diff --git a/ffcv/tokenizer_example.py b/ffcv/tokenizer_example.py
new file mode 100644
index 00000000..08a93177
--- /dev/null
+++ b/ffcv/tokenizer_example.py
@@ -0,0 +1,19 @@
+from transformers import AutoTokenizer
+
+# Load a pre-trained tokenizer (BERT in this case)
+tokenizer = AutoTokenizer.from_pretrained("bert-base-uncased")
+
+# Sample input text
+text = ["Hello, how are you?", "I am learning NLP with Hugging Face!"]
+
+# Tokenize text with padding
+tokenized_output = tokenizer(
+    text,
+    padding=True,  # Apply padding
+    truncation=True,  # Truncate if needed
+    max_length=16,  # Set max length
+    return_tensors="pt"  # Return PyTorch tensors
+)
+
+# Print tokenized output
+print(tokenized_output)