Skip to content
Images

Marigold V2 albedo (ComfyUI)

extracting shadow-free surface color

2 min readupdated

Ready-to-use weights for estimating albedo (the base color of a surface without lighting, specular highlights, or shadows) from photos. This release combines the Marigold V2 model from researchers at EPFL, HUAWEI Bayer Lab, and the University of Bologna with the Qwen-Image-Edit-2509 base architecture. Unlike the original release, the base transformer here is quantized to 4-bit NF4 (weight compression to reduce VRAM usage), and all components are organized into standard ComfyUI folders. It is useful for 3D artists, photogrammetry specialists, and anyone working on scene relighting in ComfyUI or the Askiagraphy app.

Key features

  • Compressed NF4 backbone: the Qwen-Image-Edit-2509 transformer is quantized to 4-bit via bitsandbytes (blocksize 64), while output projections and non-linear layers are kept in bf16 to preserve accuracy.
  • ComfyUI-ready folder structure: files are pre-sorted into the diffusion_models, loras, vae, and embeddings directories.
  • Comfy-Org compatibility: the LoRA adapter (a lightweight add-on for the base model), fine-tuned VAE decoder, and text embedding are byte-for-byte identical to the official Comfy-Org release.
  • Permissive license: released under the Apache License 2.0, which allows commercial use.

What you can do

  • Extract clean textures (albedo maps) from regular photos by stripping away cast shadows and light reflections.
  • Use the resulting base color maps to relight objects or build materials for 3D scenes.
  • Run the albedo estimation pipeline locally in ComfyUI on GPUs with limited VRAM thanks to the 4-bit transformer.

Example

To get started in ComfyUI, download the files from the repository and place them into their matching model directories:

models/diffusion_models/qwen_image_edit_2509_nf4.safetensors
models/loras/marigold_v2_albedo.safetensors
models/vae/marigold_v2_albedo_vae.safetensors
models/embeddings/marigold_v2_albedo_conditioning.safetensors
Read moreDetails
Base models: huawei-bayerlab/marigold-v2-0 and Qwen/Qwen-Image-Edit-2509
Quantization: bitsandbytes 0.48.2 (NF4, blocksize 64, without double quantization)
License: Apache License 2.0
Discuss Telegram 0

Will it run?

Runs in NF4

Fits without quantisation.

  • NF4weights 13 GB · ready files15 GB

What to get for this model

Massed ComputeOpen Massed Compute

Your own GPU: NF4 — from RTX 5080

Massed Compute hourly prices, November 2026. The discounted price applies with the coupon at checkout.

Estimate: weights plus working memory. Long videos, high resolutions and long contexts need more.

Run it on Massed Compute

The script downloads the model to the rented machine. How to run it is in the model card.

  1. 1. Rent a machine with a fitting GPU

    Open Massed Compute
  2. 2. Paste the script into the machine's terminal

    Connect over SSH (the Massed Compute panel shows the command) or open a terminal on the machine's desktop.

    HF_TOKEN / CIVITAI_TOKENThe script starts with token lines: paste yours between the quotes. The Hugging Face token is for gated models, the Civitai one for Civitai models.

    # askiagraphy/marigold-v2-albedo-comfyui
    mkdir -p ~/models
    curl -LsSf https://astral.sh/uv/install.sh | sh
    source "$HOME/.local/bin/env"
    uv tool install "huggingface_hub[cli]"
    hf download askiagraphy/marigold-v2-albedo-comfyui --local-dir ~/models/marigold-v2-albedo-comfyui
    # The model is in ~/models. How to run it is in the model card (link above).
  3. 3. Use it

    The model is in ~/models. How to run it is in the model card (link above).

Stop the machine in the panel when you are done: it bills by the minute while it runs.

How to speed it up

Picked for this model type. Choose your hardware above to see what fits you exactly.

  • SageAttention 3v2.2.0from October 28, 2025

    even faster on RTX 50 / Blackwell

    FP4 attention for Blackwell. Faster than SageAttention 2, but the authors suggest v2 when accuracy matters.

    How to enable: Build from the sageattention3_blackwell folder of the repo; in ComfyUI via KJNodes.

    Earlier generations (2)
    • SageAttention 2++ — Quantised (INT8/FP8) attention with no visible quality loss. The main speed-up for image and video generation.
    • SageAttention 1 — First version: INT8 for Q·K. Still works, but 2++ is faster and more accurate.
  • FlashAttention 44.0.0b40from November 18, 2026new release

    fastest exact attention on B200

    Kernels for Blackwell (beta). Exact attention without quantisation — fine for everything, LLMs included.

    How to enable: pip install flash-attn-4; vLLM and SGLang pick it up themselves.

    Earlier generations (3)
    • FlashAttention 3 — The Hopper (H100/H200) version with FP8.
    • FlashAttention 2 — The standard for RTX 30/40/50 and A100: exact attention without the memory cost of long contexts.
    • xFormers — The older way to speed up attention, works almost everywhere. Sage or Flash are faster where available.
  • cache-dit (DBCache)v1.5.6from November 20, 2026new release

    1.5–3× at nearly the same quality

    Reuses block outputs between diffusion steps. Supports most DiT models.

    How to enable: pip install cache-dit; one line cache_dit.enable_cache(pipe) for diffusers.

    Earlier generations (2)
    • MagCache — Skips steps by magnitude of change. More accurate than TeaCache, no calibration.
    • TeaCache — The first popular diffusion step cache. Works, newer methods are more accurate.
  • Nunchaku (SVDQuant 4-bit)v1.2.1from January 25, 2026

    ≈3× faster and 3.5× less memory

    4-bit versions of popular image models with almost no quality loss (FP4 on RTX 50).

    How to enable: Ready weights for supported models; in ComfyUI use ComfyUI-nunchaku.

  • torch.compile2.15.0from October 28, 2026new release

    10–30% for free

    Compiles the model for your GPU. Slower first run, faster afterwards.

    How to enable: model = torch.compile(model); in ComfyUI the TorchCompileModel node.

Model family

  1. Qwen-Image-Edit-2509family root
  2. marigold-v2-0
  3. marigold-v2-albedo-comfyui

New posts about versions of Qwen-Image-Edit-2509 (quants, finetunes, LoRAs) will show up under the bell.

Quantised versions 12All on Hugging Face
Fine-tunes 12All on Hugging Face
LoRAs and adapters 12All on Hugging Face

About the model

askiagraphy/marigold-v2-albedo-comfyui
Author
askiagraphy
Category
Images
Released
Posted

On NeyroLearn

Hugging Face

00

Ask AI about this model

Answered by Gemini 4 · the chat is kept only in this tab

Comments