Description

This LoRA is made to improve overall detail, texture, and rendering quality on Illustrious, for both characters and environments.

It helps create richer scenes without needing overly complex prompts, while also improving character textures, clothing details, materials, and hair definition. It is especially useful for semi-realistic anime renders where you want less flatness, less glossy/plastic rendering, more detailed hair, and more visually dense backgrounds.

Trigger words

semi-realistic

Often enough to bring out the intended effect by itself, especially for a natural semi-realistic look.

azm_envstyle

Mainly reinforces the environment/detail side of the LoRA. It can help with richer scenery, more detailed backgrounds, atmosphere, and overall visual density.

azm_chardet

Mainly reinforces the character detail side of the LoRA. It can help with character texture, hair detail, clothing detail, and reducing the glossy/plastic look.

The trigger words can be used separately or together depending on how strongly you want to guide the result.

Usage notes

The LoRA was designed to work with simple prompting. For backgrounds, even basic location prompts or adding detailed background to a character prompt can already help create a more complete scene.

For known characters already recognized by Illustrious, the LoRA may have a noticeable bias toward lighter fabric outfits. To preserve the official outfit more accurately, increasing the weight of the character name can help. For example, I had to use (Furina:1.2) to get Furina’s actual outfit more consistently.

The showcase images are raw outputs, without refining, retouching, upscaling, or any additional post-processing.

Results may vary depending on the model version, LoRA weight, prompt, and other LoRAs used.