Update handler.py

handler.py

@@ -7,8 +7,8 @@ import base64
 class EndpointHandler:
     def __init__(self, model_dir):
         self.model_id = "zesquirrelnator/moondream2-finetuneV2"
-        self.model = AutoModelForCausalLM.from_pretrained(self.model_id, trust_remote_code=True, ignore_mismatched_sizes=True)
-        self.tokenizer = AutoTokenizer.from_pretrained("vikhyatk/moondream2", trust_remote_code=True, revision="2024-05-08")
+        self.model = AutoModelForCausalLM.from_pretrained(self.model_id, trust_remote_code=True)
+        self.tokenizer = AutoTokenizer.from_pretrained("vikhyatk/moondream2", trust_remote_code=True, revision="2024-05-20")
 
         # Check if CUDA (GPU support) is available and then set the device to GPU or CPU
         self.device = torch.device("cuda" if torch.cuda.is_available() else "cpu")