Complete Guide to Installing and Setting Up ControlNet Plugin for Automatic1111

2023-06-20 02:32:51
In this article, Professor Shannon will explain the installation method of ControlNet, the most important plug-in of Automatic1111. ControlNet provides very powerful functions, but because it is also a manager of a complete AI generation model, the installation step is one more step than the general plug-in. First, as described in Stable Diffusion Basics — Extensions, you must first switch to the Extension tab, click the Available sub-tab, click the Load from button to read the list of extensions, and then find the ControlNet item. Click the Install button to install. This basic ControlNet installation is complete. The official name of ControlNet is sd-webui-controlnet, but this is not enough. ControlNet will use many special models to process the input pictures and parameters. These additional models have a small size of 80MB and a large size of 1.5GB, totaling about There are more than a dozen of them, all of which will occupy more than 10GB of hard disk space when downloaded. The main ControlNet models are placed on the download page of Huggingface: For the models you want, there is a download button next to them to download. After downloading, transfer these model files with the extension .pth to the folder of Automatic1111: /extensions/ sd-webui-controlnet/models/ And if you need a model like t2iadapter, you need to go to the Huggingface webpage of the model author to download it in the same way: After the installation is complete, please go to the Settings tab, and look for the ControlNet project on the left, on this page Find the Multi ControlNet: Max models amount (requires restart) item, and adjust it to a value of at least 3, and then restart Automatic1111, so that multiple ControlNet models can be manipulated. Multi ControlNet: Max models amount I usually select four of these models and download and complete the settings before Automatic1111 can be regarded as a productive tool!
1687231196
#Stable #Diffusion #BasicsInstalling #ControlNet