Update app.py

app.py

@@ -2,6 +2,10 @@ import gradio as gr
 from PIL import Image
 import io
 import base64
+from huggingface_hub import InferenceClient
+
+# Initialize the Hugging Face Inference Client
+client = InferenceClient("microsoft/llava-med-7b-delta")
 
 # Function to encode image as base64
 def image_to_base64(image):
@@ -11,7 +15,15 @@ def image_to_base64(image):
     return img_str
 
 # Function to interact with LLAVA model
-def chat_with_llava(message, history, system_message, max_tokens, temperature, top_p, image=None):
+def respond(
+    message,
+    history: list[tuple[str, str]],
+    system_message,
+    max_tokens,
+    temperature,
+    top_p,
+    image=None
+):
     messages = [{"role": "system", "content": system_message}]
 
     for val in history:
@@ -20,34 +32,43 @@ def chat_with_llava(message, history, system_message, max_tokens, temperature, t
         if val[1]:
             messages.append({"role": "assistant", "content": val[1]})
 
+    messages.append({"role": "user", "content": message})
+
     if image:
         # Convert image to base64
         image_b64 = image_to_base64(image)
         messages.append({"role": "user", "content": "Image uploaded", "image": image_b64})
 
-    messages.append({"role": "user", "content": message})
-
-    # Simulate a response (replace with your logic)
+    # Call Hugging Face model for response
     response = ""
-    for i in range(3):  # Simulate 3 responses
-        response += f"Response {i+1}\n"
+    for message in client.chat_completion(
+        messages,
+        max_tokens=max_tokens,
+        stream=True,
+        temperature=temperature,
+        top_p=top_p,
+    ):
+        token = message.choices[0].delta.content
 
-    return response
+        response += token
+        yield response
 
 # Create a Gradio interface
 demo = gr.Interface(
-    fn=chat_with_llava,
+    fn=respond,
     inputs=[
-        gr.Textbox(label="Message"),
-        gr.Image(label="Upload Medical Image"),
-        gr.Textbox(label="System message", default="You are a friendly Chatbot."),
-        gr.Slider(minimum=1, maximum=2048, default=512, step=1, label="Max new tokens"),
-        gr.Slider(minimum=0.1, maximum=4.0, default=0.7, step=0.1, label="Temperature"),
-        gr.Slider(minimum=0.1, maximum=1.0, default=0.95, step=0.05, label="Top-p (nucleus sampling)")
+        gr.inputs.Textbox(label="Message"),
+        gr.inputs.Image(label="Upload Medical Image", type="pil")
     ],
     outputs=gr.outputs.Textbox(label="Response", placeholder="Model response will appear here..."),
     title="LLAVA Model - Medical Image and Question",
-    description="Upload a medical image and ask a specific question about the image for a medical description."
+    description="Upload a medical image and ask a specific question about the image for a medical description.",
+    additional_inputs=[
+        gr.Textbox(label="System message", default="You are a friendly Chatbot."),
+        gr.Slider(minimum=1, maximum=2048, value=512, step=1, label="Max new tokens"),
+        gr.Slider(minimum=0.1, maximum=4.0, value=0.7, step=0.1, label="Temperature"),
+        gr.Slider(minimum=0.1, maximum=1.0, value=0.95, step=0.05, label="Top-p (nucleus sampling)")
+    ]
 )
 
 # Launch the Gradio interface