参考此帖进行部署香橙派AI PRO 20T(昇腾310B1)使用docker容器教程_香橙派_昇腾论坛

发现以前的310B的镜像仓库已经没了，转用此仓库镜像ascend-infer

后使用帖子中的docker启动容器命令

docker run -it --net=host \ 

--device=/dev/upgrade:/dev/upgrade \ 

--device=/dev/davinci0:/dev/davinci0 \ 

--device=/dev/davinci_manager \ 

--device=/dev/vdec:/dev/vdec \ 

--device=/dev/vpc:/dev/vpc \ 

--device=/dev/pngd:/dev/pngd \ 

--device=/dev/venc:/dev/venc \ 

--device=/dev/sys:/dev/sys \ 

--device=/dev/svm0 \ 

--device=/dev/ts_aisle:/dev/ts_aisle \ 

--device=/dev/dvpp_cmdlist:/dev/dvpp_cmdlist \ 

-v /etc/sys_version.conf:/etc/sys_version.conf:ro \ 

-v /etc/hdcBasic.cfg:/etc/hdcBasic.cfg:ro \ 

-v /usr/lib64/libaicpu_processer.so:/usr/lib64/libaicpu_processer.so:ro \ 

-v /usr/lib64/libaicpu_prof.so:/usr/lib64/libaicpu_prof.so:ro \ 

-v /usr/lib64/libaicpu_sharder.so:/usr/lib64/libaicpu_sharder.so:ro \ 

-v /usr/lib64/libadump.so:/usr/lib64/libadump.so:ro \ 

-v /usr/lib64/libtsd_eventclient.so:/usr/lib64/libtsd_eventclient.so:ro \ 

-v /usr/lib64/libaicpu_scheduler.so:/usr/lib64/libaicpu_scheduler.so:ro \ 

-v /usr/lib/aarch64-linux-gnu/libcrypto.so.1.1:/usr/lib/aarch64-linux-gnu/libcrypto.so.1.1:ro \ 

-v /usr/lib/aarch64-linux-gnu/libyaml-0.so.2:/usr/lib/aarch64-linux-gnu/libyaml-0.so.2:ro \ 

-v /usr/lib64/libdcmi.so:/usr/lib64/libdcmi.so:ro \ 

-v /usr/lib64/libmpi_dvpp_adapter.so:/usr/lib64/libmpi_dvpp_adapter.so:ro \ 

-v /usr/lib64/aicpu_kernels/:/usr/lib64/aicpu_kernels/:ro \ 

-v /usr/local/sbin/npu-smi:/usr/local/sbin/npu-smi:ro \ 

-v /usr/lib64/libstackcore.so:/usr/lib64/libstackcore.so:ro \ 

-v /usr/local/Ascend/driver/lib64:/usr/local/Ascend/driver/lib64:ro \ 

-v /var/slogd:/var/slogd:ro \ 

-v /var/dmp_daemon:/var/dmp_daemon:ro \ 

-v /etc/slog.conf:/etc/slog.conf:ro \ 

--name ascend-infer-310b \ 

swr.cn-south-1.myhuaweicloud.com/ascendhub/ascend-infer-310b:24.0.RC1-dev-arm \ 

/bin/bash 

报错如下：

安装帖子昇腾镜像仓库中310B镜像挂载后npu-smi命令报错_昇腾主版块_昇腾论坛中的解决方法：

mkdir /dev/shm/dmp

nohup /var/dmp_daemon -I -M -U 8087 >&/dev/null &

发现环境在启动/var/dmp_daemon缺少环境配置，报错如下：

然后自己修改了docker的启动命令：

(base) root@davinci-mini:~# docker run -it --name test12 \ 

  --privileged \ 

  --network=host \ 

  --shm-size=50g \ 

  --user=root \ 

  --device=/dev/davinci0 \ 

  --device=/dev/davinci_manager \ 

  --device=/dev/davinci_manager_docker \ 

  --device=/dev/ascend_manager \ 

  --device=/dev/devmm_svm \ 

  --device=/dev/hisi_hdc \ 

  --device=/dev/upgrade \ 

  --device=/dev/vdec \ 

  --device=/dev/vpc \ 

  --device=/dev/venc \ 

  --device=/dev/ai \ 

  --device=/dev/ao \ 

  --device=/dev/vo \ 

  --device=/dev/hdmi \ 

  --device=/dev/dvpp_cmdlist \ 

  -v /usr/local/sbin/npu-smi:/usr/local/sbin/npu-smi:ro \ 

  -v /usr/local/Ascend:/usr/local/Ascend:ro \ 

  -v /var/log:/var/log \ 

  -v /dev/shm:/dev/shm \ 

  -v /etc/sys_version.conf:/etc/sys_version.conf:ro \ 

  -v /etc/hdcBasic.cfg:/etc/hdcBasic.cfg:ro \ 

  -v /usr/lib64:/usr/lib64:ro \ 

  -v /usr/local/dcmi:/usr/local/dcmi:ro \ 

  -v /var/davinci:/var/davinci \ 

  -v /var/slogd:/var/slogd \ 

  -v /var/dmp_daemon:/var/dmp_daemon \ 

  -v /lib/firmware:/lib/firmware:ro \ 

  -e LD_LIBRARY_PATH=/usr/local/Ascend/driver/lib64:/usr/local/Ascend/ascend-toolkit/latest/lib64:$LD_LIBRARY_PATH \ 

  -e PATH=/usr/local/sbin:/usr/local/bin:/usr/local/Ascend/ascend-toolkit/latest/bin:$PATH \ 

  -e ASCEND_VISIBLE_DEVICES=0 \ 

  --entrypoint=/bin/bash \ 

  swr.cn-south-1.myhuaweicloud.com/ascendhub/ascend-infer:24.0.0-ubuntu20.04 

现可以调到npu-smi info，但是还是有一些报错，不知怎么解决：

注：

请问现在有没有官方可以直接使用的310B镜像哈？