ComfyUI Node: Nunchaku FLUX DiT Loader

Authored by nunchaku-tech

Created

Updated

2340 stars

Category

Nunchaku

Inputs

model_path
    cache_threshold FLOAT
    attention
    • nunchaku-fp16
    • flash-attention2
    cpu_offload
    • auto
    • enable
    • disable
    device_id INT
    data_type
    • bfloat16
    • float16
    i2f_mode
    • enabled
    • always

    Outputs

    MODEL

    Extension: ComfyUI-nunchaku

    Nunchaku ComfyUI Node. Nunchaku is the inference that supports SVDQuant. SVDQuant is a new post-training training quantization paradigm for diffusion models, which quantize both the weights and activations of FLUX.1 to 4 bits, achieving 3.5× memory and 8.7× latency reduction on a 16GB laptop 4090 GPU. See more details: https://github.com/mit-han-lab/nunchaku

    Authored by nunchaku-tech

    View Nodes

    Run ComfyUI workflows in the Cloud!

    No downloads or installs are required. Pay only for active GPU usage, not idle time. No complex setups and dependency issues

    Learn more