🎨 Boogu-Image-0.1: Open-Source Models for Image Generation
The Boogu-Image-0.1 family, based on Qwen3-VL-8B-Instruct, has been introduced. The architecture, utilizing U-Net and CLIP, ensures high-quality text rendering in both English and Chinese. The lineup includes Base, Turbo (fast), Edit, and fp8 versions.
🌍 The emergence of such efficient open-source models sets a new direction for the development of compact multimodal systems that are comparable in quality to closed-source solutions.
👤 You can run powerful image generation and editing tools locally, using the fast Turbo versions for instantaneous results.
Source 1: https://github.com/boogu-project/Boogu-Image Source 2: https://huggingface.co/Boogu