From f05a1f954b57572a972798fe72a427f56bc2b6ae Mon Sep 17 00:00:00 2001 From: zstar <65890619+zstar1003@users.noreply.github.com> Date: Mon, 16 Jun 2025 18:15:08 +0800 Subject: [PATCH] =?UTF-8?q?docs(faq):=20=E6=9B=B4=E6=96=B0=20GPU=20?= =?UTF-8?q?=E5=8A=A0=E9=80=9F=E7=9B=B8=E5=85=B3=E9=97=AE=E9=A2=98=E7=9A=84?= =?UTF-8?q?=E8=A7=A3=E5=86=B3=E6=96=B9=E6=A1=88?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit --- README.md | 3 +++ README_EN.md | 2 ++ docs/question/README.md | 15 +++++++++++++++ docs/skill/README.md | 9 +++++++-- 4 files changed, 27 insertions(+), 2 deletions(-) diff --git a/README.md b/README.md index c8796b7..059b89f 100644 --- a/README.md +++ b/README.md @@ -3,8 +3,11 @@
+ stars 版本 许可证 + docker pulls +

🇨🇳 中文 | diff --git a/README_EN.md b/README_EN.md index 9a84a93..c5c15c0 100644 --- a/README_EN.md +++ b/README_EN.md @@ -3,8 +3,10 @@

+ stars Version License + docker pulls

🇨🇳 Chinese | diff --git a/docs/question/README.md b/docs/question/README.md index 3f04697..4f30452 100644 --- a/docs/question/README.md +++ b/docs/question/README.md @@ -50,6 +50,21 @@ netsh int ipv4 add excludedportrange protocol=tcp startport=5455 numberofport net start winnat ``` +## 问题 9:MinerU GPU 加速似乎只调用了第一张显卡,如何指定其它显卡? + +**回答:** MinerU 1.x 本身无法指定具体所用显卡,且不支持多显卡部署,可通过以下方式去限定后端容器所能利用的显卡id。 + +修改`docker\docker-compose_gpu.yml`: + +```bash +deploy: + resources: + reservations: + devices: + - driver: nvidia + capabilities: [gpu] + device_ids: ["2"] # 使用索引号指定id为2的显卡 +``` --- diff --git a/docs/skill/README.md b/docs/skill/README.md index 0bbc2fd..77b3c22 100644 --- a/docs/skill/README.md +++ b/docs/skill/README.md @@ -12,7 +12,13 @@ docker启动: docker compose -f docker/docker-compose_gpu.yml up -d ``` -若启动失败,可通过以下方式增加docker对gpu加速的支持。 +若启动后,发现容器找不到显卡信息,则需要再单独安装nvidia-container-runtime: + +```bash +sudo apt install nvidia-container-runtime +``` + +若上述未能解决,可考虑使用以下备用方案: ```bash # 575为具体版本号,可根据具体gpu型号选择合适的版本 @@ -20,7 +26,6 @@ sudo apt install nvidia-cuda-toolkit sudo apt install nvidia-container-toolkit sudo apt install nvidia-fabricmanager-575 sudo apt install libnvidia-nscq-575 -sudo apt install nvidia-container-runtime sudo systemctl start nvidia-fabricmanager sudo systemctl enable nvidia-fabricmanager sudo systemctl status nvidia-fabricmanager