docs(faq): 更新 GPU 加速相关问题的解决方案
This commit is contained in:
parent
5d718428c4
commit
f05a1f954b
|
@ -3,8 +3,11 @@
|
|||
</div>
|
||||
|
||||
<div align="center">
|
||||
<a href="https://github.com/zstar1003/ragflow-plus/stargazers"><img src="https://img.shields.io/github/stars/zstar1003/ragflow-plus?style=social" alt="stars"></a>
|
||||
<img src="https://img.shields.io/badge/版本-0.4.3-blue" alt="版本">
|
||||
<a href="LICENSE"><img src="https://img.shields.io/badge/许可证-AGPL3.0-green" alt="许可证"></a>
|
||||
<a href="https://hub.docker.com/r/zstar1003/ragflowplus/tags"><img src="https://img.shields.io/docker/pulls/zstar1003/ragflowplus" alt="docker pulls"></a>
|
||||
|
||||
<h4>
|
||||
<a href="README.md">🇨🇳 中文</a>
|
||||
<span> | </span>
|
||||
|
|
|
@ -3,8 +3,10 @@
|
|||
</div>
|
||||
|
||||
<div align="center">
|
||||
<a href="https://github.com/zstar1003/ragflow-plus/stargazers"><img src="https://img.shields.io/github/stars/zstar1003/ragflow-plus?style=social" alt="stars"></a>
|
||||
<img src="https://img.shields.io/badge/version-0.4.3-blue" alt="Version">
|
||||
<a href="LICENSE"><img src="https://img.shields.io/badge/license-AGPL3.0-green" alt="License"></a>
|
||||
<a href="https://hub.docker.com/r/zstar1003/ragflowplus/tags"><img src="https://img.shields.io/docker/pulls/zstar1003/ragflowplus" alt="docker pulls"></a>
|
||||
<h4>
|
||||
<a href="README.md">🇨🇳 Chinese</a>
|
||||
<span> | </span>
|
||||
|
|
|
@ -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
|
||||
```
|
||||
|
||||
---
|
||||
|
||||
|
|
|
@ -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
|
||||
|
|
Loading…
Reference in New Issue