Nodes/Mecha Merge Node Pack/SD1-SUPERMERGER_BLOCKS Components Params
ComfyUI Node

SD1-SUPERMERGER_BLOCKS Components Params

One slider for all of SD1's blocks — the uniform block-weight shortcut

By ljleb·Created 3 years ago·Updated about a month ago· 101
SD1-SUPERMERGER_BLOCKS Components Params
    • recipe
    blocks0.00

    SD1-SUPERMERGER_BLOCKS Components Params (class SD1-SUPERMERGER_BLOCKS Defaults Hyper) is the minimal member of the block-weight family: a single blocks float that applies the same weight to all of SD1's merge blocks at once. One input, one output - it's the "uniform block weight" node, and its job is to be the boring, predictable option next to the 26-slider SD1-LDM Blocks node.

    Why does this exist when the 26-block node does strictly more? Because the sd1-supermerger_blocks config is its own merge space, and some recipes and merge methods are simply in block space. If you're building a recipe that already lives in the SuperMerger blocks layout, or you want a single scalar that scales every block together without leaving block space, this node is the clean way to express "same weight for all blocks" as a param recipe - no conversion step, no 26 widgets to babysit.

    How it works

    The node builds a scalar recipe in the sd1-supermerger_blocks config, so its one float applies uniformly to BASE, IN00IN11, M00, and OUT00OUT11. Emitted as a MECHA_RECIPE in param space, you can wire it into any merge method's weight/alpha input. Because it stays in block space, it's also handy as a building block (sorry) for manual recipe assembly and for testing whether a uniform block tweak helps before you commit to per-block tuning.

    The inputs that matter:

    • blocks - the uniform weight for all 26 SD1 blocks. Default 0 ("no override").
    • recipe (output, param space) - the block-space param.

    Honestly, that's the whole node. If your workflow already uses the 26-block node you probably never need this one - but the family of components params nodes exists precisely so every config's params are first-class, and this is the SuperMerger-blocks config's param node. When a recipe or conversion expects block-space input, reaching for the matching Components Params node is the reliable move.

    Install

    It ships in the Mecha Merge Node Pack (ljleb/comfy-mecha):

    • ComfyUI Manager → Install Custom Nodes → search "mecha" → Mecha Merge Node Pack, or manually:
    cd ComfyUI/custom_nodes
    git clone https://github.com/ljleb/comfy-mecha.git
    pip install -r comfy-mecha/requirements.txt
    

    Restart ComfyUI; it's under mecha. The pack's only dependency is sd-mecha==1.1.7; no model downloads.

    Common issues

    • It's uniform, not granular. If you came here looking for per-block control, the 26-slider SD1-LDM Blocks node is the one you actually want. This node is the "same value everywhere" convenience, not a replacement.
    • Match the config. The output is in sd1-supermerger_blocks space; if your downstream recipe is in LDM space, use the LDM blocks node (or a converter) instead of forcing this one in.
    • 0 means "no override." As with every components params node, a zero doesn't strip blocks - it leaves them at the method's default treatment.

    Every config in sd-mecha gets a Components Params node, and this is SD1's block config getting its fair share. It's not the star of the block-weight show - that's the 26-block node - but when you need a uniform block-space scalar, it's the right tool and it takes exactly one second to wire.

    Categorymecha

    Inputs (1)

    NameTypeDefaultDescription
    blocksFLOAT0.00-18446744073709550000–18446744073709550000

    Outputs (1)

    NameTypeDescription
    recipeMECHA_RECIPE