Illustrious XL 1.0 Improved Uncensored v3.0 drops on HuggingFace for anime generation
ChouiRRV released an uncensored SDXL fine-tune targeting anime-style generation, now live on HuggingFace with diffusers and safetensors support.
ChouiRRV released Illustrious XL 1.0 Improved Uncensored v3.0, an uncensored Stable Diffusion XL fine-tune for anime-style character generation. The checkpoint landed on HuggingFace on July 3 tagged for text-to-image workflows, with weights distributed in both diffusers and safetensors formats. The model card carries HuggingFace's "not-for-all-audiences" flag, signaling unrestricted output capability.
Open-weight SDXL derivatives like this run locally without server-side safety filters, a draw for practitioners building custom pipelines or running private inference. Dual-format distribution—diffusers for HuggingFace Transformers workflows and safetensors for ComfyUI, Automatic1111, and Forge—reflects standard practice in the open-weight image-model community. Safetensors offer faster load times and a security advantage over legacy pickle formats, while diffusers integration simplifies scripting for Python-native pipelines.
The Illustrious lineage has carved out a niche among anime-focused SDXL users who prize stylistic consistency over photorealism. Fine-tunes in this category typically layer additional training on curated image sets, trading general-purpose capability for tighter control over character design, line weight, and color palette.
