Inputs
| Parameter | Description | Comfy dtype |
|---|---|---|
clip_vision | Processes visual data to align with the model’s requirements, enhancing the model’s understanding of visual context. | CLIP_VISION |
init_image | Serves as the initial image input for the model, setting the baseline for further image-based operations. | IMAGE |
vae | Integrates variational autoencoder outputs, facilitating the model’s ability to generate or modify images. | VAE |
width | Specifies the width of the output image, allowing for dynamic resizing according to model needs. | INT |
height | Determines the height of the output image, enabling customization of the output dimensions. | INT |
batch_size | Controls the number of images processed in a single batch, optimizing computational efficiency. | INT |
elevation | Adjusts the elevation angle for 3D model rendering, enhancing the model’s spatial understanding. | FLOAT |
azimuth | Modifies the azimuth angle for 3D model visualization, improving the model’s perception of orientation. | FLOAT |
Outputs
| Parameter | Description | Data Type |
|---|---|---|
positive | Generates positive conditioning vectors, aiding in the model’s positive feature reinforcement. | CONDITIONING |
negative | Produces negative conditioning vectors, assisting in the model’s avoidance of certain features. | CONDITIONING |
latent | Creates latent representations, facilitating deeper model insights into the data. | LATENT |
This documentation was AI-generated. If you find any errors or have suggestions for improvement, please feel free to contribute! Edit on GitHub