set cache to false for examples

app.py

@@ -18,8 +18,6 @@ This Space demonstrates the dialogue models for Nigerian Pidgin, an African lang
 🔎 For more about visit [our homepage](https://www.masakhane.io/).
 
 """
-#🔨 Looking for an even more powerful model? Check out the large [**70B** model demo](https://huggingface.co/spaces/ysharma/Explore_llamav2_with_TGI).
-#🐇 For a smaller model that you can run on many GPUs, check our [7B model demo](https://huggingface.co/spaces/huggingface-projects/llama-2-7b-chat).
 
 
 if not torch.cuda.is_available():
@@ -149,7 +147,7 @@ with gr.Blocks(css='style.css') as demo:
         inputs=textbox,
         outputs=[textbox, chatbot],
         fn=process_example,
-        cache_examples=True,
+        cache_examples=False,
     )
 
     #gr.Markdown(LICENSE)

model.py

@@ -55,7 +55,7 @@ def run(message: str,
     inputs = tokenizer([prompt], return_tensors='pt', add_special_tokens=False)
 
     streamer = TextIteratorStreamer(tokenizer,
-                                    timeout=60.,
+                                    timeout=60,
                                     skip_prompt=True,
                                     skip_special_tokens=True)
     generate_kwargs = dict(