llava-cpp-server icon indicating copy to clipboard operation
llava-cpp-server copied to clipboard

API call with python requests

Open kubs0ne opened this issue 1 year ago • 1 comments

Hello, I was able to successfully run the server using the repo. However I am wondering how to perform an API call with python requests library. I was trying to do it like this:

`def _query():
    API_URL = 'http://localhost:8080/completion'
    
    image_path = ...
    user_prompt = ...
    
    with open(image_path, 'rb') as image_file:
        files = {'image_file': image_file}
        data = {'user_prompt': user_prompt}

        response = requests.post(API_URL, files=files, data=data)
        
        if response.status_code == 200:
            response_data = response.json()
            text = response_data['content']
            return text
        else:
            return f"Error: {response.status_code}"

response = _query()
print("Response:", response)`

I always get 404 error.

kubs0ne avatar May 23 '24 11:05 kubs0ne

I haven’t tested in a while but this repo should have examples for both this llava cpp server as well as llama.cpp (which I recommend you switch to because it is being actively developed still): https://github.com/trzy/multimodal-cloud-tests/blob/main/llava-cpp-test.py

trzy avatar May 23 '24 23:05 trzy