From db6ed71fa3aed299de4344c9d56a55d44699966f Mon Sep 17 00:00:00 2001 From: Ben Sima Date: Fri, 23 May 2025 20:35:29 -0400 Subject: add counttokens script --- counttokens | 30 ++++++++++++++++++++++++++++++ 1 file changed, 30 insertions(+) create mode 100755 counttokens (limited to 'counttokens') diff --git a/counttokens b/counttokens new file mode 100755 index 0000000..e1c7442 --- /dev/null +++ b/counttokens @@ -0,0 +1,30 @@ +#! /usr/bin/env nix-shell +#! nix-shell -i python3 -p 'python3.withPackages (ps: with ps; [ transformers argparse ])' + +import sys +import argparse +from transformers import AutoTokenizer + +def main(): + parser = argparse.ArgumentParser( + description="Tokenize input text using specified model tokenizer." + ) + parser.add_argument( + '-m', '--model', + type=str, + default='gpt2', + help='Model name for tokenizer, e.g., "gpt2", "distilgpt2". Default is "gpt2".' + ) + args = parser.parse_args() + + # Read text from standard input + text = sys.stdin.read().strip() + + # Load tokenizer for the specified model + tokenizer = AutoTokenizer.from_pretrained(args.model) + tokens = tokenizer.encode(text) + + print(len(tokens)) + +if __name__ == "__main__": + main() -- cgit v1.2.3