I’m testing gradio API access for a HuggingFace Space; for an API that requires we post one image via its URL, I have successfully used:
data = gr.processing_utils.encode_url_or_file_to_base64(image_url)
r = requests.post(url=‘https://hf.space/embed/tmabraham/fastai_pet_classifier/+/api/predict/’, json={“data”:[data]})
Can anyone provide an example of posting a concatenated data request to HF in a similar way? The API below expects (2) images (as str, as above) and an additional string. What is the best practice?
Input Payload
{
“data”: [
str | Dict, // represents base64 url data, or (if tool == “sketch”) a dict of image and mask base64 url data of ‘Image1’ image component
str | Dict, // represents base64 url data, or (if tool == “sketch”) a dict of image and mask base64 url data of ‘Image2’ image component
str, // represents selected choice of ‘Task’ radio component
]
}
Response Object
{
“data”: [
str, // represents base64 url data of ‘flow/disparity’ image component
],
“duration”: (float) // number of seconds to run function call
}