Have TOTAL CONTROL with this AI Animation Workflow in AnimateLCM! // Civitai Vid2Vid Tutorial Stream

The speaker, Tyler, welcomes the audience to the Civii office hours and introduces himself. He explains that the session will be a tutorial or walkthrough of new animate diff workflows that he has recently released on his Civii profile. Tyler mentions that he will share links to these workflows in the Twitch and Discord chats. He also discusses the importance of having a certain amount of VRAM for the workflows and suggests that viewers follow along and ask questions.

Tyler guides the audience on how to navigate the workflow, especially for first-time users. He advises viewers to zoom out to see the entire workflow and move nodes if necessary. He emphasizes the importance of following the order of operation in the workflow and explains the purpose of each group, including the video source, resolution settings, and the use of different models. Tyler also discusses the benefits of using the Photon LCM model and the stable diffusion 1.5 LCM Laura for efficient video rendering.

Tyler explains the use of control nets in the workflow, detailing how they can be toggled on and off using fast bypassers. He shares his preference for using depth and open pose control nets combined with the control GIF control net for smoother animations. Tyler also provides tips on using hotkeys for bypassing nodes and emphasizes the importance of utilizing fast bypassers for efficiency.

Tyler discusses the innovative aspect of the workflow that involves using two separate IP adapters – one for the subject or character and another for the background. He explains the benefits of this approach, which allows for more control over the AI-generated content. Tyler also provides guidance on selecting the right images for the IP adapters and emphasizes the importance of having a clear alpha mask for the subject.

Tyler addresses a common issue faced by users when installing the reactor face swapper node. He explains that users need to have Visual Studio Code and C++ installed to run the reactor. Tyler provides a link to the IP adapter plus GitHub page for downloading necessary models and thanks the community members who contributed to solving the installation problem.

Tyler demonstrates the power of the workflow by combining different characters and backgrounds provided by the audience. He explains the process of selecting images and using control nets to achieve the desired outcome. Tyler also discusses the impact of image quality and texture on the final result, highlighting the importance of high-quality IP adapter images.

Tyler talks about the upscaling process using the highres fix or upscaler. He shares his preferred settings for upscaling and discusses the potential for CUDA errors when using the NN latent upscaler. Tyler also mentions alternative methods for upscaling, such as using Topaz for social media-ready quality.

Tyler compares the LCM and V3 workflows by running the same character animation through both. He discusses the differences in rendering time and output quality, sharing his preference for the LCM workflow in this instance. Tyler encourages viewers to experiment with both workflows to find the one that best suits their needs.

Tyler announces the addition of a fifth streaming day, featuring guest streams with specialists in various fields. He expresses excitement about the upcoming guests and encourages viewers to follow the channels for updates. Tyler also invites viewers to share their creations using the workflows and to engage with the community on social media platforms.

Tyler concludes the tutorial by reiterating the importance of the alpha mask and the right models for the workflow. He encourages viewers to replay the tutorial as needed and reassures them that the workflow is designed to be straightforward. Tyler also promotes the Civii community and social media channels, thanking viewers for their participation and support.