Update handler.py

handler.py

@@ -6,12 +6,12 @@ class EndpointHandler():
         # init
         # load the model
         tokenizer = AutoTokenizer.from_pretrained("verseAI/vai-GPT-NeoXT-Chat-Base-20B")
-        # model = AutoModelForCausalLM.from_pretrained("verseAI/vai-GPT-NeoXT-Chat-Base-20B", device_map="auto", load_in_8bit=True)        
+        model = AutoModelForCausalLM.from_pretrained("verseAI/vai-GPT-NeoXT-Chat-Base-20B", device_map="auto", load_in_8bit=True)        
         # THROWS ERROR model = AutoModelForQuestionAnswering.from_pretrained("verseAI/vai-GPT-NeoXT-Chat-Base-20B", device_map="auto", load_in_8bit=True)        
-        model = AutoModel.from_pretrained("verseAI/vai-GPT-NeoXT-Chat-Base-20B", device_map="auto", load_in_8bit=True)
+        # model = AutoModel.from_pretrained("verseAI/vai-GPT-NeoXT-Chat-Base-20B", device_map="auto", load_in_8bit=True)
         # create inference pipeline
-        # self.pipeline = pipeline("text-generation", model=model, tokenizer=tokenizer)
-        self.pipeline = pipeline("question-answering", model=model, tokenizer=tokenizer)
+        self.pipeline = pipeline("text-generation", model=model, tokenizer=tokenizer)
+        #self.pipeline = pipeline("question-answering", model=model, tokenizer=tokenizer)
 
     def __call__(self, data: Dict[str, Any]) -> List[List[Dict[str, float]]]:
         """