Qwen-Image's Lora is currently being continuously trained, and the current LORA is a test version.

Style Introduction:

「Landscape Narrative」: Lighting has an atmosphere, composition combines the sense of integration between the subject and the environment.

Qwen-Image Prompt Suggestions:

Clarity: Prompts should clearly and accurately describe the content of the image you want to generate, avoiding vague or ambiguous expressions. For example, if you want to generate an image of a cat, you should not just say "animal," but specify "a orange Persian cat lying on a soft white blanket, with blue eyes and a lazy expression."

Detailed Richness: Provide as many details as possible about the elements in the image, including the characteristics, actions, environment, lighting, colors, and style of the main subject. For example, when generating an image of a person, you can describe the hairstyle, clothing style and color, expression, posture, as well as whether the background is indoors or outdoors, and any specific decorations or weather conditions.

Style Specification: If there are specific artistic style requirements, such as realistic style, cartoon style, ancient style, cyberpunk style, etc., they should be clearly stated in the prompt. For example, "An ancient-style illustration, a woman in a light blue ru skirt standing under a tree full of peach blossoms, holding a round fan, with delicate and beautiful painting style, and elegant colors."

Logical Reasonableness: The content described in the prompt should be logically reasonable, and the elements should be reasonably combined. For example, it should not include descriptions that contradict common sense, such as "a fish in space surrounded by flowing seawater."

Language Simplicity: The language should be concise and clear while ensuring complete information, avoiding long-winded and redundant expressions, which may affect the model's ability to capture key information.