Update app.py

app.py

@@ -8,7 +8,6 @@ import secrets
 from pathlib import Path
 from pydub import AudioSegment
 
-# Initialize the model and tokenizer
 torch.manual_seed(420)
 tokenizer = AutoTokenizer.from_pretrained("Qwen/Qwen-Audio-Chat", trust_remote_code=True)
 model = AutoModelForCausalLM.from_pretrained("Qwen/Qwen-Audio-Chat", device_map="cuda", trust_remote_code=True).eval()
@@ -45,6 +44,8 @@ def _parse_text(text):
     return text
     
 def predict(_chatbot, task_history, user_input):
+    if task_history is None or not isinstance(task_history, list):
+        task_history = [] 
     print("Predict - Start: task_history =", task_history)
     if not isinstance(task_history, list) or not all(isinstance(item, tuple) and len(item) == 2 for item in task_history):
         print("Error: task_history should be a list of tuples of length 2.")
@@ -107,6 +108,8 @@ def predict(_chatbot, task_history, user_input):
 
 
 def regenerate(_chatbot, task_history):
+    if task_history is None or not isinstance(task_history, list):
+        task_history = [] 
     print("Regenerate - Start: task_history =", task_history) 
     if not task_history:
         return _chatbot
@@ -123,6 +126,8 @@ def regenerate(_chatbot, task_history):
     return predict(_chatbot, task_history)
 
 def add_text(history, task_history, text):
+    if task_history is None or not isinstance(task_history, list):
+        task_history = [] 
     print("Add Text - Before: task_history =", task_history) 
     if not isinstance(task_history, list):
         task_history = []
@@ -132,6 +137,8 @@ def add_text(history, task_history, text):
     return history, task_history
 
 def add_file(history, task_history, file):
+    if task_history is None or not isinstance(task_history, list):
+        task_history = [] 
     print("Add File - Before: task_history =", task_history)
     history.append(((file.name,), None))
     task_history.append(((file.name,), None))
@@ -139,6 +146,8 @@ def add_file(history, task_history, file):
     return history, task_history
 
 def add_mic(history, task_history, file):
+    if task_history is None or not isinstance(task_history, list):
+        task_history = [] 
     print("Add Mic - Before: task_history =", task_history)
     if file is None:
         return history, task_history
@@ -153,6 +162,8 @@ def reset_user_input():
     return gr.update(value="")
 
 def reset_state(task_history):
+    if task_history is None or not isinstance(task_history, list):
+        task_history = [] 
     print("Reset State - Before: task_history =", task_history)  
     task_history = []  
     print("Reset State - After: task_history =", task_history)
@@ -163,11 +174,11 @@ iface = gr.Interface(
     inputs=[
         gr.Audio(label="Audio Input"),
         gr.Textbox(label="Text Query"),
-        gr.State(default=[]) 
+        gr.State() 
     ],
     outputs=[
         "text",  
-        gr.State()  
+        gr.State() 
     ],
     title="Audio-Text Interaction Model",
     description="This model can process an audio input along with a text query and provide a response.",