# Hunyuan LoRA Loader Nodes This collection provides four additional nodes for loading and managing Hunyuan Video LoRAs in ComfyUI: ## HunyuanLoadAndEditLoraBlocks Interactive node for editing and applying individual LoRA block settings. At the moment, it allows you to specifically enable/disable different double blocks when loading the lora. **Inputs:** - model: The base model to apply the LoRA to - lora_name: Select from available LoRAs - strength: LoRA strength (default: 1.0, range: -10.0 to 10.0) - save_settings: Whether to save block settings to cache (default: true) - use_single_blocks: Enable/disable single block layers (default: false) - double_blocks.0-19: Individual toggles for each double block ## HunyuanLoadFromBlockCache Loads a LoRA using previously cached block settings. **Inputs:** - model: The base model to apply the LoRA to - lora_name: Select from available LoRAs - strength: LoRA strength (default: 1.0, range: -10.0 to 10.0) - use_single_blocks: Enable/disable single block layers (default: false) ## HunyuanLoraFromJson Applies multiple LoRAs from a JSON configuration. **Inputs:** - model: The base model to apply the LoRA to - json_data: JSON configuration string **Example:** ```json [ { "filename": "kerbal30.safetensors", "strength": 0.75, "use_block_cache": true, "use_single_blocks": false }, { "filename": "other_lora.safetensors", "strength": 0.5, "use_block_cache": false, "use_single_blocks": true } ] ``` ## HunyuanLoraFromPrompt Automatically applies LoRAs based on trigger words in the prompt. **Inputs:** - model: The base model to apply the LoRA to - prompt: The input prompt text - config_file: YAML config file name (default: "lora_triggers.yaml") **Example Prompt:** **Example Config File (lora_triggers.yaml):** ```yaml loras: - filename: kerbal30.safetensors strength: 0.75 use_block_cache: true use_single_blocks: false keywords: - kerbal - astronaut - spacesuit ``` ## Cache Directory Block settings are stored in the cache directory at: ``` ComfyUI_PixelFun/cache/ ``` Each LoRA's block settings are saved as `[lora_name]_blocks.yaml` in this directory.