ComfyUI Node

Qwen_TE_LLM

A ComfyUI node in ListHelper/LLM with 11 inputs and 2 outputs.

By dseditor·Created about a year ago·Updated 6 months ago· 71
Qwen_TE_LLM
    • text
    • used_seed
    user_promptA girl in a coffee shop
    prompt_templateCustom
    system_prompt
    max_new_tokens2048
    temperature0.7
    seed0
    keep_model_loadedfalse
    use_flash_attentionfalse
    use_quantizationfalse
    top_p0.90
    top_k50
    CategoryListHelper/LLM

    Inputs (11)

    NameTypeDefaultDescription
    user_promptSTRINGA girl in a coffee shop
    prompt_templateCOMBOCustom8 options: Custom, extract_person_features.md, image_to_prompt.md, photography_en.md, photography_zh.md, qwen2512_en.md, +2
    system_promptSTRING
    max_new_tokensINT20481–4096
    temperatureFLOAT0.70–2
    seedoptINT00–18446744073709550000
    keep_model_loadedoptBOOLEANfalseWARNING: Set to False to unload model after generation. Required for low VRAM workflows.
    use_flash_attentionoptBOOLEANfalseEnable FlashAttention-2 (may not improve speed for small batch inference)
    use_quantizationoptBOOLEANfalseEnable INT8 quantization for lower memory usage (slower inference)
    top_poptFLOAT0.900–1
    top_koptINT500–100

    Outputs (2)

    NameTypeDescription
    textSTRING
    used_seedINT