Upload app.py with huggingface_hub

app.py

@@ -3,25 +3,18 @@ import gradio as gr
 from llama_cpp import Llama
 import os
 
-# تحديد المسار المحلي للملف المرفوع في الـ Space
+# المسار المحلي داخل المساحة
 model_path = "gemma-4-medical.gguf"
 
-try:
-    if os.path.exists(model_path):
-        print("--- Loading local GGUF model ---")
-        llm = Llama(model_path=model_path, n_ctx=2048, n_threads=2)
-        
-        def chat(message, history):
-            output = llm(f"USER: {message}\nASSISTANT: ", max_tokens=512, stop=['USER:'])
-            return output['choices'][0]['text']
-
-        demo = gr.ChatInterface(fn=chat, title='Gemma 4 Medical (Internal)')
-    else:
-        with gr.Blocks() as demo:
-            gr.Markdown(f'### ⚠️ الملف {model_path} غير موجود في المساحة.')
-except Exception as e:
+if os.path.exists(model_path):
+    llm = Llama(model_path=model_path, n_ctx=1024, n_threads=2)
+    def predict(message, history):
+        response = llm(f"USER: {message}\nASSISTANT: ", max_tokens=256, stop=['USER:'])
+        return response['choices'][0]['text']
+    demo = gr.ChatInterface(fn=predict, title='Gemma 4 Medical - Local Edition')
+else:
     with gr.Blocks() as demo:
-        gr.Markdown(f'### ❌ خطأ: {e}')
+        gr.Markdown(f'# Error: Model file {model_path} not found in Space storage.')
 
 if __name__ == '__main__':
     demo.launch()