Changed auto tokenizer to granite tokenizer

main.py

@@ -11,7 +11,7 @@
 
 from fastapi import FastAPI
 from pydantic import BaseModel
-from transformers import AutoModelForCausalLM, AutoTokenizer
+from transformers import AutoModelForCausalLM, AutoTokenizer, GraniteTokenizer
 import uvicorn
 import torch
 # torch.mps.empty_cache()
@@ -62,7 +62,7 @@ def generate(item: Item):
     device = "cuda" if torch.cuda.is_available() else "cpu"
 
     model_path = "ibm-granite/granite-34b-code-instruct-8k"
-    tokenizer = AutoTokenizer.from_pretrained(model_path)
+    tokenizer = GraniteTokenizer.from_pretrained(model_path)
     # drop device_map if running on CPU
     model = AutoModelForCausalLM.from_pretrained(model_path, device_map=device)
     model.eval()