1
Upload the image you need Omnigen to reference, either a character, an item, or a condition map, up to a maximum of 3 images. If you wish to merge 2 characters into one image, you can optionally upload photos of 2 characters. If you wish to have 2 items in the same image, you can upload images of 2 items.
2
Describe the image you want Omnigen to generate in the prompt. For the parts that involve elements in the image, you need to introduce them in the prompt in the format of <img><|image_i|></img>, such as the people in this photo <img><|image_i|></img>, or the people in this photo <img><| image_i|></img>. i represents the serial number of the image you uploaded. You can refer to the example for the prompt.
3
You can adjust Omnigen's generation parameters, such as the ratio of the images generated, in the settings. The rest of the settings are recommended to be left as default.
4
After the settings are finished, click Generate and you will enter the queue to wait for the image to be generated.
OmniGen enables you to craft images with simple text prompts or multi-modal inputs, using placeholders for image-conditioned prompts. Perfect for creating diverse and contextually rich visuals.
Maintain subject identity with OmniGen's advanced model, making it ideal for recognizable figures and consistent character rendering.
This feature allows for personalized and highly tailored imagery creation.
Edit previously generated images with OmniGen’s flexible seed handling. For example, switch to a new seed when modifying an existing image for unique edits. Perfect for image refinement and experimentation.
Detailed prompts yield clearer and higher-quality results. OmniGen’s model adapts to specific prompts to generate visuals that suit professional and creative needs alike.
FAQS