🌀 IAMCCS-nodes

<img src="icon.png" width="150" height="150">

Author: IAMCCS (Carmine Cristallo Scalzi)

Category: ComfyUI Custom Nodes

Main Feature: Fix for LoRA loading in native WANAnimate workflows

Version: 1.3.0

UPDATE VERSION 1-3-0

🆕 Version 1.3.0 — New MODEL IO LoRA Stack + Qwen Loader Instructions

Highlights:

• Added LoRA Stack (Model In→Out) WAN node: directly applies up to 4 WAN / Flow / Standard LoRAs to an incoming MODEL and outputs a patched MODEL (ideal for WAN 2.2 workflows where a single node step is preferred).
• Qwen Image LoRA (IAMCCS QwenImgLoadFix) – Updated for 1.3.0: a fixed Qwen Image LoRA loader with improved offload controls and UX.
• Extended internal WAN key remapping for seamless WAN 2.2 (Flow) + WAN 2.1 cross-compatibility.
• Documentation updated with explicit Qwen Image LoRA loader prerequisites for low VRAM users (nunchaku based).
• Version bump across project files.

New Node: LoRA Stack (Model In→Out) WAN

lora_stack_model_I_O.png

Use this node when you already have a base MODEL loaded (WAN 2.2, Flow, SDXL, etc.) and want a single pass application of multiple LoRAs without an intermediate stack/output hand-off. It mirrors the behavior of the classic stack + apply pair but merges them for simpler graphs (especially animation or chained sampler pipelines).

Inputs:

• model: base diffusion MODEL.
• lora1..lora4 + strength1..strength4 (skips if "no" or strength == 0.0)
• model_type: choose flow, wan2x, or standard to control remapping logic.
• Optional lora (LORA) input: allows concatenating a previously built stack from IAMCCS_WanLoRAStack for more than 4 total LoRAs.

Output:

• Patched MODEL ready for samplers / video pipelines.

Recommended Use (WAN 2.2 workflows):

1. Load base WAN 2.2 / LightX2V model.
2. Add LoRA Stack (Model In→Out) WAN and select up to 4 LoRAs.
3. (Optional) Chain a classic IAMCCS_WanLoRAStack into the optional lora input if you need >4.
4. Connect output to KSampler / Animate nodes.

Why this node: Eliminates one extra node hop, reduces graph complexity and clarifies model lineage in large animation workflows.

Previous Versions

Version 1.2.3 — New input lora - add another StackLoraModel (concatenate) + Extended Wan 2.1 Compatibility

Version 1.2.1

UPDATE VERSION 1-2-1

🆕 Version 1.2.1 — Extended Wan 2.1 Compatibility

The WAN-style remap node now supports LightX2V 2.1 LoRA models.
This version extends overall compatibility to all LoRA types — even those without dedicated weight tensors (these will simply display a non-critical “missing optional weights” message).

This ensures smoother cross-compatibility between LightX2V 2.1 / 2.2 and any WAN-based or character LoRA setup.

See full changelog → CHANGELOG.md

Overview

The IAMCCS-nodes package introduces a fix for a key limitation in native WANAnimate workflows: when users run animation pipelines without the WanVideoWrapper, LoRA models fail to load correctly — most weights are ignored, and the visual consistency breaks.

This package contains two complementary nodes that work together to fix this problem and restore full LoRA functionality while keeping the workflow lightweight and modular.

The IAMCCS Native LoRA System introduces an optimized way to handle multiple LoRAs inside native ComfyUI workflows. It is composed of two interconnected nodes designed to work seamlessly together.

1. LoRA Stack (WAN-style remap)

This node lets you combine several LoRA models—especially those made for WAN2.x / Animate / Flow architectures—into one unified output.

Each LoRA slot includes:

an independent strength control,

automatic WAN-style key remapping for full compatibility,

and support for .safetensors files across any model type (flow, wan, sdxl, etc).

It produces a stacked LoRA bundle that merges all the active LoRAs and prepares them for efficient native injection.

2. Apply LoRA to MODEL (Native)

This node applies the generated LoRA stack directly to a loaded diffusion model at the Torch level, without relying on older ComfyUI Apply LoRA wrappers.

Works natively with FP16 accumulation (recommended when paired with Model Patch Torch Settings)

Maintains precision and speed

Fully compatible with WAN2.x and other Flow-type diffusion models

Output: a ready-to-run patched model for image or video generation.

This structure replaces multiple chained LoRA nodes with a single modular system, improving both stability and performance. Ideal for WANAnimate, WANVideo, or any Flow-based cinematic model.

LoRA Concatenation (1.2.3)

lora_concatenatel.png

LoRA Stack (WAN-style remap)

Supports the following LoRAs:

• New Moe distill WAN 2.2 LightX2V High Model
• New Moe distill WAN 2.2 LightX2V Low Model
• WAN 2.2 LightX2V High Model
• WAN 2.2 LightX2V Low Model
• WAN Boost Realism
• WAN 2.2 LightX2V 4-Step High
• WAN 2.2 LightX2V 4-Step Low
• Character LoRAs
• WAN 2.1 LightX2V Model

Installation

The node has now been officially accepted on ComfyUI Manager, You can install it directly from there (just search for IAMCCS).

or

You can grab it manually:

digit in your terminal:

cd ComfyUI/custom_nodes git clone https://github.com/IAMCCS/IAMCCS-nodes.git

Compatibility

ComfyUI ≥ 0.3.0

Python ≥ 3.12

Torch ≥ 2.8 (CUDA 12.6 or 12.8)

Compatible with: WAN2.1, WAN2.2, WANAnimate, WANAnimate_relight, Pulid, Flux, and multi-LoRA setups.

Technical Insight

LoRA weights fail to load in native WANAnimate pipelines because the model initialization bypasses the internal LoRA merge functions used in the wrapper. By separating the process into two modular nodes, IAMCCS-nodes restores full LoRA compatibility without depending on WanVideoWrapper, keeping performance high and structure clean.

Node 1 replaces the missing LoRA-loading phase.

Node 2 reintroduces dynamic LoRA reapplication and blending inside the animation graph.

This modular architecture makes LoRA management in WANAnimate flexible, transparent, and fully native.

If my work helped you, and you’d like to say thanks — grab me a coffee ☕

<a href="https://www.buymeacoffee.com/iamccs" target="_blank"> <img src="https://cdn.buymeacoffee.com/buttons/v2/default-yellow.png" alt="Buy Me A Coffee" width="200" /> </a>

---

Qwen Image LoRA (IAMCCS QwenImgLoadFix) – Updated for 1.3.0

This repo ships a fixed Qwen Image LoRA loader with improved offload controls and UX.

LORAQWEN.png

Prerequisites (Low VRAM Friendly)

Required:

• ComfyUI (≥ 0.3.0)
• ComfyUI-nunchaku (Qwen Image / Sana transformer implementation)
• ComfyUI-QwenImageLoraLoader (wrappers.qwenimage module with ComfyQwenImageWrapper)
• Qwen Image model weights placed as per nunchaku loader instructions
• LoRA files (.safetensors) in ComfyUI/models/loras

Recommended for low VRAM systems:

• Enable offload_auto_tune (auto scales transformer block residency)
• Set offload_policy=disable only if you experience instability during composition (keeps everything on GPU while LoRAs active)
• Keep offload_num_blocks_on_gpu low (1–2) if VRAM < 12 GB

Node Overview

IAMCCS QwenImgLoadFix

• Wraps a NunchakuQwenImageTransformer2DModel / NunchakuSanaTransformer2DModel if not already wrapped.

• Adds LoRA via internal wrapper (model_wrapper.loras.append(...)) with strength preset.

• Supports composition modes:

  • append: legacy shape-changing approach
  • merge_v2: in-place delta merge (preferred; no rank expansion)

• Offload Controls: dynamic, rebuild or disable strategies; pin-memory toggle; auto-tune; VRAM margin.

• Device Safety: ensures params & buffers migrate to correct CUDA device before forward.

Dependency Checklist

| Component | Purpose | |-----------|---------| | ComfyUI-nunchaku | Provides transformer class loaded by base model node | | ComfyUI-QwenImageLoraLoader | Supplies wrappers/qwenimage.py used for wrapping | | IAMCCS-nodes | Adds fixed loader + WAN LoRA stack system | | LoRA safetensors | User-provided style/character adapters |

If wrappers/qwenimage.py is not found the node will raise an import error. Ensure ComfyUI-QwenImageLoraLoader repository resides in custom_nodes/.

Quick Start (Low VRAM Scenario)

1. Load Qwen Image model (nunchaku loader node).
2. Place IAMCCS QwenImgLoadFix directly after it.
3. Select lora_name & preset (start with 0.76–1.00 for natural balance).
4. Set composition_mode=merge_v2 unless you need legacy behavior.
5. Keep offload_auto_tune=ON; reduce offload_num_blocks_on_gpu if memory errors occur.
6. Run sampler; adjust preset or switch to a higher strength if effect too weak.

External Requirements Recap

This loader will not function standalone — both ComfyUI-nunchaku and ComfyUI-QwenImageLoraLoader must be installed, and the Qwen Image weights must be correctly placed. The IAMCCS node auto-wraps only if it detects a supported transformer class. If wrapping fails, verify repository folder names match expected conventions.