Upload app.py with huggingface_hub

app.py

@@ -0,0 +1,26 @@
+import gradio as gr
+from llama_cpp import Llama
+import os
+
+# تحميل النموذج
+print("Loading model...")
+llm = Llama(
+    model_path="gemma-4-medical.gguf",
+    n_ctx=2048,
+    n_threads=2
+)
+
+def generate_response(message, history):
+    prompt = f"USER: {message}
+ASSISTANT: "
+    response = llm(prompt, max_tokens=512, stop=["USER:"], echo=False)
+    return response["choices"][0]["text"]
+
+demo = gr.ChatInterface(
+    fn=generate_response,
+    title="Gemma 4 Medical GGUF",
+    description="Powered by llama-cpp-python inside HF Spaces"
+)
+
+if __name__ == "__main__":
+    demo.launch()