Update app.py

app.py

@@ -353,11 +353,10 @@ def process_llm_response(llm_response):
 
 
 
-
+tok, model = build_model()
 
 @spaces.GPU(duration=60)
 def llm_ans(message, history):
-    tok, model = build_model()
     terminators = [tok.eos_token_id, 32007, 32011, 32001, 32000]
     pipe = pipeline(task="text-generation", model=model, tokenizer=tok, eos_token_id=terminators, do_sample=True, max_new_tokens=CFG.max_new_tokens, temperature=CFG.temperature, top_p=CFG.top_p, repetition_penalty=CFG.repetition_penalty)
     llm = HuggingFacePipeline(pipeline=pipe)