【CANN训练营第三季】特邀直播openmmlab

pip install torch&＃61;&＃61;1.8.1 torchvision&＃61;&＃61;0.9.1 --extra-index-url https://download.pytorch.org/whl/cpu
pip install openmim
mim install mmcv-full
sudo apt-get install libopencv-dev
git clone --recursive https://github.com/open-mmlab/mmdeploy.git
cd mmdeploy
#安装 mmdeploy 楼型转换器
pip install -v -e .
#编译 mmdeploy 推理 SDK
## 设置 ascend-toolkit 的环境变量
#source /usr/local/Ascend/ascend-toolkit/set_env.sh
source /home/HwHiAiUser/Ascend/ascend-toolkit/set_env.shmkdir -p build && cd build
cmake .. -DMMDEPLOY_BUILD_SDK&＃61;ON -DMMDEPLOY_BUILD_SDK_PYTHONAPI&＃61;ON -DMMDEPLOY_TARGET_BACKEND&＃61;acl
make -j$(nproc) && make install
cd ..

2、编写测试脚本&＃xff0c;验证resnet50、retinanet适配cann的转换和推理过程

#安装 openmmlab 算法库 mmcls
pip install mmcls
# 使用 mim 下载 mmcls 中 resnet18 的模型和 ckpt
mim download mmcls --config resnet18_8xb32_in1k --dest .
#执行 mmdeploy 的转换命令&＃xff0c;得到 om 模型以及 mmdeploy model
python tools/deploy.py \
configs/mmcls/classification_ascend_static-224x224.py \
resnet18_8xb32_in1k.py \
resnet18_8xb32_in1k_20210831-fbbb1da6.pth \
tests/data/tiger.jpeg \
--work-dir mmdeploy_models/mmcls/resnet18/cann \
--device cpu\
--dump-info


fast-rnn

#安装 openmmlab 算法库 mmdet
pip install mmdet
#使用mim 下载 mmdet 中 faster r-cnn 的模型和ckpt
mim download mmdet --config faster_rcnn_r50_fpn_1x_coco --dest .
#执行 mmdeploy 的转换命令&＃xff0c;得到 .om 模型以及 mmdeploy model
python tools/deploy.py \
configs/mmdet/detection/detection_ascend_static-800x1344.py \
faster_rcnn_r50_fpn_1x_coco.py \
faster_rcnn_r5e_fpn_1x_coco20200130-047c8118.pth \
demo/resources/det.jpg \
--work-dir mmdeploy_models/mmdet/faster-rcnn/cann \
--device cpu \
--dump-info

3、升级到 CANN 6.0&＃xff0c;验证 mmdeploy 部署功能的正确性

直接在cann6.0测试即可