datong-new
/
lora-doll
- Public
- 20 runs
Run datong-new/lora-doll with an API
Use one of our client libraries to get started quickly. Clicking on a library will take you to the Playground tab where you can tweak different inputs, see the results, and copy the corresponding code to use in your own project.
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 |
|---|---|---|---|
| image |
string
|
An image of a person to be converted
|
|
| prompt |
string
|
a person
|
None
|
| negative_prompt |
string
|
|
Things you do not want in the image
|
| denoising_strength |
number
|
0.65
Max: 1 |
How much of the original image to keep. 1 is the complete destruction of the original image, 0 is the original image
|
| prompt_strength |
number
|
4.5
Max: 20 |
Strength of the prompt. This is the CFG scale, higher numbers lead to stronger prompt, lower numbers will keep more of a likeness to the original.
|
| control_depth_strength |
number
|
0.8
Max: 1 |
Strength of depth controlnet. The bigger this is, the more controlnet affects the output.
|
| human_face_strength |
number
|
1
Max: 1 |
How strong the face similarity.
|
| seed |
integer
|
Fix the random seed for reproducibility
|
|
| custom_lora_url |
string
|
URL to a Replicate custom LoRA. Must be in the format https://replicate.delivery/pbxt/[id]/trained_model.tar or https://pbxt.replicate.delivery/[id]/trained_model.tar
|
|
| lora_scale |
number
|
1
Max: 1 |
How strong the LoRA will be
|
{
"type": "object",
"title": "Input",
"properties": {
"seed": {
"type": "integer",
"title": "Seed",
"x-order": 7,
"description": "Fix the random seed for reproducibility"
},
"image": {
"type": "string",
"title": "Image",
"format": "uri",
"x-order": 0,
"description": "An image of a person to be converted"
},
"prompt": {
"type": "string",
"title": "Prompt",
"default": "a person",
"x-order": 1
},
"lora_scale": {
"type": "number",
"title": "Lora Scale",
"default": 1,
"maximum": 1,
"minimum": 0,
"x-order": 9,
"description": "How strong the LoRA will be"
},
"custom_lora_url": {
"type": "string",
"title": "Custom Lora Url",
"x-order": 8,
"description": "URL to a Replicate custom LoRA. Must be in the format https://replicate.delivery/pbxt/[id]/trained_model.tar or https://pbxt.replicate.delivery/[id]/trained_model.tar"
},
"negative_prompt": {
"type": "string",
"title": "Negative Prompt",
"default": "",
"x-order": 2,
"description": "Things you do not want in the image"
},
"prompt_strength": {
"type": "number",
"title": "Prompt Strength",
"default": 4.5,
"maximum": 20,
"minimum": 0,
"x-order": 4,
"description": "Strength of the prompt. This is the CFG scale, higher numbers lead to stronger prompt, lower numbers will keep more of a likeness to the original."
},
"denoising_strength": {
"type": "number",
"title": "Denoising Strength",
"default": 0.65,
"maximum": 1,
"minimum": 0,
"x-order": 3,
"description": "How much of the original image to keep. 1 is the complete destruction of the original image, 0 is the original image"
},
"human_face_strength": {
"type": "number",
"title": "Human Face Strength",
"default": 1,
"maximum": 1,
"minimum": 0,
"x-order": 6,
"description": "How strong the face similarity."
},
"control_depth_strength": {
"type": "number",
"title": "Control Depth Strength",
"default": 0.8,
"maximum": 1,
"minimum": 0,
"x-order": 5,
"description": "Strength of depth controlnet. The bigger this is, the more controlnet affects the output."
}
}
}
Output schema
The shape of the response you’ll get when you run this model with an API.
Schema
{
"type": "object",
"title": "Output",
"required": [
"audio"
],
"properties": {
"audio": {
"type": "string",
"title": "Audio",
"format": "uri"
}
}
}