Update app.py

app.py

@@ -5,6 +5,7 @@ import threading
 import queue
 import time
 import spaces
+
 # Model configuration
 model_name = "HelpingAI/Dhanishtha-2.0-preview"
 
@@ -31,10 +32,12 @@ def load_model():
 
 class GradioTextStreamer(TextStreamer):
     """Custom TextStreamer for Gradio integration"""
-    def __init__(self, tokenizer, skip_prompt=True, skip_special_tokens=True):
-        super().__init__(tokenizer, skip_prompt, skip_special_tokens)
+    def __init__(self, tokenizer, skip_prompt=True):
+        # TextStreamer only accepts tokenizer and skip_prompt parameters
+        super().__init__(tokenizer, skip_prompt)
         self.text_queue = queue.Queue()
         self.generated_text = ""
+        self.skip_special_tokens = True  # Handle this manually if needed
         
     def on_finalized_text(self, text: str, stream_end: bool = False):
         """Called when text is finalized"""
@@ -56,6 +59,7 @@ class GradioTextStreamer(TextStreamer):
                 self.text_queue.get_nowait()
             except queue.Empty:
                 break
+
 @spaces.GPU()
 def generate_response(message, history, max_tokens, temperature, top_p):
     """Generate streaming response"""
@@ -86,7 +90,7 @@ def generate_response(message, history, max_tokens, temperature, top_p):
     model_inputs = tokenizer([text], return_tensors="pt").to(model.device)
     
     # Create and setup streamer
-    streamer = GradioTextStreamer(tokenizer, skip_prompt=True, skip_special_tokens=True)
+    streamer = GradioTextStreamer(tokenizer, skip_prompt=True)
     streamer.reset()
     
     # Start generation in a separate thread