Try to fix streaming

app.py

@@ -2,30 +2,25 @@ import os
 import gradio as gr
 from huggingface_hub import InferenceClient
 
+api_key = os.getenv("HF_API_KEY")
+client = InferenceClient(api_key=api_key)
+
 def analyze(project_data, question):
     try:
-        api_key = os.getenv("HF_API_KEY")
-        client = InferenceClient(
-            model="Qwen/Qwen2.5-72B-Instruct",
-            token=api_key
-        )
-    
         prompt = f"Analyze this project: {project_data}\n\nQuestion: {question}"
     
-        response_generator = client.text_generation(
-            prompt,
-            max_new_tokens=512,
-            temperature=0.7,
-            top_p=0.95,
-            repetition_penalty=1.1,
-            do_sample=True,
+        response = client.chat.completions.create(
+            model="Qwen/Qwen2.5-72B-Instruct",
+            messages=messages,
+            max_tokens=1000,
             stream=True
         )
 
-        response_text = ""
-        for token in response_generator:
-            response_text += token.text
-            yield response_text
+        answer = ""
+        
+        for chunk in response:
+            answer += chunk['choices'][0]['delta']['content']
+            yield answer
         
     except Exception as e:
         print(f"Error details: {str(e)}")