ComfyUI Node: SDXL Analyzer + Selective Loader V2

Authored by shootthesound

Created

Updated

538 stars

Run ComfyUI workflows without the setup

No installs, no CUDA version roulette, no GPU sitting idle on your bill. Bring a workflow and run it in the browser.

Category

loaders/lora

Inputs

model MODEL
positive CONDITIONING
negative CONDITIONING
lora_name
    strength FLOAT
    preset
    • Default
    • All Off
    • Half Strength
    • UNet Only
    • High Impact
    • Text Encoders Only
    • Decoders Only
    • Encoders Only
    • Style Focus
    • Composition Focus
    • Face Focus
    • Custom
    text_encoder_1 BOOLEAN
    text_encoder_1_str FLOAT
    text_encoder_2 BOOLEAN
    text_encoder_2_str FLOAT
    input_4 BOOLEAN
    input_4_str FLOAT
    input_5 BOOLEAN
    input_5_str FLOAT
    input_7 BOOLEAN
    input_7_str FLOAT
    input_8 BOOLEAN
    input_8_str FLOAT
    unet_mid BOOLEAN
    unet_mid_str FLOAT
    output_0 BOOLEAN
    output_0_str FLOAT
    output_1 BOOLEAN
    output_1_str FLOAT
    output_2 BOOLEAN
    output_2_str FLOAT
    output_3 BOOLEAN
    output_3_str FLOAT
    output_4 BOOLEAN
    output_4_str FLOAT
    output_5 BOOLEAN
    output_5_str FLOAT
    other_weights BOOLEAN
    other_weights_str FLOAT
    lora_path_opt STRING
    schedule_preset
    • Custom
    • Constant 1.0 (No Change)
    • Linear In (0→1)
    • Linear Out (1→0)
    • Ease In (slow start)
    • Ease Out (slow end)
    • Ease In-Out
    • Bell Curve (peak middle)
    • Wide Bell
    • Sharp Bell
    • High Start → Cut Low
    • High Start → Cut Off
    • High Early → Fade
    • Low Start → Ramp Late
    • Off → Kick In Late
    • Low → Boost End
    • Step Up Mid
    • Step Down Mid
    • Two Steps Up
    • Pulse Early
    • Pulse Mid
    • Pulse Late
    • Constant Half
    • Constant Low
    • INV: Linear In (1→0)
    • INV: Linear Out (0→1)
    • INV: Ease In
    • INV: Ease Out
    • INV: Ease In-Out
    • INV: Bell (dip middle)
    • INV: Wide Bell
    • INV: Sharp Bell
    • INV: Low Start → Boost High
    • INV: Low Start → Full On
    • INV: Low Early → Build
    • INV: High Start → Drop Late
    • INV: Full → Cut Off Late
    • INV: High → Drop End
    • INV: Step Down Mid
    • INV: Step Up Mid
    • INV: Two Steps Down
    • INV: Dip Early
    • INV: Dip Mid
    • INV: Dip Late
    strength_schedule STRING
    save_refined_lora BOOLEAN
    save_path STRING
    save_filename STRING

    Outputs

    MODEL

    CONDITIONING

    CONDITIONING

    STRING

    STRING

    Extension: Realtime LoRA Trainer

    Train, analyze, selectively load, and extract/merge LoRAs inside ComfyUI. Supports Z-Image, Qwen Image, Qwen Image Edit, SDXL, FLUX, FLUX Klein, Wan 2.2, and SD 1.5. Includes LoRA layer analyzers, selective block loaders, model layer editors, and text-encoder debiasing/inspection tools.

    Authored by shootthesound

    View Nodes

    Looking for a different node?

    More nodes in Realtime LoRA Trainer

    Other loaders/lora nodes

    Run ComfyUI workflows without the setup

    No installs, no CUDA version roulette, no GPU sitting idle on your bill. Bring a workflow and run it in the browser.

    Learn more