Dev builds free Colab Gradio UI for Z-Image Turbo with upscaling and face fix
OkEnvironment1732 · reddit · 2026-09-23
A developer released an open-source project (YassineMenai/easyai) that wraps Z-Image Turbo in a simple Gradio UI running on a free Colab (L4 recommended), aimed at users who find ComfyUI overwhelming.
Features:
- Text-to-image (standard, enhanced, and Z-Sampler Turbo modes)
- SeedVR2 upscaler with a refine pass
- Automatic face/hand fix via YOLO detection + inpainting
- LoRA support with A/B testing
- Prompt enhancement via QwenVL
No nodes, wires, or workflow files — just generate.
More from Multimodal
- Tilly Norwood debate: synthetic performers as production infrastructure beats novelty deepfakes — alexvoica · 2026-09-23
- PixVerse R2 tackles the real-time vs. generality tradeoff with a dual-engine world model — nikola_mr64990 · 2026-09-23
- First Krea 2 LoRA for controllable human camera angles generates 360° orbits from one image — Upbeat_Birthday_6123 · 2026-09-23
- Opus 5.5 tested with a vague 'draw something beautiful' prompt — mimi10v3 · 2026-09-23
- Vivago R1 adds Cinematic Story Director for consistent multi-shot AI films — socialwithaayan · 2026-09-23
- QwenImage 2.1 vs Edit 2511 tested: the 7GB model doesn't lose every battle — DiagramAwesome · 2026-09-23