LaunchDevelopersOctober 17, 2026

MOSS-Image LoRA training added to AcademiaSD LoRAlab Trainer Studio

Read original source →reddit.com

AcademiaSD LoRAlab Trainer Studio now supports MOSS-Image, the new Apache 2.0 open text-to-image model, for both Base and SFT LoRA training under 6 GB VRAM. It becomes the 12th model family in the Studio, alongside Qwen-Image 2.1, FLUX.2 Klein 9B, FLUX.3 Image Klein, Krea 2, Z-Image, ERNIE-Image, Ideogram 4, Anima, SDXL, LTX-2.3 and MiniMax-H3.

1 source

More stories today

Open the live feed