@mindstudio-ai/remy 0.1.37 → 0.1.38
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
|
@@ -46,6 +46,7 @@ Lead with the visual style, then describe the content. This order helps the mode
|
|
|
46
46
|
- Describing positions of arms, legs, or specific limb arrangements.
|
|
47
47
|
- Conflicting style instructions ("photorealistic cartoon").
|
|
48
48
|
- Describing what you don't want — say "empty street" not "street with no cars."
|
|
49
|
+
- Framing prompts as physical objects ("artwork", "painting", "canvas", "print", "square digital artwork"). The model may render a photograph of the object on a surface with shadows rather than the image itself. Describe the visual content directly.
|
|
49
50
|
- It is critical to remember that image models have a high risk of rendering text. Any word or phrase in your prompt that could be interpreted as a title, label, or caption risks appearing as literal text in the image. Triggers like "magazine cover" also risk making it render a literal mockup of a magazine masthread, even if all you wanted was a certain photography stype. Common triggers: "poster", "editorial", "magazine", "cover", "sign", or brand names, industry jargon, etc. Be thoughtful, careful, and intentional with your prompt - especially when describing abtract visualizations - and describe the visual qualities you want instead of referencing formats or concepts as shorthand.
|
|
50
51
|
|
|
51
52
|
### How images work in the UI
|