How to show the output from a request's response from a huggingface text to image API?

Question

import requests
from PIL import Image


API_URL = "https://api-inference.huggingface.co/models/runwayml/stable-diffusion-v1-5"
headers = {"Authorization": f"Bearer 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"}

def query(payload):
    response = requests.post(API_URL, headers=headers, json=payload)
    return response.content

image_bytes = query({
"inputs": "Astronaut riding a horse",
})


image = Image.open(io.BytesIO(image_bytes))

I ran through Inference API, created my token.. pasted it in headers. i was expecting that as soon as i run the code .

The image will appear on screen on the terminal but it didnt happen and also no error was shown.

Answer 1

You need to use .show() to display the image in pillow, e.g.

import io

import requests
from PIL import Image


endpoint = "https://api-inference.huggingface.co/models/runwayml/stable-diffusion-v1-5"
headers = {"Authorization": f"Bearer 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"}


payload = {"inputs": "Astronaut riding a horse"}
response = requests.post(endpoint, headers=headers, json=payload)


im = Image.open(io.BytesIO(response.content))
im.show()  # This will open up the image in your OS default image viewer

To save the image to file, there are several options https://stackoverflow.com/a/75824531/610569, you can try this to save as a numpy array:

import io

import requests

from PIL import Image
import numpy as np


endpoint = "https://api-inference.huggingface.co/models/runwayml/stable-diffusion-v1-5"
headers = {"Authorization": f"Bearer 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"}


payload = {"inputs": "Astronaut riding a horse"}
response = requests.post(endpoint, headers=headers, json=payload)


img = Image.open(io.BytesIO(response.content))

# Converts and save image into numpy array.
np.save('myimage.npy', np.asarray(img))

Then to read the saved file:

import numpy as np
from PIL import Image

# Loads a npy file to Image
img_arr = np.load('dpreview.npy')
img = Image.fromarray(img_arr.astype(np.uint8))

Note: While the code snippet above works, I'm not sure if (ab)using the API to to do requests.post() on api-inference.huggingface.co/models/runwayml is encouraged.