First image Text to image

Make your first image from a description, and meet the main settings on the way.

8 stepsNothing to prepare

The same tutorial, three subjects

Step by step

Explanation detail

  1. Prompt

    1 · Prompt

    Use to describe the subject, setting, and appearance you want. describes features to reduce or avoid. This example uses:Prompt:
    a woman in her 30s with a sleek silver bob haircut, round copper goggles resting up in her silver hair, a patched teal canvas coat, a small brass compass on a cord around her neck, standing at the wooden railing of a zeppelin airship's open gondola deck at sunset, the vast canvas envelope of the airship curving overhead, brass fittings and taut rigging, wind pulling at her coat, a sea of clouds below, cinematic lighting, rich detail, professional photography, high dynamic range
    Negative prompt:
    lowres, blurry, deformed, extra limbs, watermark, text, jpeg artifacts, ocean, sea, water surface, sailing ship, boat, cruise ship, sunglasses

  2. Model & style

    2 · Model

    A is a trained network used to generate images. Models differ in the subjects they handle, their appearance, and supported settings.Select AlbedoBase XL (SDXL) for this example. You can compare other models later using the same prompt.Set to 1. The “Set this for me” button selects AlbedoBase XL (SDXL) and sets CLIP skip to 1. Models can need different CLIP skip settings; using the wrong one can give poor results.

  3. Size

    3 · Image size

    Set to 1024×1024 pixels for this example with AlbedoBase XL (SDXL).Dimensions affect composition as well as image size. Start with the model's presets or suggestions. Very small sizes may lose details; very large sizes may introduce repeated or distorted subjects.

  4. Sampling

    4 · Steps

    Set to 30. Sampling builds the image through a series of updates.Too few steps may leave an incomplete image. Extra steps take more time and can eventually add little. The useful count depends on the model and sampler.
  5. Sampling

    5 · Guidance

    Set to 7.5. Guidance controls the influence of the prompt during generation.Increasing it can improve prompt agreement, but excessive values may produce harsh contrast or distorted shapes. Follow the model's suggestions when trying a different model.
  6. Sampling

    6 · Sampler

    Choose k_euler_a in . A sampler is the method used to update the image during generation.Samplers differ in speed and in how they respond to step count. Use this one to follow the example; check the model's suggestions when experimenting.
  7. Sampling

    7 · Seed

    Set to 1234567890 to use the example's seed. This number helps determine the starting noise.Keep it fixed while comparing settings. Different worker hardware or software can still change the result.Leave it blank for a random seed. Save an image's seed and settings when you want to return to it.

  8. 8 · Generation

    Press Generate. A compatible volunteer worker will generate the image when it can accept the job.Follow the queue estimate. Waiting time depends on demand, available workers, and the request. Compare the result with the example; an exact match is not guaranteed. This run took about 20 seconds and cost 45 kudos.

Related tutorial

Try img2img to change an existing image. You can use the image you just generated as its source.