape

* use channelshuffle optimize

* change benchmark after channelshuffle

* fix yapf

* fix yapf

* fuse continuous reshape

* fix static shape deploy

* fix code

* drop pad

* only static shape

* fix static

* fix docstring

* Added mask overlay to output image, changed fprintf info messages to … (#55)

* Added mask overlay to output image, changed fprintf info messages to stdout

* Improved box filtering (filter area/score), make sure roi coordinates stay within bounds

* clang-format

* Support UNet in mmseg (#77)

* Repeatdataset in train has no CLASSES & PALETTE

* update result for unet

* update docstring for mmdet

* remove ppl for unet in docs

* fix ort wrap about input type (#81)

* Fix memleak (#86)

* delete []

* fix build error when enble MMDEPLOY_ACTIVE_LEVEL

* fix lint

* [Doc] Nano benchmark and tutorial (#71)

* add cls benchmark

* add nano zh-cn benchmark and en tutorial

* add device row

* add doc path to index.rst

* fix typo

* [Fix] fix missing deploy_core (#80)

* fix missing deploy_core

* mv flag to demo

* target link

* [Docs] Fix links in Chinese doc (#84)

* Fix docs in Chinese link

* Fix links

* Delete symbolic link and add links to html

* delete files

* Fix link

* [Feature] Add docker files (#67)

* add gpu and cpu dockerfile

* fix lint

* fix cpu docker and remove redundant

* use pip instead

* add build arg and readme

* fix grammar

* update readme

* add chinese doc for dockerfile and add docker build to build.md

* grammar

* refine dockerfiles

* add FAQs

* update Dpplcv_DIR for SDK building

* remove mmcls

* add sdk demos

* fix typo and lint

* update FAQs

* [Fix]fix check_env (#101)

* fix check_env

* update

* Replace convert_syncbatchnorm in mmseg (#93)

* replace convert_syncbatchnorm with revert_sync_batchnorm from mmcv

* change logger

* [Doc] Update FAQ for TensorRT (#96)

* update FAQ

* comment

* [Docs]: Update doc for openvino installation (#102)

* fix docs

* fix docs

* fix docs

* fix mmcv version

* fix docs

* rm blank line

* simplify non batch nms (#99)

* [Enhacement] Allow test.py to save evaluation results (#108)

* Add log file

* Delete debug code

* Rename logger

* resolve comments

* [Enhancement] Support mmocr v0.4+ (#115)

* support mmocr v0.4+

* 0.4.0 -> 0.4.1

* fix onnxruntime wrapper for gpu inference (#123)

* fix ncnn wrapper for ort-gpu

* resolve comment

* fix lint

* Fix typo (#132)

* lock mmcls version (#131)

* [Enhancement] upgrade isort in pre-commit config (#141)

* [Enhancement] upgrade isort in pre-commit config by refering to mmflow pr #87

* fix lint

* remove .isort.cfg and put its known_third_party to setup.cfg

* Fix ci for mmocr (#144)

* fix mmocr unittests

* remove useless

* lock mmdet maximum version to 2.20

* pip install -U numpy

* Fix capture_output (#125)

Co-authored-by: hanrui1sensetime <83800577+hanrui1sensetime@users.noreply.github.com>
Co-authored-by: Johannes L <tehkillerbee@users.noreply.github.com>
Co-authored-by: RunningLeon <mnsheng@yeah.net>
Co-authored-by: VVsssssk <88368822+VVsssssk@users.noreply.github.com>
Co-authored-by: lvhan028 <lvhan_028@163.com>
Co-authored-by: AllentDan <41138331+AllentDan@users.noreply.github.com>
Co-authored-by: Yifan Zhou <singlezombie@163.com>
Co-authored-by: 杨培文 (Yang Peiwen) <915505626@qq.com>
Co-authored-by: Semyon Bevzyuk <semen.bevzuk@gmail.com>

* configs for all tasks

* use torchvision roi align

* remote unnecessary code

* fix ut

* fix ut

* export

* det dynamic

* det dynamic

* add ut

* fix ut

* add ut and docs

* fix ut

* skip torchscript ut if no ops available

* add torchscript option to build.md

* update benchmark and resolve comments

* resolve conflicts

* rename configs

* fix mrcnn cuda test

* remove useless

* add version requirements to docs and comments to codes

* enable empty image exporting for torchscript and accelerate ORT inference for MRCNN

* rebase

* update example for torchscript.md

* update FAQs for torchscript.md

* resolve comments

* only use torchvision roi_align for torchscript

* fix ut

* use torchvision roi align when pool model is avg

* resolve comments

Co-authored-by: grimoire <streetyao@live.com>
Co-authored-by: grimoire <yaoqian@sensetime.com>
Co-authored-by: hanrui1sensetime <83800577+hanrui1sensetime@users.noreply.github.com>
Co-authored-by: Johannes L <tehkillerbee@users.noreply.github.com>
Co-authored-by: RunningLeon <mnsheng@yeah.net>
Co-authored-by: VVsssssk <88368822+VVsssssk@users.noreply.github.com>
Co-authored-by: lvhan028 <lvhan_028@163.com>
Co-authored-by: Yifan Zhou <singlezombie@163.com>
Co-authored-by: 杨培文 (Yang Peiwen) <915505626@qq.com>
Co-authored-by: Semyon Bevzyuk <semen.bevzuk@gmail.com>

* Update supported mmseg models (#181)

* fix ocrnet cascade decoder

* update mmseg support models

* update mmseg configs

* support emanet and icnet

* set max K of TopK for tensorrt

* update supported models for mmseg in docs

* add test for emamodule

* add configs and update docs

* Update docs

* update benchmark

* [Features]Support mmdet3d (#103)

* add mmdet3d code

* add code

* update code

* [log]This commit finish pointpillar export and evaluate on onnxruntime.The model is sample with nvidia repo model

* add tensorrt config

* fix config

* update

* support for tensorrt

* add config

* fix config`

* fix apis about torch2onnx

* update

* mmdet3d deploy version1.0

* map is ok

* fix code

* version1.0

* fix code

* fix visual

* fix bug

* tensorrt support success

* add docstring

* add docs

* fix docs

* fix comments

* fix comment

* fix comment

* fix openvino wrapper

* add unit test

* fix device about cpu

* fix comment

* fix show_result

* fix lint

* fix requirments

* remove ci about det3d

* fix ut

* add ut data

* support for new version pointpillars

* fix comment

* fix support_list

* fix comments

* fix config name

* [Enhancement] Update pad logic in detection heads (#168)

* pad with register

* fix lint

Co-authored-by: AllentDan <dongchunyu@sensetime.com>

* [Enhancement] Additional arguments support for OpenVINO Model Optimizer (#178)

* Add mo args.

* [Docs]: update docs and argument descriptions (#196)

* bump version to v0.4.0

* update docs and argument descriptions

* revert version change

* fix unnecessary change of config for dynamic exportation (#199)

* fix mmcls get classes (#215)

* fix mmcls get classes

* resolve comment

* resolve comment

* Add ModelOptimizerOptions.

* Fix merge bugs.

* Update mmpose.md (#224)

* [Dostring]add example in apis docstring (#214)

* add example in apis docstring

* add backend example in docstring

* rm blank line

* Fixed get_mo_options_from_cfg args

* fix l2norm test

Co-authored-by: RunningLeon <mnsheng@yeah.net>
Co-authored-by: Haofan Wang <frankmiracle@outlook.com>
Co-authored-by: VVsssssk <88368822+VVsssssk@users.noreply.github.com>
Co-authored-by: grimoire <yaoqian@sensetime.com>

* [Enhancement] Switch to statically typed Value::Any (#209)

* replace std::any with StaticAny

* fix __compare_typeid

* remove fallback id support

* constraint on traits::TypeId<T>::value

* fix includes

* support for centerpoint

* [Enhancement] TensorRT DCN support (#205)

* add tensorrt dcn support

* fix lint

* add docstring and dcn model support

* add centerpoint ut and docs

* add config and fix input rank

* fix merge error

* fix a bug

* fix comment

* [Doc] update benchmark add supported-model-list (#286)

* update benchmark add supported-model-list

* fix lint

* fix lint

* loc mmocr maximum version

* fix ut

Co-authored-by: maningsheng <mnsheng@yeah.net>
Co-authored-by: Yifan Zhou <singlezombie@163.com>
Co-authored-by: AllentDan <41138331+AllentDan@users.noreply.github.com>
Co-authored-by: grimoire <streetyao@live.com>
Co-authored-by: grimoire <yaoqian@sensetime.com>
Co-authored-by: hanrui1sensetime <83800577+hanrui1sensetime@users.noreply.github.com>
Co-authored-by: Johannes L <tehkillerbee@users.noreply.github.com>
Co-authored-by: lvhan028 <lvhan_028@163.com>
Co-authored-by: 杨培文 (Yang Peiwen) <915505626@qq.com>
Co-authored-by: Semyon Bevzyuk <semen.bevzuk@gmail.com>
Co-authored-by: AllentDan <dongchunyu@sensetime.com>
Co-authored-by: Haofan Wang <frankmiracle@outlook.com>
Co-authored-by: lzhangzz <lzhang329@gmail.com>

Co-authored-by: maningsheng <mnsheng@yeah.net>
Co-authored-by: Yifan Zhou <singlezombie@163.com>
Co-authored-by: AllentDan <41138331+AllentDan@users.noreply.github.com>
Co-authored-by: grimoire <streetyao@live.com>
Co-authored-by: grimoire <yaoqian@sensetime.com>
Co-authored-by: hanrui1sensetime <83800577+hanrui1sensetime@users.noreply.github.com>
Co-authored-by: Johannes L <tehkillerbee@users.noreply.github.com>
Co-authored-by: VVsssssk <88368822+VVsssssk@users.noreply.github.com>
Co-authored-by: 杨培文 (Yang Peiwen) <915505626@qq.com>
Co-authored-by: Semyon Bevzyuk <semen.bevzuk@gmail.com>
Co-authored-by: AllentDan <dongchunyu@sensetime.com>
Co-authored-by: Haofan Wang <frankmiracle@outlook.com>
Co-authored-by: lzhangzz <lzhang329@gmail.com>
Co-authored-by: Chen Xin <xinchen.tju@gmail.com>
Co-authored-by: chenxin2 <chenxin2@sensetime.com>
2022-04-01 18:14:23 +08:00

490 lines
15 KiB
C++

// Copyright (c) OpenMMLab. All rights reserved.
#ifndef MMDEPLOY_CSRC_CORE_MPL_STATIC_ANY_H_
#define MMDEPLOY_CSRC_CORE_MPL_STATIC_ANY_H_
#include <cstdint>
#include <memory>
#include <stdexcept>
#include <type_traits>
#include <utility>
// re-implementation of std::any, relies on static type id instead of RTTI.
// adjusted from libc++-10
namespace mmdeploy {
namespace traits {
using type_id_t = uint64_t;
template <class T>
struct TypeId {
static constexpr type_id_t value = 0;
};
template <>
struct TypeId<void> {
static constexpr auto value = static_cast<type_id_t>(-1);
};
// ! This only works when calling inside mmdeploy namespace
#define MMDEPLOY_REGISTER_TYPE_ID(type, id) \
namespace traits { \
template <> \
struct TypeId<type> { \
static constexpr type_id_t value = id; \
}; \
}
} // namespace traits
namespace detail {
template <typename T>
struct is_in_place_type_impl : std::false_type {};
template <typename T>
struct is_in_place_type_impl<std::in_place_type_t<T>> : std::true_type {};
template <typename T>
struct is_in_place_type : public is_in_place_type_impl<T> {};
} // namespace detail
class BadAnyCast : public std::bad_cast {
public:
const char* what() const noexcept override { return "BadAnyCast"; }
};
[[noreturn]] inline void ThrowBadAnyCast() {
#if __cpp_exceptions
throw BadAnyCast{};
#else
std::abort();
#endif
}
// Forward declarations
class StaticAny;
template <class ValueType>
std::add_pointer_t<std::add_const_t<ValueType>> static_any_cast(const StaticAny*) noexcept;
template <class ValueType>
std::add_pointer_t<ValueType> static_any_cast(StaticAny*) noexcept;
namespace __static_any_impl {
using _Buffer = std::aligned_storage_t<3 * sizeof(void*), std::alignment_of_v<void*>>;
template <class T>
using _IsSmallObject =
std::integral_constant<bool, sizeof(T) <= sizeof(_Buffer) &&
std::alignment_of_v<_Buffer> % std::alignment_of_v<T> == 0 &&
std::is_nothrow_move_constructible_v<T>>;
enum class _Action { _Destroy, _Copy, _Move, _Get, _TypeInfo };
union _Ret {
void* ptr_;
traits::type_id_t type_id_;
};
template <class T>
struct _SmallHandler;
template <class T>
struct _LargeHandler;
template <class T>
inline bool __compare_typeid(traits::type_id_t __id) {
if (__id && __id == traits::TypeId<T>::value) {
return true;
}
return false;
}
template <class T>
using _Handler = std::conditional_t<_IsSmallObject<T>::value, _SmallHandler<T>, _LargeHandler<T>>;
} // namespace __static_any_impl
class StaticAny {
public:
constexpr StaticAny() noexcept : h_(nullptr) {}
StaticAny(const StaticAny& other) : h_(nullptr) {
if (other.h_) {
other.__call(_Action::_Copy, this);
}
}
StaticAny(StaticAny&& other) noexcept : h_(nullptr) {
if (other.h_) {
other.__call(_Action::_Move, this);
}
}
template <class ValueType, class T = std::decay_t<ValueType>,
class = std::enable_if_t<
!std::is_same<T, StaticAny>::value && !detail::is_in_place_type<ValueType>::value &&
std::is_copy_constructible<T>::value && traits::TypeId<T>::value>>
explicit StaticAny(ValueType&& value);
template <
class ValueType, class... Args, class T = std::decay_t<ValueType>,
class = std::enable_if_t<std::is_constructible<T, Args...>::value &&
std::is_copy_constructible<T>::value && traits::TypeId<T>::value>>
explicit StaticAny(std::in_place_type_t<ValueType>, Args&&... args);
template <class ValueType, class U, class... Args, class T = std::decay_t<ValueType>,
class = std::enable_if_t<
std::is_constructible<T, std::initializer_list<U>&, Args...>::value &&
std::is_copy_constructible<T>::value && traits::TypeId<T>::value>>
explicit StaticAny(std::in_place_type_t<ValueType>, std::initializer_list<U>, Args&&... args);
~StaticAny() { this->reset(); }
StaticAny& operator=(const StaticAny& rhs) {
StaticAny(rhs).swap(*this);
return *this;
}
StaticAny& operator=(StaticAny&& rhs) noexcept {
StaticAny(std::move(rhs)).swap(*this);
return *this;
}
template <
class ValueType, class T = std::decay_t<ValueType>,
class = std::enable_if_t<!std::is_same<T, StaticAny>::value &&
std::is_copy_constructible<T>::value && traits::TypeId<T>::value>>
StaticAny& operator=(ValueType&& v);
template <
class ValueType, class... Args, class T = std::decay_t<ValueType>,
class = std::enable_if_t<std::is_constructible<T, Args...>::value &&
std::is_copy_constructible<T>::value && traits::TypeId<T>::value>>
T& emplace(Args&&... args);
template <class ValueType, class U, class... Args, class T = std::decay_t<ValueType>,
class = std::enable_if_t<
std::is_constructible<T, std::initializer_list<U>&, Args...>::value &&
std::is_copy_constructible<T>::value && traits::TypeId<T>::value>>
T& emplace(std::initializer_list<U>, Args&&...);
void reset() noexcept {
if (h_) {
this->__call(_Action::_Destroy);
}
}
void swap(StaticAny& rhs) noexcept;
bool has_value() const noexcept { return h_ != nullptr; }
traits::type_id_t type() const noexcept {
if (h_) {
return this->__call(_Action::_TypeInfo).type_id_;
} else {
return traits::TypeId<void>::value;
}
}
private:
using _Action = __static_any_impl::_Action;
using _Ret = __static_any_impl::_Ret;
using _HandleFuncPtr = _Ret (*)(_Action, const StaticAny*, StaticAny*, traits::type_id_t info);
union _Storage {
constexpr _Storage() : ptr_(nullptr) {}
void* ptr_;
__static_any_impl::_Buffer buf_;
};
_Ret __call(_Action a, StaticAny* other = nullptr, traits::type_id_t info = 0) const {
return h_(a, this, other, info);
}
_Ret __call(_Action a, StaticAny* other = nullptr, traits::type_id_t info = 0) {
return h_(a, this, other, info);
}
template <class>
friend struct __static_any_impl::_SmallHandler;
template <class>
friend struct __static_any_impl::_LargeHandler;
template <class ValueType>
friend std::add_pointer_t<std::add_const_t<ValueType>> static_any_cast(const StaticAny*) noexcept;
template <class ValueType>
friend std::add_pointer_t<ValueType> static_any_cast(StaticAny*) noexcept;
_HandleFuncPtr h_ = nullptr;
_Storage s_;
};
namespace __static_any_impl {
template <class T>
struct _SmallHandler {
static _Ret __handle(_Action action, const StaticAny* self, StaticAny* other,
traits::type_id_t info) {
_Ret ret;
ret.ptr_ = nullptr;
switch (action) {
case _Action::_Destroy:
__destroy(const_cast<StaticAny&>(*self));
break;
case _Action::_Copy:
__copy(*self, *other);
break;
case _Action::_Move:
__move(const_cast<StaticAny&>(*self), *other);
break;
case _Action::_Get:
ret.ptr_ = __get(const_cast<StaticAny&>(*self), info);
break;
case _Action::_TypeInfo:
ret.type_id_ = __type_info();
break;
}
return ret;
}
template <class... Args>
static T& __create(StaticAny& dest, Args&&... args) {
T* ret = ::new (static_cast<void*>(&dest.s_.buf_)) T(std::forward<Args>(args)...);
dest.h_ = &_SmallHandler::__handle;
return *ret;
}
private:
template <class... Args>
static void __destroy(StaticAny& self) {
T& value = *static_cast<T*>(static_cast<void*>(&self.s_.buf_));
value.~T();
self.h_ = nullptr;
}
template <class... Args>
static void __copy(const StaticAny& self, StaticAny& dest) {
_SmallHandler::__create(dest, *static_cast<const T*>(static_cast<const void*>(&self.s_.buf_)));
}
static void __move(StaticAny& self, StaticAny& dest) {
_SmallHandler::__create(dest, std::move(*static_cast<T*>(static_cast<void*>(&self.s_.buf_))));
__destroy(self);
}
static void* __get(StaticAny& self, traits::type_id_t info) {
if (__static_any_impl::__compare_typeid<T>(info)) {
return static_cast<void*>(&self.s_.buf_);
}
return nullptr;
}
static traits::type_id_t __type_info() { return traits::TypeId<T>::value; }
};
template <class T>
struct _LargeHandler {
static _Ret __handle(_Action action, const StaticAny* self, StaticAny* other,
traits::type_id_t info) {
_Ret ret;
ret.ptr_ = nullptr;
switch (action) {
case _Action::_Destroy:
__destroy(const_cast<StaticAny&>(*self));
break;
case _Action::_Copy:
__copy(*self, *other);
break;
case _Action::_Move:
__move(const_cast<StaticAny&>(*self), *other);
break;
case _Action::_Get:
ret.ptr_ = __get(const_cast<StaticAny&>(*self), info);
break;
case _Action::_TypeInfo:
ret.type_id_ = __type_info();
break;
}
return ret;
}
template <class... Args>
static T& __create(StaticAny& dest, Args&&... args) {
using _Alloc = std::allocator<T>;
_Alloc alloc;
auto dealloc = [&](T* p) { alloc.deallocate(p, 1); };
std::unique_ptr<T, decltype(dealloc)> hold(alloc.allocate(1), dealloc);
T* ret = ::new ((void*)hold.get()) T(std::forward<Args>(args)...);
dest.s_.ptr_ = hold.release();
dest.h_ = &_LargeHandler::__handle;
return *ret;
}
private:
static void __destroy(StaticAny& self) {
delete static_cast<T*>(self.s_.ptr_);
self.h_ = nullptr;
}
static void __copy(const StaticAny& self, StaticAny& dest) {
_LargeHandler::__create(dest, *static_cast<const T*>(self.s_.ptr_));
}
static void __move(StaticAny& self, StaticAny& dest) {
dest.s_.ptr_ = self.s_.ptr_;
dest.h_ = &_LargeHandler::__handle;
self.h_ = nullptr;
}
static void* __get(StaticAny& self, traits::type_id_t info) {
if (__static_any_impl::__compare_typeid<T>(info)) {
return static_cast<void*>(self.s_.ptr_);
}
return nullptr;
}
static traits::type_id_t __type_info() { return traits::TypeId<T>::value; }
};
} // namespace __static_any_impl
template <class ValueType, class T, class>
StaticAny::StaticAny(ValueType&& v) : h_(nullptr) {
__static_any_impl::_Handler<T>::__create(*this, std::forward<ValueType>(v));
}
template <class ValueType, class... Args, class T, class>
StaticAny::StaticAny(std::in_place_type_t<ValueType>, Args&&... args) {
__static_any_impl::_Handler<T>::__create(*this, std::forward<Args>(args)...);
}
template <class ValueType, class U, class... Args, class T, class>
StaticAny::StaticAny(std::in_place_type_t<ValueType>, std::initializer_list<U> il, Args&&... args) {
__static_any_impl::_Handler<T>::__create(*this, il, std::forward<Args>(args)...);
}
template <class ValueType, class, class>
inline StaticAny& StaticAny::operator=(ValueType&& v) {
StaticAny(std::forward<ValueType>(v)).swap(*this);
return *this;
}
template <class ValueType, class... Args, class T, class>
inline T& StaticAny::emplace(Args&&... args) {
reset();
return __static_any_impl::_Handler<T>::__create(*this, std::forward<Args>(args)...);
}
template <class ValueType, class U, class... Args, class T, class>
inline T& StaticAny::emplace(std::initializer_list<U> il, Args&&... args) {
reset();
return __static_any_impl::_Handler<T>::_create(*this, il, std::forward<Args>(args)...);
}
inline void StaticAny::swap(StaticAny& rhs) noexcept {
if (this == &rhs) {
return;
}
if (h_ && rhs.h_) {
StaticAny tmp;
rhs.__call(_Action::_Move, &tmp);
this->__call(_Action::_Move, &rhs);
tmp.__call(_Action::_Move, this);
} else if (h_) {
this->__call(_Action::_Move, &rhs);
} else if (rhs.h_) {
rhs.__call(_Action::_Move, this);
}
}
inline void swap(StaticAny& lhs, StaticAny& rhs) noexcept { lhs.swap(rhs); }
template <class T, class... Args>
inline StaticAny make_static_any(Args&&... args) {
return StaticAny(std::in_place_type<T>, std::forward<Args>(args)...);
}
template <class T, class U, class... Args>
StaticAny make_static_any(std::initializer_list<U> il, Args&&... args) {
return StaticAny(std::in_place_type<T>, il, std::forward<Args>(args)...);
}
template <class ValueType>
ValueType static_any_cast(const StaticAny& v) {
using _RawValueType = std::remove_cv_t<std::remove_reference_t<ValueType>>;
static_assert(std::is_constructible<ValueType, const _RawValueType&>::value,
"ValueType is required to be a const lvalue reference "
"or a CopyConstructible type");
auto tmp = static_any_cast<std::add_const_t<_RawValueType>>(&v);
if (tmp == nullptr) {
ThrowBadAnyCast();
}
return static_cast<ValueType>(*tmp);
}
template <class ValueType>
inline ValueType static_any_cast(StaticAny& v) {
using _RawValueType = std::remove_cv_t<std::remove_reference_t<ValueType>>;
static_assert(std::is_constructible<ValueType, _RawValueType&>::value,
"ValueType is required to be an lvalue reference "
"or a CopyConstructible type");
auto tmp = static_any_cast<_RawValueType>(&v);
if (tmp == nullptr) {
ThrowBadAnyCast();
}
return static_cast<ValueType>(*tmp);
}
template <class ValueType>
inline ValueType static_any_cast(StaticAny&& v) {
using _RawValueType = std::remove_cv_t<std::remove_reference_t<ValueType>>;
static_assert(std::is_constructible<ValueType, _RawValueType>::value,
"ValueType is required to be an rvalue reference "
"or a CopyConstructible type");
auto tmp = static_any_cast<_RawValueType>(&v);
if (tmp == nullptr) {
ThrowBadAnyCast();
}
return static_cast<ValueType>(std::move(*tmp));
}
template <class ValueType>
inline std::add_pointer_t<std::add_const_t<ValueType>> static_any_cast(
const StaticAny* __any) noexcept {
static_assert(!std::is_reference<ValueType>::value, "ValueType may not be a reference.");
return static_any_cast<ValueType>(const_cast<StaticAny*>(__any));
}
template <class RetType>
inline RetType __pointer_or_func_test(void* p, std::false_type) noexcept {
return static_cast<RetType>(p);
}
template <class RetType>
inline RetType __pointer_or_func_test(void*, std::true_type) noexcept {
return nullptr;
}
template <class ValueType>
std::add_pointer_t<ValueType> static_any_cast(StaticAny* any) noexcept {
using __static_any_impl::_Action;
static_assert(!std::is_reference<ValueType>::value, "ValueType may not be a reference.");
using ReturnType = std::add_pointer_t<ValueType>;
if (any && any->h_) {
void* p = any->__call(_Action::_Get, nullptr, traits::TypeId<ValueType>::value).ptr_;
return __pointer_or_func_test<ReturnType>(p, std::is_function<ValueType>{});
}
return nullptr;
}
} // namespace mmdeploy
#endif // MMDEPLOY_CSRC_CORE_MPL_STATIC_ANY_H_