Inference Providers documentation
Image to Image
Image to Image
Image-to-image is the task of transforming a source image to match the characteristics of a target image or a target image domain.
Example applications:
- Transferring the style of an image to another image
- Colorizing a black and white image
- Increasing the resolution of an image
For more details about the
image-to-imagetask, check out its dedicated page! You will find examples and related materials.
Recommended models
- black-forest-labs/FLUX.1-Kontext-dev: Powerful image editing model.
- kontext-community/relighting-kontext-dev-lora-v3: Image re-lighting model.
Explore all available models and find the one that suits you best here.
Using the API
Language
Client
Provider
import os
from huggingface_hub import InferenceClient
client = InferenceClient(
provider="fal-ai",
api_key=os.environ["HF_TOKEN"],
)
with open("cat.png", "rb") as image_file:
input_image = image_file.read()
# output is a PIL.Image object
image = client.image_to_image(
input_image,
prompt="Turn the cat into a tiger.",
model="black-forest-labs/FLUX.2-dev",
)API specification
Request
| Headers | ||
|---|---|---|
| authorization | string | Authentication header in the form 'Bearer: hf_****' when hf_**** is a personal user access token with “Inference Providers” permission. You can generate one from your settings page. |
| Payload | ||
|---|---|---|
| inputs* | string | The input image data as a base64-encoded string. If no parameters are provided, you can also provide the image data as a raw bytes payload. |
| parameters | object | |
| prompt | string | The text prompt to guide the image generation. |
| guidance_scale | number | For diffusion models. A higher guidance scale value encourages the model to generate images closely linked to the text prompt at the expense of lower image quality. |
| negative_prompt | string | One prompt to guide what NOT to include in image generation. |
| num_inference_steps | integer | For diffusion models. The number of denoising steps. More denoising steps usually lead to a higher quality image at the expense of slower inference. |
| target_size | object | The size in pixels of the output image. This parameter is only supported by some providers and for specific models. It will be ignored when unsupported. |
| width* | integer | |
| height* | integer |
Response
| Body | ||
|---|---|---|
| image | unknown | The output image returned as raw bytes in the payload. |