use gpt-large and optimum package for faster cpu infernce

app.py

@@ -5,7 +5,7 @@ import numpy as np
 from transformers_cfg.grammar_utils import IncrementalGrammarConstraint
 from transformers_cfg.generation.logits_process import GrammarConstrainedLogitsProcessor
 
-MODEL_NAME = "gpt2"
+MODEL_NAME = "gpt2-large"
 
 if __name__ == "__main__":
     # Define your model and your tokenizer
@@ -14,6 +14,7 @@ if __name__ == "__main__":
     if tokenizer.pad_token_id is None:
         tokenizer.pad_token_id = tokenizer.eos_token_id
         model.config.pad_token_id = model.config.eos_token_id
+    model.to_bettertransformer()
 
     # Define your color-coding labels; if prob > x, then label = y; Sorted in descending probability order!
     probs_to_label = [

requirements.txt

@@ -1,3 +1,4 @@
 torch
+optimum
 transformers>=4.26
 transformers-cfg==0.2.0