Update handler.py

handler.py

@@ -4,7 +4,7 @@ from typing import Dict, List, Any
 
 class EndpointHandler():
     def __init__(self, path=""):
-      self.quant = quantization_config = BitsAndBytesConfig(load_in_4bit=True, bnb_4bit_compute_dtype=torch.float16)
+      self.quant = quantization_config = BitsAndBytesConfig(load_in_4bit=True, bnb_4bit_compute_dtype=torch.float16, bnb_4bit_quant_type="nf4", bnb_4bit_use_double_quant=True)
       self.model = AutoModelForCausalLM.from_pretrained(path, device_map="auto", quantization_config=self.quant, trust_remote_code=True)
       self.tokenizer = AutoTokenizer.from_pretrained(path)
       self.pipeline = pipeline("text-generation", model=self.model, tokenizer=self.tokenizer)