Skip to content
latestai.eulatestai.eu
KR

Kromoss

Try itHugging Face

Community full fine-tune of MOSS-Image-Base at 1024 px, with optional 8-step turbo LoRAs; open weights under Apache-2.0.

Available in EUImage GenerationModel

Updated EU status checked Last shown

Kromoss at a glance

What is Kromoss?
Kromoss is a model. Community full fine-tune of MOSS-Image-Base at 1024 px, with optional 8-step turbo LoRAs; open weights under Apache-2.0.
Does Kromoss work in the EU?
Available in the EU. Open weights under Apache-2.0 on Hugging Face, downloadable and runnable locally with no regional restriction in the license. Last verified 14 Nov 2026.
How much does Kromoss cost?
Open source. Free to download; runs on local hardware. Licence: Apache-2.0.
Where was Kromoss shown?
In 1 creator video by AI Search. Most recently by AI Search on 20 Oct 2026, at 15:18 in “New top local AI image generator is here!”.

Sources: the vendor's own pages (linked above) and creator videos at the linked timestamps. EU status is information, not legal advice; see how we check.

About

Kromoss is a text-to-image model published on Hugging Face by the user "lodestones". It is a full fine-tune of OpenMOSS-Team/MOSS-Image-Base at 1024 px, trained on about 704k images (the card lists checkpoint update 11000, EMA weights). It keeps the MOSS-Image architecture, inference pipeline, Qwen3.5-9B text encoder and FLUX.2 small-decoder VAE, so it works as a drop-in replacement for MOSS-Image-SFT. A YouTube reviewer described it as looking better and more consistent than the base model.

The repo includes diffusers weights and single-file ComfyUI checkpoints (bf16 and an int8 ConvRot W8A8 version of about 5.4 GB). It also ships two 8-step distilled LoRAs (DMD and DMD2) that allow generation at 8 steps with CFG 1 instead of the default 50 steps at CFG 3. Conversion and quantization scripts are provided for turning other MOSS-Image fine-tunes into the ComfyUI format.

It is run locally, either through diffusers (needs diffusers>=0.38, transformers>=5.3 and trust_remote_code) or through ComfyUI, which has native MOSS-Image support. The card documents English prompting only and the vendor or organisation behind it is an individual community author.

Key features
  • Full fine-tune of MOSS-Image-Base at 1024 px
  • Drop-in replacement for MOSS-Image-SFT in diffusers and ComfyUI workflows
  • Two 8-step turbo LoRAs (DMD, DMD2) for CFG 1 generation
  • bf16 and int8 ConvRot (5.4 GB) ComfyUI checkpoints
  • Includes scripts to convert and quantize MOSS-Image fine-tunes
  • Apache-2.0 licensed open weights
1 moment

As seen on YouTube

Watch the moment: the exact second top creators show it. Videos load only when you press play, via youtube-nocookie.com.

Keep exploring

Similar AI tools

Closest neighbours by meaning, not just by category.

MI

MOSS Image 8-step DMD LoRA

Model

1

8-step distillation LoRA for MOSS-Image-SFT that runs without CFG, for ComfyUI.

Available in EUImage Generation
16 Oct 2026Open source
MO

MOSS-Image

Model · OpenMOSS

3

Open-weights 5B text-to-image model family from OpenMOSS with native JSON bounding-box layout control and text rendering.

Available in EUImage Generation
16 Oct 2026Open source
K2

Krea 2 LoRAs

App · Krea AI

1

LoRA training on Krea's platform to teach Krea 2 a custom style, character, face or product from a few images.

Available in EUImage Generation
Release date n/aFreemium
MF

Mage Flow

Model · Microsoft

1

Microsoft's open-weights 4B model for image generation and editing, with a 4-step turbo variant.

Available in EUImage Generation
Release date n/aOpen source
F2

Fibo 2.0

Model · Bria

1

Bria's open-weights image generation and natural-language editing model with multi-reference support, non-commercial license.

Available in EUImage Generation
18 Oct 2026Open source
F1

Fibo 1.5

Model · Bria AI

1

Open-weights text-to-image model from Bria AI, fine-tuned from FIBO, supporting structured JSON prompts and image editing.

Available in EUImage Generation
13 Jul 2026Open source