Changes in app.py

app.py

@@ -38,6 +38,7 @@ if torch.cuda.is_available():
 
 @spaces.GPU
 def generate(
+    tokenizer,
     message: str,
     chat_history: list[tuple[str, str]],
     system_prompt: str,
@@ -62,6 +63,7 @@ def generate(
 
     streamer = TextIteratorStreamer(tokenizer, timeout=10.0, skip_prompt=True, skip_special_tokens=True)
     generate_kwargs = dict(
+        tokenizer,
         inputs,
         streamer=streamer,
         max_new_tokens=max_new_tokens,
@@ -84,6 +86,7 @@ def generate(
 chat_interface = gr.ChatInterface(
     fn=generate,
     additional_inputs=[
+        tokenizer,
         gr.Textbox(label="System prompt", value=DEFAULT_SYSTEM_PROMPT, lines=6),
         gr.Slider(
             label="Max new tokens",