Try using path...

handler.py

@@ -2,6 +2,8 @@ from typing import Dict, List, Any
 from transformers import AutoTokenizer, AutoModel
 from optimum.pipelines import pipeline
 from optimum.onnxruntime import ORTModelForFeatureExtraction
+from pathlib import Path
+
 
 import torch
 
@@ -16,7 +18,7 @@ class EndpointHandler():
         # self.device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
         task = "feature-extraction"
         tokenizer = AutoTokenizer.from_pretrained('sentence-transformers/msmarco-MiniLM-L-6-v3')
-        model_regular = ORTModelForFeatureExtraction.from_pretrained("onnx", file_name="model.onnx", from_transformers=False)
+        model_regular = ORTModelForFeatureExtraction.from_pretrained(Path("onnx"), file_name="model.onnx", from_transformers=False)
         self.onnx_extractor = pipeline(task, model=model_regular, tokenizer=tokenizer)
         # self.model.to(self.device)
         # print("model will run on ", self.device)