llama vision

tutorials/meta/docker-compose.yml

@@ -1,15 +1,14 @@
-version: '3.8'
 services:
   ollama:
     volumes:
       - ~/ollama/ollama:/root/.ollama
     container_name: ollama
     #pull_policy: always
     tty: true
-    restart: unless-stopped
+    #restart: unless-stopped
     image: ollama/ollama:0.4.0-rc5
     ports:
-      - 11434:11434
+      - 21434:11434
     deploy:
       resources:
         reservations:

tutorials/meta/llama_vision.py

@@ -0,0 +1,37 @@
+from PIL import Image
+import base64
+import io
+from ollama import Client
+def image_to_base64(image_path):
+    # Open the image file
+    with Image.open(image_path) as img:
+        # Create a BytesIO object to hold the image data
+        buffered = io.BytesIO()
+        # Save the image to the BytesIO object in a specific format (e.g., PNG)
+        img.save(buffered, format="PNG")
+        # Get the byte data from the BytesIO object
+        img_bytes = buffered.getvalue()
+        # Encode the byte data to base64
+        img_base64 = base64.b64encode(img_bytes).decode('utf-8')
+        return img_base64
+
+
+def test_code():
+    image_path = '/home/gaganyatri/Downloads/books-rec-1.jpeg'  # Replace with your image path
+    base64_image = image_to_base64(image_path)
+    client = Client(host='http://localhost:21434')
+    response = client.chat(
+    model="x/llama3.2-vision:latest",
+    messages=[{
+    "role": "user",
+    "content": "Describe this image?",
+    "images": [base64_image]
+    }],
+)
+
+    # Extract the model's response about the image
+    cleaned_text = response['message']['content'].strip()
+    print(f"Model Response: {cleaned_text}")
+    # Example usage
+
+test_code()

tutorials/meta/requirements.txt

@@ -0,0 +1,11 @@
+anyio==4.6.2.post1
+certifi==2024.8.30
+exceptiongroup==1.2.2
+h11==0.14.0
+httpcore==1.0.6
+httpx==0.27.2
+idna==3.10
+ollama==0.3.3
+pillow==11.0.0
+sniffio==1.3.1
+typing_extensions==4.12.2