tinyConditioning

Documentation

• Class name: ttN conditioning
• Category: 🌏 tinyterra/base
• Output node: False

The node focuses on applying conditioning transformations to input data, leveraging advanced encoding techniques to enhance or modify the data based on specified conditions. It abstracts complex conditioning logic into a more accessible interface for users, aiming to streamline the process of data manipulation and enrichment.

Input types

Required

• model
  • Specifies the model to be used for conditioning, playing a central role in how the input data is processed and transformed.
  • Comfy dtype: MODEL
  • Python dtype: ModelPatcher | None
• clip
  • Defines the CLIP model parameters to be used in conjunction with the main model for conditioning, affecting the outcome of the transformation.
  • Comfy dtype: CLIP
  • Python dtype: CLIP | None
• loras
  • Specifies the LoRA configurations to be applied, influencing the conditioning process by adjusting the model's behavior.
  • Comfy dtype: STRING
  • Python dtype: str | None
• positive
  • The positive text input for conditioning, used to guide the model towards desired attributes or outcomes.
  • Comfy dtype: STRING
  • Python dtype: str
• positive_token_normalization
  • Normalization method for the positive text, affecting how the text is processed before conditioning.
  • Comfy dtype: COMBO[STRING]
  • Python dtype: str
• positive_weight_interpretation
  • Determines how the positive text's weight is interpreted during the conditioning process, influencing the model's adjustments.
  • Comfy dtype: COMBO[STRING]
  • Python dtype: str
• negative
  • The negative text input for conditioning, used to steer the model away from undesired attributes or outcomes.
  • Comfy dtype: STRING
  • Python dtype: str
• negative_token_normalization
  • Normalization method for the negative text, affecting how the text is processed before conditioning.
  • Comfy dtype: COMBO[STRING]
  • Python dtype: str
• negative_weight_interpretation
  • Determines how the negative text's weight is interpreted during the conditioning process, influencing the model's adjustments.
  • Comfy dtype: COMBO[STRING]
  • Python dtype: str

Optional

• optional_lora_stack
  • An optional stack of LoRA configurations to be applied, providing additional customization to the conditioning process.
  • Comfy dtype: LORA_STACK
  • Python dtype: List[Tuple[str, int, float]] | None
• prepend_positive
  • Text to prepend to the positive input, further customizing the conditioning process.
  • Comfy dtype: STRING
  • Python dtype: str | None
• prepend_negative
  • Text to prepend to the negative input, further customizing the conditioning process.
  • Comfy dtype: STRING
  • Python dtype: str | None

Output types

• model
  • Comfy dtype: MODEL
  • Returns the conditioned model after applying the specified transformations.
  • Python dtype: ModelPatcher | None
• positive
  • Comfy dtype: CONDITIONING
  • Outputs the positive conditioning effects applied to the model.
  • Python dtype: torch.Tensor
• negative
  • Comfy dtype: CONDITIONING
  • Outputs the negative conditioning effects applied to the model.
  • Python dtype: torch.Tensor
• clip
  • Comfy dtype: CLIP
  • Returns the CLIP model parameters used during the conditioning process.
  • Python dtype: CLIP | None
• pos_string
  • Comfy dtype: STRING
  • Outputs the final positive string after conditioning, reflecting the combined and processed input.
  • Python dtype: str
• neg_string
  • Comfy dtype: STRING
  • Outputs the final negative string after conditioning, reflecting the combined and processed input.
  • Python dtype: str

Usage tips

• Infra type: GPU
• Common nodes: unknown

Source code

class ttN_conditioning:
    version = '1.0.0'
    @classmethod
    def INPUT_TYPES(cls):
        return {"required": { 
                        "model": ("MODEL",),
                        "clip": ("CLIP",),

                        "loras": ("STRING", {"placeholder": "<lora:loraName:weight:optClipWeight>", "multiline": True}),

                        "positive": ("STRING", {"default": "Positive","multiline": True, "dynamicPrompts": True}),
                        "positive_token_normalization": (["none", "mean", "length", "length+mean"],),
                        "positive_weight_interpretation": (["comfy", "A1111", "compel", "comfy++", "down_weight"],),

                        "negative": ("STRING", {"default": "Negative", "multiline": True, "dynamicPrompts": True}),
                        "negative_token_normalization": (["none", "mean", "length", "length+mean"],),
                        "negative_weight_interpretation": (["comfy", "A1111", "compel", "comfy++", "down_weight"],),
                        },                
                "optional": {
                    "optional_lora_stack": ("LORA_STACK",),
                    "prepend_positive": ("STRING", {"default": None, "forceInput": True}),
                    "prepend_negative": ("STRING", {"default": None, "forceInput": True}),
                    },
                "hidden": {"ttNnodeVersion": ttN_conditioning.version, "my_unique_id": "UNIQUE_ID"},}

    RETURN_TYPES = ("MODEL", "CONDITIONING", "CONDITIONING", "CLIP", "STRING", "STRING")
    RETURN_NAMES = ("model", "positive", "negative", "clip", "pos_string", "neg_string")

    FUNCTION = "condition"
    CATEGORY = "🌏 tinyterra/base"

    def condition(self, model, clip, loras,
                       positive, positive_token_normalization, positive_weight_interpretation, 
                       negative, negative_token_normalization, negative_weight_interpretation, 
                       optional_lora_stack=None, prepend_positive=None, prepend_negative=None,
                       my_unique_id=None):

        if optional_lora_stack is not None:
            for lora in optional_lora_stack:
                model, clip = loader.load_lora(lora[0], model, clip, lora[1], lora[2])

        if loras not in [None, "None"]:
            model, clip = loader.load_lora_text(loras, model, clip)

        positive_embedding = loader.embedding_encode(positive, positive_token_normalization, positive_weight_interpretation, clip, title='ttN Conditioning Positive', my_unique_id=my_unique_id, prepend_text=prepend_positive)
        negative_embedding = loader.embedding_encode(negative, negative_token_normalization, negative_weight_interpretation, clip, title='ttN Conditioning Negative', my_unique_id=my_unique_id, prepend_text=prepend_negative)

        final_positive = (prepend_positive + ' ' if prepend_positive else '') + (positive + ' ' if positive else '')
        final_negative = (prepend_negative + ' ' if prepend_negative else '') + (negative + ' ' if negative else '')

        return (model, positive_embedding, negative_embedding, clip, final_positive, final_negative)