adirik/realvisxl-v3.0-turbo
Photorealism with RealVisXL V3.0 Turbo based on SDXL
Capabilities
Cost
Community model (estimated from hardware time)
Input Parameters
| Name | Type | Description | Default | Constraints |
|---|---|---|---|---|
apply_watermark | boolean | Applies a watermark to enable determining if an image is generated in downstream applications. If you have other provisions for generating or deploying images safely, you can use this to disable watermarking. | false | — |
disable_safety_checker | boolean | Disable safety checker for generated images. This feature is only available through the API. See [https://replicate.com/docs/how-does-replicate-work#safety](https://replicate.com/docs/how-does-replicate-work#safety) | false | — |
guidance_scale | number | Scale for classifier-free guidance | 2 | min: 1, max: 50 |
height | integer | Height of output image | 768 | — |
high_noise_frac | number | For expert_ensemble_refiner, the fraction of noise to use | 0.8 | min: 0, max: 1 |
image | string(uri) | Input image for img2img or inpaint mode | — | — |
mask | string(uri) | Input mask for inpaint mode. Black areas will be preserved, white areas will be inpainted. | — | — |
negative_prompt | string | Input Negative Prompt | "(worst quality, low quality, illustration, 3d, 2d, painting, cartoons, sketch), open mouth" | — |
num_inference_steps | integer | Number of denoising steps | 25 | min: 1, max: 500 |
num_outputs | integer | Number of images to output. | 1 | min: 1, max: 4 |
prompt | string | Input prompt | "An astronaut riding a rainbow unicorn" | — |
prompt_strength | number | Prompt strength when using img2img / inpaint. 1.0 corresponds to full destruction of information in image | 0.8 | min: 0, max: 1 |
refine | string | Which refine style to use | "no_refiner" | no_refinerexpert_ensemble_refinerbase_image_refiner |
refine_steps | integer | For base_image_refiner, the number of steps to refine, defaults to num_inference_steps | — | — |
scheduler | string | Scheduler to use, DPM++ SDE Karras is recommended | "DPM++_SDE_Karras" | DDIMDPMSolverMultistepHeunDiscreteKarrasDPMK_EULER_ANCESTRALK_EULERPNDMDPM++_SDE_Karras |
seed | integer | Random seed. Leave blank to randomize the seed | — | — |
width | integer | Width of output image | 768 | — |
apply_watermarkbooleanApplies a watermark to enable determining if an image is generated in downstream applications. If you have other provisions for generating or deploying images safely, you can use this to disable watermarking.
falsedisable_safety_checkerbooleanDisable safety checker for generated images. This feature is only available through the API. See [https://replicate.com/docs/how-does-replicate-work#safety](https://replicate.com/docs/how-does-replicate-work#safety)
falseguidance_scalenumberScale for classifier-free guidance
2min: 1, max: 50heightintegerHeight of output image
768high_noise_fracnumberFor expert_ensemble_refiner, the fraction of noise to use
0.8min: 0, max: 1imagestringInput image for img2img or inpaint mode
maskstringInput mask for inpaint mode. Black areas will be preserved, white areas will be inpainted.
negative_promptstringInput Negative Prompt
"(worst quality, low quality, illustration, 3d, 2d, painting, cartoons, sketch), open mouth"num_inference_stepsintegerNumber of denoising steps
25min: 1, max: 500num_outputsintegerNumber of images to output.
1min: 1, max: 4promptstringInput prompt
"An astronaut riding a rainbow unicorn"prompt_strengthnumberPrompt strength when using img2img / inpaint. 1.0 corresponds to full destruction of information in image
0.8min: 0, max: 1refinestringWhich refine style to use
"no_refiner"refine_stepsintegerFor base_image_refiner, the number of steps to refine, defaults to num_inference_steps
schedulerstringScheduler to use, DPM++ SDE Karras is recommended
"DPM++_SDE_Karras"seedintegerRandom seed. Leave blank to randomize the seed
widthintegerWidth of output image
7683dc73c805b11Updated: 7/25/2026615.3K runs
cinemasetfree.com