You're looking at a specific version of this model. Jump to the model overview.

edenartlab /sdxl-pipelines:4df329a3

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
mode
string (enum)
generate

Options:

generate, remix, upscale, blend, controlnet, interpolate, real2real, interrogate

Mode
stream
boolean
False
yield individual results if True
stream_every
integer
1

Min: 1

Max: 25

for mode generate, how many steps per update to stream (steam must be set to True)
width
integer
1024

Min: 512

Max: 2048

Width
height
integer
1024

Min: 512

Max: 2048

Height
checkpoint
string (enum)
sdxl-v1.0

Options:

sdxl-v1.0

Which Stable Diffusion checkpoint to use
lora
string
(optional) URL of Lora finetuning
lora_scale
number
0.8

Max: 1.2

Lora scale (how much of the Lora finetuning to apply)
sampler
string (enum)
euler

Options:

ddim, ddpm, klms, euler, euler_ancestral, dpm, kdpm2, kdpm2_ancestral, pndm

Which sampler to use
steps
integer
35

Min: 10

Max: 100

Diffusion steps
guidance_scale
number
7.5

Min: 1

Max: 20

Strength of text conditioning guidance
upscale_f
number
1

Min: 1

Max: 2

Upscaling resolution
init_image_data
string
Load initial image from file, url, or base64 string
init_image_strength
number
0

Max: 1

Strength of initial image
adopt_aspect_from_init_img
boolean
True
Adopt aspect ratio from init image
controlnet_type
string (enum)
off

Options:

off, canny-edge, depth, luminance

Controlnet type
ip_image_data
string
Load ip_adapter image from file, url, or base64 string
ip_image_strength
number
0.65

Max: 1

Strength of image conditioning from ip_adapter (vs txt conditioning from clip-interrogator or prompt) (used in remix, upscale, blend and real2real)
text_input
string
Text input (mode==generate)
uc_text
string
nude, naked, text, watermark, low-quality, signature, padding, margins, white borders, padded border, moiré pattern, downsampling, aliasing, distorted, blurry, blur, jpeg artifacts, compression artifacts, poorly drawn, low-resolution, bad, grainy, error, bad-contrast
Negative text input (mode==all)
seed
integer
13

Max: 10000000000

random seed (mode==generate)
n_samples
integer
1

Min: 1

Max: 4

batch size (mode==generate)
n_frames
integer
40

Min: 3

Max: 1000

Total number of frames (mode==interpolate)
interpolation_texts
string
Interpolation texts (mode==interpolate)
interpolation_seeds
string
Seeds for interpolated texts (mode==interpolate)
interpolation_init_images
string
Interpolation init images, file paths or urls (mode==interpolate)
interpolation_init_images_power
number
2.5

Min: 0.5

Max: 5

Power for interpolation_init_images prompts (mode==interpolate)
interpolation_init_images_min_strength
number
0.05

Max: 1

Minimum init image strength for interpolation_init_images prompts (mode==interpolate)
interpolation_init_images_max_strength
number
0.95

Max: 1

Maximum init image strength for interpolation_init_images prompts (mode==interpolate)
loop
boolean
True
Loops (mode==interpolate & real2real)
smooth
boolean
True
Smooth (mode==interpolate & real2real)
latent_blending_skip_f
string
0.05|0.6
What fraction of the denoising trajectory to skip at the start and end of each interpolation phase, two floats, separated by a pipe (|)
n_anchor_imgs
integer
3

Min: 3

Max: 6

Number of anchor frames to render (including keyframes) before activating latent blending
n_film
integer
1

Max: 3

Number of times to smooth final frames with FILM (default is 0) (mode==interpolate)
fps
integer
12

Min: 1

Max: 30

Frames per second (mode==interpolate & real2real)

Output schema

The shape of the response you’ll get when you run this model with an API.

Schema
{'items': {'format': 'uri', 'type': 'string'},
 'title': 'Output',
 'type': 'array',
 'x-cog-array-type': 'iterator'}