fix tokenizer cache file name bug (#36)

rapidsai · Dec 8, 2023 · 1f7d20a · 1f7d20a
1 parent 6749ca3
commit 1f7d20a
Showing 1 changed file with 1 addition and 1 deletion.
diff --git a/crossfit/op/tokenize.py b/crossfit/op/tokenize.py
@@ -159,7 +159,7 @@ def from_pretrained(cls, name, cache_dir=None):
 
             # Save vocabulary to disk
             # `save_vocabulary()` automatically appends `-vocab.txt` suffix.
-            vocab_path = tokenizer.save_vocabulary(cache_dir, "{tokenizer_class}")[0]
+            vocab_path = tokenizer.save_vocabulary(cache_dir, f"{tokenizer_class}")[0]
 
             # Hash the vocabulary and save it
             hash_vocab(vocab_path, hashed_vocab_path)