Nodes/ComfyUI-FunPack/FunPack Story Writer
ComfyUI Node

FunPack Story Writer

A ComfyUI node in FunPack with 24 inputs and 5 outputs.

By digital-garbage·Created about a year ago·Updated 24 days ago· 16
FunPack Story Writer
    • prompt1
    • prompt2
    • prompt3
    • prompt4
    • prompt5
    user_promptA photo of a [subject] in a [setting]. [action].
    prompt1
    prompt2
    prompt3
    prompt4
    prompt5
    story_system_prompt
    sequence_system_prompt
    model_path_type
    model_pathmlabonne/NeuralLlama-3-8B-Instruct-abliterated
    llm_safetensors_file
    prompt_count3
    top_p0.75
    top_k40
    min_p0.10
    temperature0.60
    max_new_tokens512
    repetition_penalty1.00
    mode
    vision_inputPut outputs of your VL model here to make the Story Writer aware of the starting image.
    sanity_checktrue
    sanity_check_system_promptAnalyze the given sequence and perform a correction, if the sequence does not match the given requirements: 1. The sequence is related to given user's prompt. 2. The sequence contains only physically possible actions. 3. The sequence contains information about characters, their appearances, positioning, actions, camera angle, focus and zoom. 4. The sequence is fully describing the requested action. Output ONLY corrected sequence, or return it unchanged if it matches the requirements. No additional text except for sequence is allowed.
    disable_continuityfalse
    provide_current_idtrue
    CategoryFunPack

    Inputs (24)

    NameTypeDefaultDescription
    user_promptSTRINGA photo of a [subject] in a [setting]. [action].
    prompt1STRING
    prompt2STRING
    prompt3STRING
    prompt4STRING
    prompt5STRING
    story_system_promptSTRING
    sequence_system_promptSTRING
    model_path_typeCOMBO2 options: Local Safetensors, HuggingFace Pretrained
    model_pathSTRINGmlabonne/NeuralLlama-3-8B-Instruct-abliterated
    llm_safetensors_fileCOMBO0 options:
    prompt_countINT31–5
    top_pFLOAT0.750–2
    top_kINT400–1000
    min_pFLOAT0.100–1
    temperatureFLOAT0.600–2
    max_new_tokensINT51264–4096
    repetition_penaltyFLOAT1.000–3
    modeCOMBO2 options: Sequences from story, Sequences from user prompt
    vision_inputSTRINGPut outputs of your VL model here to make the Story Writer aware of the starting image.
    sanity_checkBOOLEANtrue
    sanity_check_system_promptSTRINGAnalyze the given sequence and perform a correction, if the sequence does not match the given requirements: 1. The sequence is related to given user's prompt. 2. The sequence contains only physically possible actions. 3. The sequence contains information about characters, their appearances, positioning, actions, camera angle, focus and zoom. 4. The sequence is fully describing the requested action. Output ONLY corrected sequence, or return it unchanged if it matches the requirements. No additional text except for sequence is allowed.
    disable_continuityBOOLEANfalse
    provide_current_idBOOLEANtrue

    Outputs (5)

    NameTypeDescription
    prompt1STRING
    prompt2STRING
    prompt3STRING
    prompt4STRING
    prompt5STRING