-
Notifications
You must be signed in to change notification settings - Fork 0
RunPod: Deploy and launch a pod
- Create a RunPod account
- Add credits to your RunPod account
You need to select the Secure Stable Diffusion by Storj template. The easiest way to do it is by opening the below link in your web browser:
https://runpod.io/gsc?template=3jgnr4mriw&ref=se98m3j8
This will open the RunPod dashboard and select the "Secure Stable Diffusion by Storj" template.
As this RunPod template is focused on security, we recommend deploying it on a VM provided by the Secure Cloud. As an additional benefit, these servers usually have better network speeds than the VMs from the Community Cloud, which saves you time and money when loading the Docker image and AI models during startup.
There are plenty of VMs to select from. The RTX A5000 GPU is good enough for training and inference with this template. But feel free to experiment with other GPUs.
Click on the Deploy button of the selected VM.
By default, the Docker container will have 10GB of Container Disk (Temporary) storage and 20GB of Volume Disk (Persistent) storage. This should be enough for basic scenarios. If you want to load a large number of AI models and train a lot of checkpoints, consider increasing the size of the Volume Disk. This is done with the Customize Deployment button.
As this RunPod template is focused on security, we recommend turning the Encrypted Volume checkbox on.
You can choose between On Demand (Non-Interruptible) and Spot (Interruptible) deployment. The Spot option is a good way to save money if you don't mind your VM being accidentally stopped. It would be easy to start it up again. All data will be preserved on the Volume Disk.
Optionally, configure the pod to download your AI models from Storj automatically.
Review your selection and deploy the pod using the Deploy button.
The pod is now starting. You can observe the logs using the Logs button.