Migrate_master_to_1.3.2 (#6)

* docs(project): deploee introduction (open-mmlab#2120)

* docs(project): deploee introduction

* Update README_zh-CN.md

* fix mmseg output sync (open-mmlab#2122)

* Update docs (open-mmlab#2114)

* update docs from 1.x to main

* fix dead links

* fix tag_name

* Revert "fix tag_name"

This reverts commit fcf0c58.

* fix readthedocs for zh_cn

* support mscv statically linked CRT (open-mmlab#2118)

* support mscv mt

* prevent modification of flag

* fix uploading assets (open-mmlab#2130)

* Fit for MMPretrain EfficientFormer (open-mmlab#2108)

* add ef in rgtest metafile

* fix for EfficientFormerClsHead to LinearClsHead

* fix lint

* update benchmark

* fix lint

* fix docs

* fix docs

* add openvino test

* fix export info for mmyolo (open-mmlab#2146)

* fix

* fix

* fix

* Update build_ubuntu_x64_ncnn.py (open-mmlab#2154)

* update document for docker installation (open-mmlab#2142)

* update

* update

* add docker ci

* update

* fix

* [ADD] pplcv area interpolation (open-mmlab#2153)

* ADD: area interpolation for PPL.CV (cuda) backend

* ADD: "area" entry for unit testing

* [Fix] fix satrn onnxruntime batch inference (open-mmlab#2139)

* fix satrn onnxruntime batch inference

* add ut

* better compare

* Update readme intro image and docs (open-mmlab#2175)

* update logo

* update

* update

* update

* fix ci

* Revert "update logo"

This reverts commit 6935ff0.

* update intro

* fix

* Add Sdk Doxygen document. (open-mmlab#2157)

* doxygen c api

* fix ci

* update en doc

* update zh doc

* Add coco-wholebody-hand skeleton (open-mmlab#2186)

* update

* update

* Fix NMS Return Index (open-mmlab#2196)

* fix nms index return

* fix

* fix mmseg out_channels=1 (open-mmlab#2179)

* fix

* fix

* fix

* fix

* fix

* fix android library size (open-mmlab#2095)

* fix trt>=8.6 api usage error (open-mmlab#2211)

* add sdk profiler docs (open-mmlab#2189)

* add sdk profiler docs

* fix typos

* [Feature] Enable read zip model in prebuild package. (open-mmlab#2185)

* support libzip

* support windows

* fix

* update rpath

* default build sdk zip model

* Fix torch2onnx for pointpillars with multi-level outputs (open-mmlab#2210)

* temp fix

* fix

* update

* [Fix] Fix text recog task postprocess (open-mmlab#2209)

* update ocr_recog postprocess export

* update ocr_recog sdk postprocess

* fix read unknown_token

* Fix some ut (open-mmlab#2187)

* fix vit ut

* fix

* fix

* update logo

* Revert "update logo"

This reverts commit 2e4fa24.

* fix dcn ut

* fix

* update

* fix

* Revert "fix"

This reverts commit c902cf8.

* fix

* fix

* use torch from docker image

* Revert "use torch from docker image"

This reverts commit fdb4840.

* fix

* update

* update

* update

* update

* Support deploy of YoloX-Pose (open-mmlab#2184)

* dev_mmpose

* tide

* fix lint

* del redundant task and model

* fix

* test ut

* test ut

* upload configs

* fix

* remove debug

* fix lint

* use mmcv.ops.nms

* fix lint

* remove loop

* debug

* test modified ut

* fix lint

* fix return type

* fix

* fix rescale

* fix

* fix pack_result

* update batch inference

* fix nms and pytorch show_box

* fix lint

* modify ut

* add docstring

* modify nms

* fix

* add openvino config

* update docs

* fix test_mmpose

---------

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

* Fix reg test for maskrcnn (open-mmlab#2230)

* Fix yolox-pose ut (open-mmlab#2231)

* update yolox-pose ut

* fix lint

* fix

* bump version to v1.2.0 (open-mmlab#2223)

* add is_resize_mask cfg in inst seg (open-mmlab#2229)

* add is_resize_mask cfg in inst seg

* Update csrc/mmdeploy/codebase/mmdet/instance_segmentation.cpp

set is_resize_mask default fasle

Co-authored-by: Chen Xin <xinchen.tju@gmail.com>

* Update mmdeploy/codebase/mmdet/deploy/object_detection.py

set is_resize_mask default fasle

Co-authored-by: Chen Xin <xinchen.tju@gmail.com>

* fix format

---------

Co-authored-by: Chen Xin <xinchen.tju@gmail.com>

* fix mmdeploy.version_info (open-mmlab#2215)

* fix mmdeploy.version_info

* fix mmdeploy.version_info

* fix lint

---------

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

* fix yolox-pose link (open-mmlab#2248)

* update rknn-toolkit2 version to 1.5 (open-mmlab#2278)

* update rknn-toolkit2 version to 1.5

* fix lint

* cmake ..

* fix int type

* fix

* Update model-support-list  (open-mmlab#2259)

* Support onnxruntime fp16 (open-mmlab#2269)

* support ort-fp16

* update configs

* update

* update reg ci

* fix mmrotate mmdet3d ort fp16

* fix dead links

* Fix rotated nms onnx forward indices err (open-mmlab#2305)

* fix pytorch deepcopy trace error (open-mmlab#2319)

* feat(demo): add time profile (open-mmlab#2307)

* Fix CPP demo code in get_started docs (open-mmlab#2321)

* Translate onnxruntime.md and tensorrt.md (open-mmlab#2320)

* add kwargs to _scaled_dot_product_attention__tensorrt (open-mmlab#2332)

* [Fix] Fix random color for Palette (open-mmlab#2356)

* fix(demo): add warmup for wrong profiling result (open-mmlab#2339)

* fix(demo): add warmup for wrong profiling result

* style(demo/csrc): format cpp code

* fix(demo): compile error

* Update build_ubuntu_x64_ncnn.py

* Update build_ubuntu_x64_ncnn.py

* [Docs] Fix mmseg document (open-mmlab#2363)

* Update mmseg.md

* Update mmseg docs

* remove

* [MMSIG-MMDeploy-557]Support the deployment of 3D object detection algorithm SMOKE on Jetson (open-mmlab#2312)

* [Doc]: Update QR Code (open-mmlab#2371)

* [Feature]: Support deployment of panoptic segmentation models (open-mmlab#2347)

* add semantic segmentation head

 Author:    Daigo Hirooka <daigo.hirooka@gmail.com>
 Date:      Sun Jun 26 18:25:29 2022 +0900

* add panoptic detection model

* add panoptic segmentation configs

* support panoptic-fpn

* remove interpolate

* update

* support panoptic-fpn mask2former maskformer

* update

* support dynamic

* update

* remove unused rewritings for mask2former

* Revert "remove unused rewritings for mask2former"

This reverts commit 2b6d24a.

* update configs and regs

* debug dynamic

* fix for panoptic-fpn

* update

* remove rewritings for mask2former

* update reg test config

* fix

* update docs

* fix comments

* fix

---------

Co-authored-by: Daigo Hirooka <daigo.hirooka@gmail.com>

* fix urls in CI and readthedocs (open-mmlab#2364)

* debug

* fix ut

* fix ut for end2end model

* fix

* update readthedocs

* fix

* fix skip

* enlarge space

* fix mmdet export for coreml backend (open-mmlab#2392)

* fix mmdet export

* fix lint

* [Fix]: fix rtmdet-inst sdk (open-mmlab#2343)

* support rtmdet-inst sdk

* fix batch infer

* fix

* fix mask resize

* fix

* update

* fix segment fault

* fix

* fix lint

* fix

* fix

* fix

* resolve comments

* support deployment of mobilenetv3 (open-mmlab#2256)

* Mobilenet V3 to ONNX support

* Update supported_models.md

* Update supported_models.md

* Update mmpretrain.md

* Update mmpretrain.md

* Update mmpretrain.yml

* Update classification.py

---------

Co-authored-by: RunningLeon <maningsheng@sensetime.com>

* docs(build): add more example (open-mmlab#2378)

* docs(build): add more example

* Update linux-x86_64.md

* Update linux-x86_64.md

* docs(format): update install doc

* Update linux-x86_64.md

* Update linux-x86_64.md

* Update lint.yml

* fix lint

---------

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

* refine gemm in ort mdcn (open-mmlab#2292)

* refine gemm in ort mdcn

* int64_t -> int32_t

* Fix ONNX Runtime Op memory leak (open-mmlab#2406)

* Fix sdk error for multi-gpu execution (open-mmlab#2411)

* Rename MMEdit to MMagic (open-mmlab#2418)

* Support dino from mmdet (open-mmlab#2410)

* detr batch infer

* support dino

* remove dynamic batch

* update doc

* disable exporting masks for image paddings in multi-batch inference

* fix

* remove rewriting and move changes to mmdet

* fix img load for super-resolution in sdk (open-mmlab#2426)

* fix load for super-resolution in sdk

* fix lint

* fix building deformconv for onnxrubtime on windows(open-mmlab#2432)

* fix roi align symbolic function in onnx opset>=16 (open-mmlab#2428)

* fix torch2onnx for mmdet3d (open-mmlab#2425)

* fix mmdet3d

* fix

* resolve comments

* fix rtmdet ncnn

* update docs

* Upgrade CI and prebuild to cu118 (open-mmlab#2447)

* update

* fix

* update docs

* fix

* remove ort-fp16 in reg test

* fix build

* debug

* debug

* Revert "debug"

This reverts commit 2505464.

* Revert "debug"

This reverts commit 7779f0b.

* update

* load ort lib for prebuilt mmdeploy

* support coco-wholebody visualization in pose_tracker python demo (open-mmlab#2450)

* update

* update

* Fix conversion with rtmdet-inst, vit, conformer (open-mmlab#2453)

* fix

* fix scaled_dot_product_attention

* Bump version v1.3.0 (open-mmlab#2451)

* update

* bump version

* export get_simcc_maximum for simcc (open-mmlab#2449)

* update

* update for simcc csrc

* fix docker ci

* update simcc_label

* CodeCamp2023-670 (open-mmlab#2476)

* Update the document of building mmdeploy with ncnn backend on Windows platform

* fix a bug

* fix lint error

* fix lint error

* CodeCamp2023-671 (open-mmlab#2422)

* add nms ops

* add some file

* new file

* some change

* Update nms_match.cpp

* Update nms_match.cpp

* Update __init__.py

* Delete test_onnx_match.onnx

* Delete tests/test_ops/test_onnx_match.onnx

* Update test_nms_match_small.py

* Update test_nms_match_small.py

* Update nms_match.cpp

remove allocate

* Update nms_match.py

remove some test print

* Update test_nms_match_small.py

* Update nms_match.cpp

* Update nms_match.py

* Update test_nms_match_small.py

* fix the lint

fix the lint

* Update test_nms_match_small.py

* Update test_nms_match_small.py

* Update nms_match.cpp

* Update test_nms_match_small.py

* Update test_nms_match_small.py

add input_names

* Update onnxruntime.md

* Update onnxruntime.md

* Update test_nms_match_small.py

* Update onnxruntime.md

* Update onnxruntime.md

* Update test_nms_match_small.py

Add UT in nmsmatch

* Update test_nms_match_small.py

* Update test_nms_match_small.py

* CodeCamp2023-555 (open-mmlab#2469)

* support condinst from mmdet

* remove

* update

* update

* support batch inference

* add condinst head unit testing

* fix lint error

* remove

* fix bug in postprocess

* remove

* update

---------

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

* add condinst ut & update docs (open-mmlab#2481)

* support htc (open-mmlab#2438)

* support htc

* update mmdet.yml

* [Fix] fix the onnx exportation for yoloxpose in mmpose (open-mmlab#2466)

* fix the onnx exportation for yoloxpose

* remove deprecated func

* refine code

* fix the rescaling process of top-down models

* fix ut

* add yoloxpose in regression test

* fix comment

* rebase & fix conflict

* Roman duris/fix jetson wheel platform (open-mmlab#2497)

* generate_build_config.py now selects <any> wheel platform for system flag <jetson>

* Revert "generate_build_config.py now selects <any> wheel platform for system flag <jetson>"

This reverts commit d722bee.

* generate_build_config.py now selects <any> wheel platform for system flag <jetson>

* use different stream when create handle (open-mmlab#2526)

* fix ort wrapper (open-mmlab#2560)

* Fix mask dimensions mismatch in rtmdet-ins models after exporting to onnx (open-mmlab#2574)

* Added the ability to build a project with PyTorch 2.0. (open-mmlab#2553)

* Added the ability to build a project with PyTorch 2.0.
Namely, I added the flag -std=c++17 to extra_compile_args
 depending on the version of Torch.

* Lost the condition for the presence of nvcc

* Lost the condition for the presence of nvcc

* Add parse_version

* fix lint

---------

Co-authored-by: Xin Chen <irexyc@gmail.com>

* There was an error when using a model from the mmyolo repository during image preprocessing, provided that an array was supplied as input. If you input the path to the image, everything works fine. (open-mmlab#2554)

* [Feature] Support ONNX and TensorRT exportation of RTMO models (open-mmlab#2597)

* support ONNX&TensorRT exportation of RTMO

* add configs for rtmo

* replace bbox expansion factor with parameter bbox_padding

* refine code

* refine comment

* apply model.switch_to_deploy in BaseTask.build_pytorch_model

* fix lint

* add rtmo into regression test

* add rtmo with trt backend into regression test

* add rtmo into supported model list

* Fix "has no attribute startswith" (open-mmlab#2498)

The value needs to be converted to string first.

* Fix error when default_scope is None (open-mmlab#2499)

* Fix inference on list of ndarrays (open-mmlab#2506)

* bugfix (open-mmlab#2581)

* fix ci (open-mmlab#2609)

* fix ci

* change to ort fp32

* bump version to v1.13.1 (open-mmlab#2613)

* add new dockerfile, add run cmd

* change Dockerfile for building mmdeploy image

* ls fixes to use new mmpose image

* add branch for building image

* fix

* remove 1.3.2-branch

* add mmpose to mmdeploy repo

* remove mmpose branch checkout

---------

Co-authored-by: tpoisonooo <khj.application@aliyun.com>
Co-authored-by: Chen Xin <xinchen.tju@gmail.com>
Co-authored-by: RunningLeon <mnsheng@yeah.net>
Co-authored-by: huayuan4396 <110151316+huayuan4396@users.noreply.github.com>
Co-authored-by: Carmine Ianni <ianni.carmine@outlook.com>
Co-authored-by: huangjiyi <43315610+huangjiyi@users.noreply.github.com>
Co-authored-by: Chen Xin <irexyc@gmail.com>
Co-authored-by: ruisv <27017322+ruisv@users.noreply.github.com>
Co-authored-by: xiabo123 <40659418+xiabo123@users.noreply.github.com>
Co-authored-by: Tau <taujiang@outlook.com>
Co-authored-by: AllentDan <41138331+AllentDan@users.noreply.github.com>
Co-authored-by: YH <yhna940@gmail.com>
Co-authored-by: fanyunchao <55267877+fanyunchao@users.noreply.github.com>
Co-authored-by: blue-q <85726493+blue-q@users.noreply.github.com>
Co-authored-by: Tianlong Ai <50650583+AI-Tianlong@users.noreply.github.com>
Co-authored-by: zhjunqin <zhjunqin@users.noreply.github.com>
Co-authored-by: zhengjie.xu <jerryxuzhengjie@gmail.com>
Co-authored-by: Daigo Hirooka <daigo.hirooka@gmail.com>
Co-authored-by: mpocwierz <maciej.pocwierz@semantive.com>
Co-authored-by: RunningLeon <maningsheng@sensetime.com>
Co-authored-by: q.yao <streetyao@live.com>
Co-authored-by: Junhwa Song <ethan9867@gmail.com>
Co-authored-by: Chen Xin <chenxin@pjlab.org.cn>
Co-authored-by: Yang Nie <lenny.n@outlook.com>
Co-authored-by: CescMessi <tommymuniesa@gmail.com>
Co-authored-by: nino <shuyaoyimei@users.noreply.github.com>
Co-authored-by: yinfan98 <32722923+yinfan98@users.noreply.github.com>
Co-authored-by: Renhao Liu <69111729+Boomerl@users.noreply.github.com>
Co-authored-by: Peng Lu <penglu2097@gmail.com>
Co-authored-by: roman-duris <141549520+roman-duris@users.noreply.github.com>
Co-authored-by: LuukvandenBent <94459309+LuukvandenBent@users.noreply.github.com>
Co-authored-by: Danil <Danil328@yandex.ru>
Co-authored-by: Mohammed Yasin <32206511+Y-T-G@users.noreply.github.com>
Co-authored-by: Yang Nie <flytocc@outlook.com>
Co-authored-by: Andrii-Sheba <andriy.shebeko@logivations.com>

.drone.yml

@@ -2,7 +2,7 @@ kind: pipeline
 type: docker
 name: Label Studio main Build
 
-platform: 
+platform:
   arch: amd64
   os: linux
 
@@ -26,7 +26,7 @@ steps:
       registry: quay.io
       repo: quay.io/logivations/ml_all
       privileged: true
-      build_args: 
+      build_args:
         - BUILDKIT_INLINE_CACHE=1
       no_cache: true
       tags:
@@ -43,7 +43,7 @@ kind: pipeline
 type: docker
 name: PR Pre-commit check
 
-platform: 
+platform:
   arch: amd64
   os: linux
 
@@ -67,7 +67,7 @@ steps:
       registry: quay.io
       repo: quay.io/logivations/ml_all
       privileged: true
-      build_args: 
+      build_args:
         - BUILDKIT_INLINE_CACHE=1
       cache_from:
         - quay.io/logivations/ml_all:LS_mmdeploy_pr${DRONE_PULL_REQUEST}

.github/md-link-config.json

@@ -15,6 +15,9 @@
     },
     {
       "pattern": "^http://localhost"
+    },
+    {
+      "pattern": "^https://openmmlab.com"
     }
   ],
   "httpHeaders": [

.github/scripts/linux/test_full_pipeline.sh

@@ -58,7 +58,7 @@ if [ $backend == "trt" ]; then
 fi
 
 # prepare dataset
-wget -P data/ https://github.com/open-mmlab/mmdeploy/files/9401216/imagenet-val100.zip
+wget -P data/ https://github.com/open-mmlab/mmdeploy/releases/download/v0.1.0/imagenet-val100.zip
 unzip data/imagenet-val100.zip -d data/
 
 echo "Running test with $backend"

.github/scripts/prepare_reg_test.py

@@ -70,19 +70,19 @@ def run_cmd(cmd_lines, log_path=None, raise_error=True):
 
 
 def prepare_codebases(codebases):
-    run_cmd(['python -m mim install "mmcv>=2.0.0"'])
     for codebase in codebases:
         full_name = REPO_NAMES[codebase]
         target_dir = os.path.join(MMDEPLOY_DIR, '..', full_name)
         branch = 'main'
-        if codebase in ['mmrotate', 'mmdet3d']:
-            branch = 'dev-1.x'
-        if not osp.exists(target_dir):
-            run_cmd([
-                'git clone --depth 1 ', f'-b {branch} '
-                f'https://github.com/open-mmlab/{full_name}.git '
-                f'{target_dir} '
-            ])
+        if codebase == 'mmrotate':
+            branch = 'v1.0.0rc1'
+        if osp.exists(target_dir):
+            shutil.rmtree(target_dir)
+        run_cmd([
+            'git clone --depth 1 ', f'-b {branch} '
+            f'https://github.com/open-mmlab/{full_name}.git '
+            f'{target_dir} '
+        ])
         run_cmd([
             'python -m mim install ',
             f'-r {target_dir}/requirements/mminstall.txt ',
@@ -98,6 +98,12 @@ def prepare_codebases(codebases):
                             f'{MMDEPLOY_DIR}/configs/mmyolo')
             shutil.copy(f'{target_dir}/tests/regression/mmyolo.yml',
                         f'{MMDEPLOY_DIR}/tests/regression/mmyolo.yml')
+        elif codebase == 'mmdet':
+            # for panoptic
+            run_cmd([
+                'python -m pip install ',
+                'git+https://github.com/cocodataset/panopticapi.git',
+            ])
 
 
 def install_torch(torch_version):
@@ -106,8 +112,12 @@ def install_torch(torch_version):
     tv = version.parse(torch_version)
     if tv < version.parse('1.10.0'):
         cuda_int = '111'
-    elif tv >= version.parse('1.13.0'):
+    elif tv < version.parse('1.13.0'):
+        cuda_int = '113'
+    elif tv < version.parse('2.0.0'):
         cuda_int = '117'
+    else:
+        cuda_int = '118'
 
     is_torch_v2 = tv >= version.parse('2.0.0')
     if is_torch_v2:

.github/scripts/test_onnx2ncnn.py

@@ -19,13 +19,13 @@
         'mmpretrain/configs/resnet/resnet18_8xb32_in1k.py',
         'https://download.openmmlab.com/mmclassification/v0/resnet/resnet18_8xb32_in1k_20210831-fbbb1da6.pth',  # noqa: E501
         'resnet18.onnx',
-        'https://media.githubusercontent.com/media/tpoisonooo/mmdeploy-onnx2ncnn-testdata/main/resnet18.onnx',  # noqa: E501
+        'https://github.com/open-mmlab/mmdeploy/releases/download/v0.1.0/resnet18.onnx',  # noqa: E501
     ),
     (
         'mmpretrain/configs/mobilenet_v2/mobilenet-v2_8xb32_in1k.py',
         'https://download.openmmlab.com/mmclassification/v0/mobilenet_v2/mobilenet_v2_batch256_imagenet_20200708-3b2dc3af.pth',  # noqa: E501
         'mobilenet-v2.onnx',
-        'https://media.githubusercontent.com/media/tpoisonooo/mmdeploy-onnx2ncnn-testdata/main/mobilenet-v2.onnx',  # noqa: E501
+        'https://github.com/open-mmlab/mmdeploy/releases/download/v0.1.0/mobilenet-v2.onnx',  # noqa: E501
     )
 ]
 

.github/scripts/windows/test_full_pipeline.ps1

@@ -56,7 +56,7 @@ python tools\deploy.py `
   --dump-info
 
 # prepare dataset
-Invoke-WebRequest -Uri https://github.com/open-mmlab/mmdeploy/files/9401216/imagenet-val100.zip -OutFile $pwd\data\imagenet-val100.zip
+Invoke-WebRequest -Uri https://github.com/open-mmlab/mmdeploy/releases/download/v0.1.0/imagenet-val100.zip -OutFile $pwd\data\imagenet-val100.zip
 Add-Type -AssemblyName System.IO.Compression.FileSystem ; [System.IO.Compression.ZipFile]::ExtractToDirectory("$pwd\data\imagenet-val100.zip", "$pwd\data\")
 
 Write-Host "Running test with ort"