How to have custom output size for inference API

jetpackjules · February 16, 2023, 9:47pm

I am using the free version of inference API’s with a custom Stable Diffusion model. When trying to use the endpoint, I am unable to figure out how to set a custom size for the output (like how in the pipeline it’s just height=x and width=x)…

Is this possible?

radames · February 16, 2023, 10:22pm

Hi @jetpackjules, the Inference API only accepts the prompt parameter such as inputs, and most recently we have also added support for negative_prompts. If you need access to extra params you can deploy it as a Space or a custom Inference Endpoints.

github.com

huggingface/api-inference-community/blob/main/docker_images/diffusers/app/pipelines/text_to_image.py#L57


      
                  self.ldm.unet.load_attn_procs(
                      model_id, use_auth_token=os.getenv("HF_API_TOKEN")
                  )
          
          
    if isinstance(self.ldm, (StableDiffusionPipeline, AltDiffusionPipeline)):
                  self.ldm.scheduler = DPMSolverMultistepScheduler.from_config(
                      self.ldm.scheduler.config
                  )
          
          
def __call__(
              self, inputs: str, inference_steps=25, negative_prompt=None, **kwargs
          ) -> "Image.Image":
              """
              Args:
                  inputs (:obj:`str`):
                      a string containing some text
              Return:
                  A :obj:`PIL.Image` with the raw image representation as PIL.
              """
          
          
    if kwargs:

jetpackjules · February 16, 2023, 10:30pm

Thanks for the quick reply!

Does this mean that there is no way to accomplish this for free?

radames · February 16, 2023, 10:41pm

You can try to find a Space with our community granted GPU that has an open API.

jetpackjules · February 16, 2023, 10:48pm

Not sure how to do that…

Also would it still work with a custom model?

Topic		Replies	Views
Adapting a model from Spaces to Inference Endpoint Inference Endpoints on the Hub	3	1464	November 25, 2022
Custom containers - setting args Inference Endpoints on the Hub	0	171	November 7, 2023
Negative prompts for the inference api 🤗Hub	10	2090	January 10, 2024
No custom pipeline found at /repository/handler.py Inference Endpoints on the Hub	4	555	April 3, 2023
Are there any parameters for Stable DIffiusion inference api Beginners	0	762	February 6, 2023

How to have custom output size for inference API

Related Topics