LoRA Training for Fibo 2.0 Added
The LoRAlab Trainer studio is a web app that trains LoRAs for a lot of models (Qwen-Image, Krea, Z-Image, Anima, SDXL, LTX, MiniMax-H3 and so on) with an automatic captioner, live previews, step-accurate resume and one-click export to ComfyUI.
Proof1 community thread
Key points
- Fibo 2.0 Dev joins it, with normal LoRAs and edit LoRAs (before/after pairs).
- Pre-cache, training, previews, edit mode and resume are all verified on Fibo 2.0 Dev, but it is young and the defaults will probably move.
- And worth saying: Fibo 2.0 Dev is CC BY-NC 4.0, so non-commercial use only.
- While you are at it, the studio also got Ming-Image (transparent designs), Boogu-Image-Edit 0.1, ERNIE-Image, FLUX.3 Image Klein, Z-Image, Anima, MOSS-Image and Qwen-Image 2512 / Edit 2511.
Sources (1)
- [1]LoRA Training for Fibo 2.0 Addedr/StableDiffusion (top, day) · Oct 21, 02:27 PM
The LoRAlab Trainer studio is a web app that trains LoRAs for a lot of models (Qwen-Image, Krea, Z-Image, Anima, SDXL, LTX, MiniMax-H3 and so on) with an automatic captioner, live previews, step-accurate resume and one-click export to ComfyUI.
Fibo 2.0 Dev joins it, with normal LoRAs and edit LoRAs (before/after pairs).
Extractive summary: sentences quoted from the sources.
Before this
- Oct 21, 2026Original character LoRA for MiniMax H3 learnings. Workflows included.
- Oct 20, 2026Fizgig 8.8.1: a new MOSS-Image training technique
- Oct 20, 2026Gradium Open-Sources Phonon, a 31M-Parameter On-Device TTS Model Under Apache 2.0
- Oct 20, 2026How to train MOSS-Image LoRAs locally (8 GB VRAM): settings, tips and workflow
- Oct 20, 2026MPS in PyTorch 2.15
- Oct 19, 2026Moss image speed boost with spectrum moss node