ComfyUI Node: Batch Prompt Schedule (Latent Input) 📅🅕🅝

Batch Prompt Schedule (Latent Input) 📅🅕🅝:

The BatchPromptScheduleLatentInput node is designed to streamline the process of applying composable diffusion across a batch of latent inputs, specifically for animation prompts. This node takes in a set of animation prompts, processes them to remove unnecessary whitespace and newlines, and then splits them into positive and negative prompts. It interpolates the weights of these prompts over frames, ensuring smooth transitions and consistent conditioning across the batch. By leveraging this node, you can achieve more dynamic and coherent animations, as it handles the complex task of prompt scheduling and conditioning seamlessly.

Batch Prompt Schedule (Latent Input) 📅🅕🅝 Input Parameters:

settings

The settings parameter is an instance of ScheduleSettings that contains various configuration options for the prompt scheduling process. This includes text prompts, pre-text, and app-text, which are used to generate and split the animation prompts. The settings also dictate how the weights are interpolated over frames, impacting the smoothness and consistency of the animation. This parameter is crucial for customizing the behavior of the node to fit specific animation needs.

clip

The clip parameter represents the clip model used for conditioning. It is essential for applying composable diffusion across the batch, as it influences how the prompts are interpreted and conditioned. The clip model helps in generating more accurate and contextually relevant animations by providing a robust framework for prompt conditioning.

latents

The latents parameter is the batch of latent inputs that will be processed by the node. These latents are the initial states that will be conditioned and transformed based on the interpolated prompts. The quality and characteristics of the final animation heavily depend on the input latents, making this parameter a key component of the node's functionality.

Batch Prompt Schedule (Latent Input) 📅🅕🅝 Output Parameters:

p

The p output parameter represents the positive conditioning applied to the batch of latents. This includes the current and next positive prompts, as well as the interpolated weights. The positive conditioning helps in enhancing the desired features and characteristics in the animation, making it more aligned with the intended prompts.

n

The n output parameter represents the negative conditioning applied to the batch of latents. Similar to the positive conditioning, this includes the current and next negative prompts, along with the interpolated weights. The negative conditioning helps in suppressing unwanted features and characteristics, ensuring that the animation remains focused and coherent.

latents

The latents output parameter is the batch of latents after they have been conditioned by the positive and negative prompts. These latents are now ready for further processing or rendering, having been transformed based on the specified prompt schedule.

Batch Prompt Schedule (Latent Input) 📅🅕🅝 Usage Tips:

• Ensure that your settings parameter is well-configured with clear and concise prompts to achieve the best results.
• Use high-quality latents as input to maximize the effectiveness of the conditioning process.
• Experiment with different clip models to find the one that best suits your animation needs.

Batch Prompt Schedule (Latent Input) 📅🅕🅝 Common Errors and Solutions:

"Invalid settings configuration"

• Explanation: This error occurs when the settings parameter is not properly configured or contains invalid values.
• Solution: Double-check the settings parameter to ensure all required fields are correctly filled and valid.

"Clip model not found"

• Explanation: This error indicates that the specified clip model could not be located or loaded.
• Solution: Verify that the clip model path is correct and that the model is accessible.

"Latents shape mismatch"

• Explanation: This error happens when the input latents have incompatible shapes or dimensions.
• Solution: Ensure that all input latents have consistent shapes and dimensions before passing them to the node.