diff --git a/tokenizer.py b/tokenizer.py
deleted file mode 100644
index 9d04293..0000000
--- a/tokenizer.py
+++ /dev/null
@@ -1,12 +0,0 @@
-from gemini_torch.tokenizer import MultimodalSentencePieceTokenizer
-
-# Example usage
-tokenizer_name = "hf-internal-testing/llama-tokenizer"
-tokenizer = MultimodalSentencePieceTokenizer(tokenizer_name=tokenizer_name)
-
-# Encoding and decoding examples
-encoded_audio = tokenizer.encode("Audio description", modality="audio")
-decoded_audio = tokenizer.decode(encoded_audio)
-
-print("Encoded audio:", encoded_audio)
-print("Decoded audio:", decoded_audio)