Skip to content
Images

RMSharmonize

Object harmonization LoRA for FLUX.2 Klein 9B

2 min readupdated

RMSharmonize is a dedicated adapter (LoRA) for FLUX.2 Klein 9B designed to blend cut-out or composited objects seamlessly into a unified scene. It automatically matches an object's lighting, shadows, colors, and reflections to the surrounding environment. This tool is built for designers and content creators who put together complex collages or composite images in ComfyUI and want realistic cohesion without hours of manual retouching.

Key features

  • Uses the trigger word rmsharmonize to activate the harmonization process.
  • Runs on FLUX.2 Klein 9B at a standard LoRA strength of 1.0.
  • Adjusts shadows, color grading, and reflections to match the existing scene conditions.
  • Supports local edits with masks for targeted compositing.

What you can do

  • Seamlessly blend an external object or cut-out subject into a real photo or illustration.
  • Generate accurate cast shadows and environmental reflections for added elements.
  • Build clean composites using compatible image-editing workflows in ComfyUI.

Example

Sample prompt for integrating an object into a scene: rmsharmonize, Organically integrate the soccer ball into the scene. Preserve the existing composition, object placement, and surroundings.

Read moreDetails
Benchmarks: designed for 2 generation steps at a LoRA strength of 1.0
Limitations: may alter fine details and object lighting, but does not guarantee preserving exact logos, text, or original object geometry
Training details: trained on FLUX.2 Klein Base 9B using Ostris AI Toolkit
Examples
Discuss Telegram 0

Will it run?

Runs in BF16

Fits without quantisation.

  • BF16weights 17 GB · ready files20 GB
  • FP8 / INT8recommendedweights 8.8 GB11 GB
  • 4-bitweights 5.1 GB7.1 GB

Tip: go 8-bit.FP8 / INT8 is within about a percent of full precision in quality, takes half the memory and runs faster on RTX 40/50 and H100. Full precision only pays off for fine-tuning.

What to get for this model

Massed ComputeOpen Massed Compute

Your own GPU: FP8 / INT8 — from RTX 5070 · BF16 — from RTX 4090

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

This is a LoRA: counted together with its base model FLUX.2-klein-base-9B.

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.

    # rmstools/RMSharmonize-FLUX.2-klein-9B-harmonization-lora
    # ===== Tokens: paste yours between the quotes =====
    export HF_TOKEN=""        # huggingface.co/settings/tokens, for gated models
    # ==========================================
    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 black-forest-labs/FLUX.2-klein-base-9B --local-dir ~/models/FLUX.2-klein-base-9B
    hf download rmstools/RMSharmonize-FLUX.2-klein-9B-harmonization-lora --local-dir ~/models/RMSharmonize-FLUX.2-klein-9B-harmonization-lora
    # The model is in ~/models. How to run it is in the model card (link above).
    Add more LoRAs to FLUX.2-klein-base-9B
  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

Requires the base model

black-forest-labs/FLUX.2-klein-base-9B

This model is a LoRA / adapter. Download the base model first, then load this one on top.

New posts about versions of FLUX.2-klein-base-9B (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

rmstools/RMSharmonize-FLUX.2-klein-9B-harmonization-lora
Author
rmstools
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