Update inference.py

inference.py

@@ -22,7 +22,7 @@ login(token = token)
 
 if PREFERED_MODEL == "pretrained":
     #print("Using pretrained model")
-    model_id = "meta-llama/Meta-Llama-3-8B"
+    model_id = "mattshumer/Llama-3-8B-16K"
     print("Loading model...")
     model = AutoModelForCausalLM.from_pretrained(model_id, device_map = "cuda", load_in_8bit = True)
     print("loading tokenizer...")