自行下载funasr-cpu/gpu和asrmodel-deploy-cpu/gpu压缩包,官方都有
cpu版
运行:
cd /home/user/data/AIModels/FunASR_docker/deploy_resources
sudo docker load -i funasr-runtime-sdk-cpu-0.4.4_w_models_v0.2.1.tar
cd asrmodel-deploy-cpu
sudo bash docker_start.sh
改名:sudo docker rename <容器ID> funasr-cpu
查信息:sudo docker inspect funasr-cpu
工作区:sudo docker exec -it funasr-cpu /bin/bash
挂载点对应目录:cd /workspace
单条推理:
cd /workspace/funasr_samples/python
bash funasr_wss.sh
批量推理:
音频文件放到/workspace/funasr_samples/audio/acc1中,可自定义
cd /workspace/funasr_samples/python
bash multi_funasr_wss.sh
gpu版
cd /home/user/data/AIModels/FunASR_docker/deploy_resources
sudo docker load -i funasr-runtime-sdk-gpu-0.1.1_w_models_v0.1.1.tar
cd asrmodel-deploy-gpu
sudo bash docker_start.sh
改名:sudo docker rename <容器ID> funasr-gpu
查信息:sudo docker inspect funasr-gpu
工作区:sudo docker exec -it funasr-gpu /bin/bash
挂载点对应目录:cd /workspace
单条推理:
cd /workspace/funasr_samples/python
bash funasr_wss.sh
批量推理:
音频文件放到/workspace/funasr_samples/audio/acc1中,可自定义
cd /workspace/funasr_samples/python
bash multi_funasr_wss.sh
其它
向docker传文件(docker外环境):
sudo docker cp /home/user/data/AIModels/FunASR_docker/deploy_resources/xxx funasr-gpu:/workspace/funasr_samples/audio/