Send a prompt with an image to the Ollama server for vision models like Llama 3.2-vision (no external packages needed)

ollama_vision.py

import argparse
import base64
import requests
import json
import sys
from pathlib import Path

def image_to_base64(image_path):
    """Convert image to base64 string"""
    try:
        with open(image_path, "rb") as image_file:
            return base64.b64encode(image_file.read()).decode('utf-8')
    except Exception as e:
        print(f"Error converting image to base64: {e}")
        sys.exit(1)

def make_ollama_request(base64_image, prompt, model, api_url):
    """Make API request to Ollama"""
    headers = {
        'Content-Type': 'application/json'
    }

    payload = {
        "model": model,
        "prompt": prompt,
        "images": [base64_image],
        "stream": False
    }
    # print("Payload:")
    # print(json.dumps(payload, indent=2))

    try:
        response = requests.post(api_url, headers=headers, json=payload)
        response.raise_for_status()
        return response.json()
    except requests.exceptions.RequestException as e:
        print(f"Error making API request: {e}")
        sys.exit(1)

def main():
    parser = argparse.ArgumentParser(description='Convert image to base64 and make Ollama API request')
    parser.add_argument('--image', required=True, help='Path to the image file')
    parser.add_argument('--prompt', required=True, help='Prompt for the model')
    parser.add_argument('--model', default='llama3.2-vision', help='Model to use (default: llama3.2-vision)')
    parser.add_argument('--api-url', default='http://localhost:11434/api/generate',
                        help='Ollama API URL (default: http://localhost:11434/api/generate)')

    args = parser.parse_args()

    # Validate image path
    image_path = Path(args.image)
    if not image_path.exists():
        print(f"Error: Image file '{args.image}' does not exist")
        sys.exit(1)

    # Convert image to base64
    print("Converting image to base64...")
    base64_image = image_to_base64(args.image)

    # Make API request
    print("Making API request to Ollama...")
    response = make_ollama_request(base64_image, args.prompt, args.model, args.api_url)

    # Print response
    print("\nResponse from Ollama:")
    print(json.dumps(response, indent=2))

    # Print response times
    if 'total_duration' in response:
        total_duration = response['total_duration'] / 1e9  # Convert to seconds
        load_duration = response.get('load_duration', 0) / 1e9
        prompt_eval_duration = response.get('prompt_eval_duration', 0) / 1e9
        eval_duration = response.get('eval_duration', 0) / 1e9

        print("\nTiming Information:")
        print(f"Total Duration: {total_duration:.2f} seconds")
        print(f"Load Duration: {load_duration:.2f} seconds")
        print(f"Prompt Evaluation Duration: {prompt_eval_duration:.2f} seconds")
        print(f"Evaluation Duration: {eval_duration:.2f} seconds")

if __name__ == "__main__":
    main()