Update app.py

app.py

@@ -47,7 +47,7 @@ model = PeftModel.from_pretrained(base_model, "lakshyaag/llama38binstruct_summar
 if torch.cuda.is_available():
     model = model.to("cuda")
 
-tokenizer = AutoTokenizer.from_pretrained("lakshyaag/llama38binstruct_summarize")
+tokenizer = AutoTokenizer.from_pretrained("NousResearch/Meta-Llama-3-8B-Instruct")
 
 tokenizer.pad_token = tokenizer.eos_token
 tokenizer.padding_side = "right"
@@ -100,7 +100,7 @@ async def main(message: cl.Message):
     print(decoded_output)
 
     # return only the generated response (not the prompt) as output
-    response = decoded_output[0].split("<|end_of_text|>")[-1]
+    response = decoded_output[0].split("<|end_header_id|>")[-1]
 
     msg = cl.Message(content=response)