replaced the terminator values to origin and replaced the transformer to instruct version

app.py

@@ -1,18 +1,11 @@
 import torch
-import pandas as pd
-import numpy as np
 import gradio as gr
-import re
 from transformers import AutoTokenizer, AutoModelForCausalLM, TextIteratorStreamer
-import re
-from huggingface_hub import login
 import os
 from threading import Thread
 
 # HF_TOKEN
-TOKEN = os.getenv('HF_AUTH_TOKEN')
-login(token=TOKEN,
-      add_to_git_credential=False)
+HF_AUTH_TOKEN = os.getenv('HF_AUTH_TOKEN')
 
 # Open ai api key
 API_KEY = os.getenv('OPEN_AI_API_KEY')
@@ -25,21 +18,13 @@ DESCRIPTION = '''
 '''
 
 # Place transformers in hardware to prepare for process and generation
-llama_tokenizer = AutoTokenizer.from_pretrained("meta-llama/Meta-Llama-3-8B")
-llama_model = AutoModelForCausalLM.from_pretrained("meta-llama/Meta-Llama-3-8B", token=TOKEN, torch_dtype=torch.float16).to('cuda')
+llama_tokenizer = AutoTokenizer.from_pretrained("meta-llama/Meta-Llama-3-8B-Instruct")
+llama_model = AutoModelForCausalLM.from_pretrained("meta-llama/Meta-Llama-3-8B-Instruct", torch_dtype=torch.float16).to('cuda')
 terminators = [
     llama_tokenizer.eos_token_id,
-    # Remove this line, as an empty string won't convert to a valid token ID
-    # llama_tokenizer.convert_tokens_to_ids("")
+    llama_tokenizer.convert_tokens_to_ids("<|eot_id|>")
 ]
 
-# Get special tokens list from the tokenizer
-special_tokens = llama_tokenizer.special_tokens_map
-eos_token = special_tokens.get("eos_token")
-
-print("Default EOS Token:", eos_token)
-
-
 
 # Place just input pass and return generation output
 def llama_generation(input_text: str,