Nodes/ComfyUI Ino Nodes/Ino Handle Load Model
ComfyUI Node

Ino Handle Load Model

The honest version is 'check the type-specific loaders instead'

By nobandegani·Created about a year ago·Updated 2 months ago· 6
Ino Handle Load Model
    • success
    • message
    • model
    enabledtrue
    model_type
    model_path
    unet_weight_dtype
    clip_type

    Sometimes the most useful thing an article can tell you about a node is that it isn't done yet. Ino Handle Load Model is the Ino Nodes pack's generic "load anything by type" loader, and as of this writing it's a stub: the author's own description says "currently not fully implemented," and the code confirms it - run it and you'll get a message saying loading that model type isn't supported yet. Know that going in and you'll save yourself a confusing afternoon.

    It's part of the Ino Nodes pack from nobandegani. The idea is right: one node with a model_type dropdown that loads a model by path and returns the loaded object, whatever type it is. The execution, right now, isn't there.

    What it exposes

    The inputs describe the ambition:

    • model_type - dropdown of model categories (diffusion model, text encoder, VAE, LoRA, ControlNet...).
    • model_path - the path to the model file.
    • unet_weight_dtype, clip_type - optional, for the type-specific loading behaviors.

    Outputs are success, message, and model (any type). And here's the thing: for every model type you pick, the current implementation returns False with a "not supported yet" message. The actual loading in this pack is done by the type-specific nodes - Ino Load VAE Model, Ino Load Clip Model, Ino Load Controlnet Model, Ino Load Diffusion Model, and the LoRA loaders - each of which wraps ComfyUI's own loaders and returns the correctly-typed object.

    What you should actually use

    If you're building an Ino workflow and need to load a model, skip this node and reach for the specific loader for your type. The one place this node participates in real flows is behind the scenes: Ino Handle Download And Load Model calls it after a download, which means that node inherits the same "not supported yet" limitation for unusual model types. For the mainstream types (VAE, CLIP, diffusion, ControlNet, LoRA) the pack has working dedicated loaders - use those.

    This node is worth knowing about for two reasons: so you don't waste time wiring it into a workflow expecting it to work, and so you know what to watch for. The pack updates fast (version 2.0.5, active development), so "not supported yet" may become "supported" in a future release - re-test it when the pack bumps.

    Installing it

    It ships with the whole Ino Nodes pack. ComfyUI Manager: search "ComfyUI Ino Nodes", install, restart. Or manually:

    cd ComfyUI/custom_nodes
    git clone https://github.com/nobandegani/ComfyUI-InoNodes.git
    cd comfyui_ino_nodes
    pip install -r requirements.txt
    

    Restart after. Pack-wide caveats apply: it's built on ComfyUI's V3 schema (README wants v0.18.1+), and requirements.txt installs inopyutils, the helper library every node imports.

    Common issues

    The only "issue" worth flagging is the known one: this node reports "not supported" for every model type - that's the current state of the implementation, not a misconfiguration on your part. Don't chase it. If success comes back false with a "not supported" message, switch to the type-specific loader (or wait for a pack update). Search "Ino Handle Load Model" in the node menu (grouped under InoModelHelper) to see it - and its working siblings - for yourself.

    CategoryInoModelHelper

    Inputs (5)

    NameTypeDefaultDescription
    enabledBOOLEANtrue
    model_typeCOMBO32 options: audio_encoders, checkpoints, clip, clip_vision, controlnet, detection, +26
    model_pathSTRING
    unet_weight_dtypeoptCOMBO4 options: default, fp8_e4m3fn, fp8_e4m3fn_fast, fp8_e5m2
    clip_typeoptCOMBO25 options: stable_diffusion, stable_cascade, sd3, stable_audio, mochi, ltxv, +19

    Outputs (3)

    NameTypeDescription
    successBOOLEAN
    messageSTRING
    model*