Meet DiffusionBear 🐻 Free, open-source macOS studio for local AI image generation on Apple Silicon.
⚡ Native MLX (4-bit)
🎨 FLUX.2, Krea 2, Z-Image, Qwen & SDXL
🧠 13B models in 16GB RAM
🔒 100% offline & MIT
https://t.co/lRtlh7N2cd
https://t.co/8JEIzXrxoj
I'm sure you'll love it !
First integration test on next DiffusionBear release, show amazing rendering and generation time on 16gb M1, Qwen-Image-2.1 and @linoy_tsaban 2 steps LoRA ( 768x512 : first image 80sec, following ones : 50sec )
✨v1.5 of my 2-step Qwen Image 2.1 distill LoRA is now on @huggingface ✨
> improved prompt adherence and edit fidelity
> Realtime space updated to use v1.5
weights: https://t.co/5OZNpGGRB9
space ��️https://t.co/lfDCIlQnYv
editing product photos on a mac without comfyui? @OuincheWinch's diffusionbear 0.4.2 sequences qwen3-vl and fibo 2.0 edit turbo in mlx so resident vram stays under 6.8gb on 16gb macs (mit code; ad-hoc signed, non-commercial flux weights). https://t.co/nZwBpSBHzi
What a week !
On the last 7 days, we added to DiffusionBear :
- MOSS Image FST 5B
- FLUX.3 Kein 5.6B
- FIBO 2.0 Edit Turbo
Fibo is only 4 steps, 4bit quantized, and support local editing, up to 4 images reference https://t.co/IRfZylBDI3 and had pretty good writting skills
@ai_hakase_ Hey, Thank you !
We are now in 0.4.2, in the last week we add :
- MOSS Image FST 5B
- FLUX.3 Kein 5.6B
- FIBO 2.0 Edit Turbo
Fibo is only 4 steps, 4bit quantized, and support local editing, up to 4 images reference https://t.co/Nv4wXPEVhX
and had pretty good writting skills
@aiwithadb DiffusionBear
A free macOS studio that runs SDXL, FLUX.1, FLUX.2 Klein 4B/9B, FLUX.3 Klein 5.6B, MOSS-Image SFT 5B, Qwen-Image-2.1 & Turbo, Krea 2 Turbo, Z-Image-Turbo, and Bria FIBO 2.0 Edit Turbo on your Mac's own Metal GPU.
https://t.co/y2ENDnyWw4
@bria_ai_ https://t.co/y2ENDnyWw4
Never seen a such writing capable model with so few steps ! Amazing works @bria_ai_
16gb deserve to run it too :)
4bit excellent after all : https://t.co/IRfZylBDI3
@bria_ai_ FIBO 2 Edit Turbo is here in DiffusionBear! 🐻
Reserving such a great model for 32GB+ Macs was a shame… Now running in MLX 4-bit, it’s 100% local and lightning-fast on Apple Silicon Macs with just 16GB RAM!
Get it now on GitHub 👇
#AppleSilicon#Mac#DiffusionBear
A free macOS studio that runs SDXL (and its derivatives), FLUX.1, FLUX.2 Klein 4B/9B, FLUX.3 Klein 5.6B, MOSS-Image SFT 5B, Qwen-Image-2.1 & Turbo, Krea 2 Turbo, and Z-Image-Turbo on your Mac's own Metal GPU.
https://t.co/IFKwtCYgxk