2026-10-05 · America/Los_Angeles · 社区动态 · #15

Fizgig 7 - Anima and SDXL arrive, fine-tuning for every model & the community can now add support for new/old models

Anima and SDXL join using the new model driver system, with LoRA and LoKR training, sliders, fine-tuning and the full workbench (Repair Studio, LoRA the Explorer, Profiler, Extract, LoRA Royale). SDXL takes any single-file checkpoint; Juggernaut XL v9 is the default. Full fine-tuning on every model: Klein 9B, Krea 2, Qwen Image 2.1, MiniMax H3, SDXL and Anima. Every model fine-tunes on a 16 GB card, Anima and SDXL from 8 GB. On a bigger card it trains the whole model at once instead of a part at a time, which got through the model about 2.6x faster in my tests. Model Fine-tunes from Whole…

热度 45.9 / 100；排名与评分保留该期记录。