added streamer in the generate_kwargs and placed params in streamer

app.py

@@ -16,7 +16,7 @@ API_KEY = os.getenv('OPEN_AI_API_KEY')
 DESCRIPTION = '''
 <div>
 <h1 style="text-align: center;">Amphisbeana 🐍</h1>
-<p>This uses Llama 3 and GPT-4o as generation, both of these make the final generation. <a href="https://huggingface.co/meta-llama/Meta-Llama-3-8B"><b>Llama3-8b</b></a>and <a href="https://platform.openai.com/docs/models/gpt-4o"><b>GPT-4o</b></a></p>
+<p>This uses Llama 3 and GPT-4o as generation, both of these make the final generation. <a href="https://huggingface.co/meta-llama/Meta-Llama-3-8B"><b>Llama3-8b</b></a> and <a href="https://platform.openai.com/docs/models/gpt-4o"><b>GPT-4o</b></a></p>
 </div>
 '''
 
@@ -54,11 +54,12 @@ def llama_generation(input_text: str,
 
     input_ids = llama_tokenizer.apply_chat_template(conversation, return_tensors='pt').to(llama_model.device)
 
-    streamer = TextIteratorStreamer(llama_tokenizer, skip_prompt=True, timeout=10.0)
+    streamer = TextIteratorStreamer(llama_tokenizer, timeout=10.0, skip_prompt=True, skip_special_tokens=True)
 
     # generation arguments to pass in llm generate() eventually
     generate_kwargs = dict(
         input_ids=input_ids,
+        streamer=streamer,
         max_new_tokens=max_new_tokens,
         do_sample=True,
         temperature=temperature,