--- tags: - Face - ReActorFace --- # Build Blended Face Model 🌌 ReActor ## Documentation - Class name: `ReActorBuildFaceModel` - Category: `🌌 ReActor` - Output node: `True` This node is designed to construct a face model from a provided image, utilizing face detection and analysis to identify and extract facial features. It aims to create a detailed face model that can be used for various face manipulation tasks within the ReActor framework. ## Input types ### Required - **`save_mode`** - Indicates whether the face model should be saved after creation, affecting the workflow of model handling and storage. - Comfy dtype: `BOOLEAN` - Python dtype: `bool` - **`send_only`** - This parameter determines if only the face model's name should be sent back, without the actual model data, to optimize data transfer. - Comfy dtype: `BOOLEAN` - Python dtype: `bool` - **`face_model_name`** - The name under which the face model will be saved. This parameter is essential for identifying and retrieving the model in subsequent operations. - Comfy dtype: `STRING` - Python dtype: `str` - **`compute_method`** - Defines the method used for computing the face model, allowing for flexibility in the face detection and analysis process. - Comfy dtype: `COMBO[STRING]` - Python dtype: `str` ### Optional - **`images`** - A batch of images from which to build face models, enabling the processing of multiple images in a single operation. - Comfy dtype: `IMAGE` - Python dtype: `List[Image.Image]` - **`face_models`** - A collection of existing face models to be processed or analyzed further, providing a way to handle multiple models efficiently. - Comfy dtype: `FACE_MODEL` - Python dtype: `List[Dict[str, Any]]` ## Output types - **`face_model`** - Comfy dtype: `FACE_MODEL` - The constructed face model or an error message if the process fails. - Python dtype: `Dict[str, Any]` ## Usage tips - Infra type: `CPU` - Common nodes: unknown ## Source code ```python # Built-in or C extension class, unable to automatically detect source code ```