Community finetunes Qwen-Image-2.1 VAE to kill checkerboard artifacts, plus realtime AE
bdsqlsz · x · 2026-09-30
Developer madebyollin released two unofficial VAE variants for Qwen-Image-2.1:
- Texture-Fix-VAE: the original VAE decoder finetuned for 5000 steps at lr 3e-5, with only the two highest-resolution decoder stages and output head unfrozen (7.5M trainable params). It removes checkerboard artifacts, with the biggest gains on detailed photo-style images.
- A distilled tiny AE for realtime encode/decode, e.g. live previews.
Both are on Hugging Face with ComfyUI (drop into models/vae and swap in the Load VAE node) and Diffusers integration snippets included.
More from Multimodal
- Testing Qwen Image 2.1 and Minimax H3 on anime-to-photo conversion and film photo restoration — Ant_6431 · 2026-11-16
- Second take of a 'Cubee run' video generated with Minimax H3 — JahJedi · 2026-11-16
- Non-artist builds 14-chapter hand-painted storybook game with Fable 5.2 in 4 hours — jradoff · 2026-11-16
- Subject LoRAs on Minimax H3 Image learn poor likenesses while identical datasets excel on Qwen 2.1 — Ad_Organic · 2026-11-16
- Seedance-generated anime train scene stuns Reddit with cinematic quality — Ok-Solution-7889 · 2026-11-16
- AI-generated 'Men in Black' reimagined in 1977 goes viral — techhalla · 2026-11-16