|
|
--- |
|
|
library_name: diffusers |
|
|
--- |
|
|
|
|
|
a simple block to prepare all your default image inputs: takes a image or url, `size` or pocessor_id in controlnet_aux |
|
|
|
|
|
used for testing/demo |
|
|
|
|
|
|
|
|
to create the pipelline |
|
|
```python |
|
|
from diffusers import ModularPipeline |
|
|
get_image_step = ModularPipeline.from_pretrained("YiYiXu/image_inputs") |
|
|
``` |
|
|
|
|
|
common image inputs |
|
|
```python |
|
|
init_image = get_image_step(image_url=url,output="image") |
|
|
control_image = get_image_step(image_url=url, processor_id="canny",output="image") |
|
|
controlnet_union_image = get_image_step(image_url=url, processor_id="lineart_anime",output="image") |
|
|
inpaint_image = get_image_step(image_url=inpaint_img_url, size=(1024, 1024),output="image") |
|
|
inpaint_mask = get_image_step(image_url=inpaint_mask_url, size=(1024, 1024),output="image") |
|
|
ip_adapter_image = get_image_step(image_url=ip_adapter_image_url,output="image") |
|
|
``` |
|
|
|
