tgohblio / instant-id-multicontrolnet

InstantID. ControlNets. More base SDXL models. And the latest ByteDance's ⚡️SDXL-Lightning !⚡️ (Updated 1 year, 1 month ago)

  • Public
  • 296.6K runs
  • GitHub
  • Paper
  • License
Iterate in playground

Input

*file
Preview
face_image_path

Image of your face

file

Reference pose image

string
Shift + Return to add a new line

Input prompt

Default: "a person"

string
Shift + Return to add a new line

Input negative prompt

Default: "(lowres, low quality, worst quality:1.2), (text:1.2), watermark, glitch, deformed, mutated, cross-eyed, ugly, disfigured, blurry, grainy"

string

Select SDXL model

Default: "AlbedoBase XL V2"

boolean

Enable SDXL-lightning fast inference. If pose, canny or depth map is used, disable it for better quality images.

Default: true

string

if enable fast mode, choose number of denoising steps

Default: "4step"

string

Scheduler options. If enable fast mode, this is not used.

Default: "DPMSolverMultistepScheduler"

number
(minimum: 0, maximum: 1)

Image adapter strength (for detail)

Default: 0.8

number
(minimum: 0, maximum: 1)

IdentityNet strength (for fidelity)

Default: 0.8

boolean

Use pose for skeleton inference

Default: false

number
(minimum: 0, maximum: 1.5)

Default: 1

boolean

Use canny for edge detection

Default: false

number
(minimum: 0, maximum: 1.5)

Default: 0.5

boolean

Use depth for depth map estimation

Default: false

number
(minimum: 0, maximum: 1.5)

Default: 0.5

integer
(minimum: 1, maximum: 50)

Number of denoising steps. If enable fast mode, this is not used.

Default: 25

number
(minimum: 0, maximum: 10)

Scale for classifier-free guidance. Optimum is 4-8. If enable fast mode, this is not used.

Default: 7

integer
(minimum: 0, maximum: 2147483647)

Seed number. Set to non-zero to make the image reproducible.

Default: 0

boolean

Enhance non-face region

Default: true

boolean

Safety checker is enabled by default. Un-tick to expose unfiltered results.

Default: true

Output

output
Generated in

Run time and cost

This model costs approximately $0.023 to run on Replicate, or 43 runs per $1, but this varies depending on your inputs. It is also open source and you can run it on your own computer with Docker.

This model runs on Nvidia L40S GPU hardware. Predictions typically complete within 24 seconds. The predict time for this model varies significantly based on the inputs.

Readme

This model doesn't have a readme.