We recommend using the `cuda128` environment for faster inference speed. If you need to use the `cuda124` environment, you can use image versions with the `-cu124` suffix:
```bash
docker pull lightx2v/lightx2v:25091903-cu124
docker pull lightx2v/lightx2v:25092201-cu124
```
#### 2. Run Container
...
...
@@ -54,7 +54,7 @@ For mainland China, if the network is unstable when pulling images, you can pull