You're looking at a specific version of this model. Jump to the model overview.
Input schema
The fields you can use to run this model with an API. If you don’t give a value for a field its default value will be used.
| Field | Type | Default value | Description |
|---|---|---|---|
| prompt |
string
|
|
None
|
| negative_prompt |
string
|
|
Things you do not want to see in your image
|
| number_of_images |
integer
|
1
Min: 1 Max: 10 |
Number of images to generate
|
| width |
integer
|
1024
|
None
|
| height |
integer
|
1024
|
None
|
| lora_strength |
number
|
1
Max: 3 |
Strength of the lora to use for the generation. Default is 1.0.
|
| image |
string
|
Optional. Image to use for img2img generation. Leave empty to use just a prompt.
|
|
| denoise_strength |
number
|
0.65
Max: 1 |
How much of the original input image to destroy when using img2img. 1 is total destruction. 0 is untouched. 0.65 is a good balance.
|
| output_format |
string
(enum)
|
webp
Options: webp, jpg, png |
Format of the output images
|
| output_quality |
integer
|
80
Max: 100 |
Quality of the output images, from 0 to 100. 100 is best quality, 0 is lowest quality.
|
| seed |
integer
|
Set a seed for reproducibility. Random by default.
|
|
| disable_safety_checker |
boolean
|
False
|
Disable safety checker for generated images.
|
| sampler |
string
(enum)
|
Default
Options: Default, euler, euler_ancestral, heun, heunpp2, dpm_2, dpm_2_ancestral, lms, dpm_fast, dpm_adaptive, dpmpp_2s_ancestral, dpmpp_sde, dpmpp_sde_gpu, dpmpp_2m, dpmpp_2m_sde, dpmpp_2m_sde_gpu, dpmpp_3m_sde, dpmpp_3m_sde_gpu, ddpm, lcm, ddim, uni_pc, uni_pc_bh2 |
Advanced. Change the sampler used for generation. Default is what we think gives the best images.
|
| scheduler |
string
(enum)
|
Default
Options: Default, normal, karras, exponential, sgm_uniform, simple, ddim_uniform |
Advanced. Change the scheduler used for generation. Default is what we think gives the best images.
|
| steps |
integer
|
Min: 1 Max: 50 |
Advanced. Leave empty to use recommended steps. Set it only if you want to customise the number of steps to run the sampler for.
|
| cfg |
number
|
Max: 20 |
Advanced. Leave empty to use recommended CFG (classifier free guidance). This changes how much the prompt influences the output. Set it only if you want to customise.
|
| controlnet_image |
string
|
Advanced. The image to use with a controlnet. Leave empty to use no controlnet.
|
|
| controlnet_model |
string
(enum)
|
None
Options: None, control_boxdepth_LooseControlfp16.safetensors, control_sd15_inpaint_depth_hand_fp16.safetensors, control_v11e_sd15_ip2p.pth, control_v11e_sd15_shuffle.pth, control_v11f1e_sd15_tile.pth, control_v11f1p_sd15_depth.pth, control_v11p_sd15_canny.pth, control_v11p_sd15_lineart.pth, control_v11p_sd15_mlsd.pth, control_v11p_sd15_normalbae.pth, control_v11p_sd15_openpose.pth, control_v11p_sd15_scribble.pth, control_v11p_sd15_seg.pth, control_v11p_sd15_softedge.pth, control_v11p_sd15s2_lineart_anime.pth, control_v11u_sd15_tile_fp16.safetensors, control_v1p_sd15_qrcode_monster.safetensors, control_v1p_sdxl_qrcode_monster.safetensors, controlnet-canny-sdxl-1.0.safetensors, controlnet-canny-sdxl-1.0_V2.safetensors, controlnet-depth-sdxl-1.0.fp16.safetensors, controlnet-openpose-sdxl-1.0.safetensors, controlnet-openpose-sdxl-1.0_twins.safetensors, controlnet-scribble-sdxl-1.0.safetensors, controlnet-sd-xl-1.0-softedge-dexined.safetensors, depth-anything.safetensors, depth-zoe-xl-v1.0-controlnet.safetensors, diffusers_xl_canny_full.safetensors, diffusers_xl_canny_mid.safetensors, diffusers_xl_canny_small.safetensors, diffusers_xl_depth_full.safetensors, diffusers_xl_depth_mid.safetensors, diffusers_xl_depth_small.safetensors, ioclab_sd15_recolor.safetensors, mistoLine_fp16.safetensors, mistoLine_rank256.safetensors, OpenPoseXL2.safetensors, sargezt_xl_depth.safetensors, sargezt_xl_depth_faid_vidit.safetensors, sargezt_xl_depth_zeed.safetensors, sargezt_xl_softedge.safetensors, t2i-adapter_diffusers_xl_canny.safetensors, t2i-adapter_diffusers_xl_depth_midas.safetensors, t2i-adapter_diffusers_xl_depth_zoe.safetensors, t2i-adapter_diffusers_xl_lineart.safetensors, t2i-adapter_diffusers_xl_openpose.safetensors, t2i-adapter_diffusers_xl_sketch.safetensors, t2i-adapter_xl_canny.safetensors, t2i-adapter_xl_openpose.safetensors, t2i-adapter_xl_sketch.safetensors, t2iadapter_canny_sd14v1.pth, t2iadapter_color_sd14v1.pth, t2iadapter_depth_sd14v1.pth, t2iadapter_keypose_sd14v1.pth, t2iadapter_openpose_sd14v1.pth, t2iadapter_seg_sd14v1.pth, t2iadapter_sketch_sd14v1.pth, t2iadapter_style_sd14v1.pth, thibaud_xl_openpose.safetensors, thibaud_xl_openpose_256lora.safetensorscontrol-lora-canny-rank128.safetensors, control-lora-canny-rank256.safetensors, control-lora-depth-rank128.safetensors, control-lora-depth-rank256.safetensors, control-lora-recolor-rank128.safetensors, control-lora-recolor-rank256.safetensors, control-lora-sketch-rank128-metadata.safetensors, control-lora-sketch-rank256.safetensors, control_lora_rank128_v11e_sd15_ip2p_fp16.safetensors, control_lora_rank128_v11e_sd15_shuffle_fp16.safetensors, control_lora_rank128_v11f1e_sd15_tile_fp16.safetensors, control_lora_rank128_v11f1p_sd15_depth_fp16.safetensors, control_lora_rank128_v11p_sd15_canny_fp16.safetensors, control_lora_rank128_v11p_sd15_inpaint_fp16.safetensors, control_lora_rank128_v11p_sd15_lineart_fp16.safetensors, control_lora_rank128_v11p_sd15_mlsd_fp16.safetensors, control_lora_rank128_v11p_sd15_normalbae_fp16.safetensors, control_lora_rank128_v11p_sd15_openpose_fp16.safetensors, control_lora_rank128_v11p_sd15_scribble_fp16.safetensors, control_lora_rank128_v11p_sd15_seg_fp16.safetensors, control_lora_rank128_v11p_sd15_softedge_fp16.safetensors, control_lora_rank128_v11p_sd15s2_lineart_anime_fp16.safetensors, sai_xl_canny_128lora.safetensors, sai_xl_canny_256lora.safetensors, sai_xl_depth_128lora.safetensors, sai_xl_depth_256lora.safetensors, sai_xl_recolor_128lora.safetensors, sai_xl_recolor_256lora.safetensors, sai_xl_sketch_128lora.safetensors, sai_xl_sketch_256lora.safetensors |
Advanced. Change the model used for the controlnet. Make sure it matches the checkpoint type (eg SD15 or SDXL) and pick an appropriate controlnet preprocessor to go with it.
|
| controlnet_preprocessor |
string
(enum)
|
None
Options: None, CannyEdgePreprocessor, AnyLineArtPreprocessor_aux, LineArtPreprocessor, LineartStandardPreprocessor, MiDaS-DepthMapPreprocessor, DepthAnythingPreprocessor, Zoe-DepthMapPreprocessor, Zoe_DepthAnythingPreprocessor, HEDPreprocessor, OpenposePreprocessor, DWPreprocessor, AnimalPosePreprocessor, AnimeFace_SemSegPreprocessor, AnimeLineArtPreprocessor, BAE-NormalMapPreprocessor, BinaryPreprocessor, ColorPreprocessor, DSINE-NormalMapPreprocessor, DensePosePreprocessor, DiffusionEdge_Preprocessor, FakeScribblePreprocessor, ImageIntensityDetector, ImageLuminanceDetector, LeReS-DepthMapPreprocessor, M-LSDPreprocessor, Manga2Anime_LineArt_Preprocessor, MediaPipe-FaceMeshPreprocessor, MeshGraphormer-DepthMapPreprocessor, Metric3D-DepthMapPreprocessor, Metric3D-NormalMapPreprocessor, MiDaS-NormalMapPreprocessor, OneFormer-ADE20K-SemSegPreprocessor, OneFormer-COCO-SemSegPreprocessor, PiDiNetPreprocessor, SAMPreprocessor, ScribblePreprocessor, Scribble_PiDiNet_Preprocessor, Scribble_XDoG_Preprocessor, SemSegPreprocessor, ShufflePreprocessor, TEEDPreprocessor, TTPlanet_TileGF_Preprocessor, TTPlanet_TileSimple_Preprocessor, TilePreprocessor, UniFormer-SemSegPreprocessor |
Advanced. Change the preprocessor used for the controlnet. Make sure it matches the controlnet model you're using.
|
| controlnet_return_preprocessor_image |
boolean
|
False
|
Advanced. Return the processed image from the controlnet preprocessor. Useful for debugging.
|
| controlnet_strength |
number
|
1
Max: 3 |
Advanced. Strength of the controlnet. Default is 1.0.
|
| controlnet_start |
number
|
0
Max: 1 |
Advanced. When to begin applying the controlnet.
|
| controlnet_end |
number
|
1
Max: 1 |
Advanced. When to stop applying the controlnet.
|
Output schema
The shape of the response you’ll get when you run this model with an API.
{'items': {'format': 'uri', 'type': 'string'},
'title': 'Output',
'type': 'array'}