Skip to main content
This node is designed for generating conditioning data for video generation tasks, specifically tailored for use with SVD_img2vid models. It takes various inputs including initial images, video parameters, and a VAE model to produce conditioning data that can be used to guide the generation of video frames.

Inputs

ParameterDescriptionComfy dtype
clip_visionRepresents the CLIP vision model used for encoding visual features from the initial image, playing a crucial role in understanding the content and context of the image for video generation.CLIP_VISION
init_imageThe initial image from which the video will be generated, serving as the starting point for the video generation process.IMAGE
vaeA Variational Autoencoder (VAE) model used for encoding the initial image into a latent space, facilitating the generation of coherent and continuous video frames.VAE
widthThe desired width of the video frames to be generated, allowing for customization of the video’s resolution.INT
heightThe desired height of the video frames, enabling control over the video’s aspect ratio and resolution.INT
video_framesSpecifies the number of frames to be generated for the video, determining the video’s length.INT
motion_bucket_idAn identifier for categorizing the type of motion to be applied in the video generation, aiding in the creation of dynamic and engaging videos.INT
fpsThe frames per second (fps) rate for the video, influencing the smoothness and realism of the generated video.INT
augmentation_levelA parameter controlling the level of augmentation applied to the initial image, affecting the diversity and variability of the generated video frames.FLOAT

Outputs

ParameterDescriptionComfy dtype
positiveThe positive conditioning data, consisting of encoded features and parameters for guiding the video generation process in a desired direction.CONDITIONING
negativeThe negative conditioning data, providing a contrast to the positive conditioning, which can be used to avoid certain patterns or features in the generated video.CONDITIONING
latentLatent representations generated for each frame of the video, serving as a foundational component for the video generation process.LATENT
This documentation was AI-generated. If you find any errors or have suggestions for improvement, please feel free to contribute! Edit on GitHub