Using the Inference API

[CemHaligur]

Search before asking

• I have searched the HUB issues and found no similar bug report.

HUB Component

No response

Bug

Hey,

I was trying to use the API to inference images in my notebook, but when i try to print my 'response' variable i have this : <Response [400]>.
I guess the API don't work, how can i handle this ?

Kind regards !

Environment

No response

Minimal Reproducible Example

No response

Additional

No response

[github-actions]

👋 Hello @CemHaligur, thank you for raising an issue about Ultralytics HUB 🚀! Please visit our HUB Docs to learn more, and see our ⭐️ HUB Guidelines to quickly get started uploading datasets and training YOLO models.

If this is a 🐛 Bug Report, please provide screenshots and steps to recreate your problem to help us get started working on a fix.

If this is a ❓ Question, please provide as much information as possible, including dataset, model, environment details etc. so that we might provide the most helpful response.

We try to respond to all issues as promptly as possible. Thank you for your patience!

[kalenmike]

Hi @CemHaligur, the response suggests that your request is invalid. Can you give us an example of a request you are sending?

You can also try printing the json of the response to see if there is any more information:

print(response.json())

[CemHaligur]

This is an example of the code :

url = "https://api.ultralytics.com/v1/predict/XXXX"

headers = {

"Content-Type": "multipart/form-data",

"x-api-key": "XXX"

}

files = {

"file": ("image", open("./data/Raw/V-00488_2023_02_03_09-23-51.jpeg", "rb"), "image/jpeg")

}

data = {

"size": 640,

"confidence": 0.20,

"iou": 0.45

}

response = requests.post(url, headers=headers, files=files, data=data)

And i try to print the json file i have this :

{'data': {},
'message': 'An image file or URL must be submitted.',
'success': False}

but i have the right path to my file

[kalenmike]

@CemHaligur There is an issue with how you are passing the files, I can see the same issue on our example code on HUB. We will get this updated. For now you can follow the docs.

The change needed is here:

files = {
"image": open("./data/Raw/V-00488_2023_02_03_09-23-51.jpeg", "rb")
}

[ape75]

Hi.
I have also tried to use the API with some images from my computer and I keep getting the same error as @CemHaligur. I have tried various ways to send the image and also noticed some differences between HUB example code and documentation.

API returns the 400 - Bad Request response status code :
{'data': {}, 'message': 'An image file or URL must be submitted.', 'success': False}

Here is my code:

`import requests

url = "https://api.ultralytics.com/v1/predict/XXXXXXXXXXXXXXXXX"
headers = {
"Content-Type": "multipart/form-data",
"x-api-key": "*********************************************************"
}
files = {
"image": open("image2.jpg", "rb")
}
data = {
"size": 640,
"confidence": 0.25,
"iou": 0.45
}
response = requests.post(url, headers=headers, files=files, data=data)

print(response.json())`

[CemHaligur]

Same issue with this doc :

import requests

API URL, use actual MODEL_ID

url = f"https://api.ultralytics.com/v1/predict/XXXXXXXXXX"

Headers, use actual API_KEY

headers = {"x-api-key": "XXXXX"}

Inference arguments (optional)

data = {"size": 640, "confidence": 0.20, "iou": 0.45}

Load image and send request

with open("./data/Raw/V-00476_2023_02_03_09-50-04.jpeg", "rb") as image_file:
files = {"image": image_file}
response = requests.post(url, headers=headers, files=files, data=data)

print(response.json())

-> {'data': {}, 'message': 'Unknown', 'success': False}

[glenn-jocher]

@CemHaligur and @CemHaligur, there seems to be an issue when using specific versions of requests package, which prevents multipart/form-data from being set up correctly. Please make sure you are using requests version 2.22.0 or newer by running !pip install --upgrade requests in your notebook.

If you still face the same error after upgrading the requests package, please try sending the image as bytes instead of a file, as in the following example:

import requests

url = "https://api.ultralytics.com/v1/predict/XXXXXXXXX"
headers = {
    "Content-Type": "multipart/form-data",
    "x-api-key": "XXXXXXXXXXXXXXXXXXXXX",
}
with open("image.jpg", "rb") as f:
    image_data = f.read()
data = {"size": 640, "confidence": 0.25, "iou": 0.45}
response = requests.post(
    url, headers=headers, data=data, files={"image": ("image.jpg", image_data)}
)
print(response.json())

Please let us know if this helps!

[ape75]

Hi.
I updated the requests package and tried also your example code, but the problem still remains the same.

{'data': {}, 'message': 'An image file or URL must be submitted.', 'success': False}

[kalenmike]

@CemHaligur We found an issue that was causing the Unknown response when not setting the normalize option. Your last example should work now since the fix.

@ape75 Do not set the content type header but allow requests to set it so that a boundary can be set.

import requests

url = "https://api.ultralytics.com/v1/predict/XXXXXXXXX"

headers = {
	"x-api-key": "XXXXXXXXXXXXXXXXXXXXX",
}

data = {
	"size": 640,
	"confidence": 0.25,
	"iou": 0.45
}

files={
	"image": open("image2.jpg", "rb")
}

response = requests.post(url, headers=headers, files=files, data=data)

print(response.json())

[ape75]

Now it is working fine. 👍
Thank you for the quick response.

[glenn-jocher]

@ape75 you're welcome! We're glad to hear that the issue has been resolved. If you have any other questions or concerns, please don't hesitate to reach out to us. We're always here to help. Thank you for using Ultralytics HUB!