Update handler.py

removing increased token limit

handler.py

@@ -11,7 +11,6 @@ class EndpointHandler:
     def __init__(self, path=""):
         # load the model
         tokenizer = AutoTokenizer.from_pretrained(path) 
-        tokenizer..max_seq_len = 8192
         model = AutoModelForCausalLM.from_pretrained(path, device_map="auto", torch_dtype=dtype, trust_remote_code=True)
         model.to('cuda:0')