Update app.py

app.py

@@ -32,8 +32,9 @@ HF_TOKEN = os.getenv("HF_TOKEN")  # Load from Render environment variable
 
 device = "cuda" if torch.cuda.is_available() else "cpu"
 
-tokenizer = LlamaTokenizerFast.from_pretrained(REPO_NAME, token=HF_TOKEN)
+#tokenizer = LlamaTokenizerFast.from_pretrained(REPO_NAME, token=HF_TOKEN)
 #tokenizer = LlamaTokenizer.from_pretrained(REPO_NAME, use_auth_token=HF_TOKEN)
+tokenizer = AutoTokenizer.from_pretrained("microsoft/Phi-3-mini-4k-instruct")
 
 model = AutoModelForCausalLM.from_pretrained(
     REPO_NAME,