Upload tokenizer.py

tokenizer.py

@@ -0,0 +1,7 @@
+import streamlit as st
+from transformers import DistilBertTokenizerFast
+
+
+@st.cache(allow_output_mutation=True)
+def get_tokenizer(num_labels=153):
+    return DistilBertTokenizerFast.from_pretrained('distilbert-base-uncased', num_labels=num_labels)