A tiny CLI wrapper around lmppl for Pre-Trained Language Models Perplexity Calculation for CSV files. Please edit your config.ini file according to your needs. For larger CSV files, it's highly recommended to have a NVIDIA GPU hardware unless processing going to be significantly slower due to the increased computational requirements.
Prompt examples are taken from fka/awesome-chatgpt-prompts
Please modify the config.ini file according to your needs:
[Config]
csv_file = /content/lmppl-cli-csv-wrapper/data/prompts.csv
csv_sentence_header = prompt
model_name = gpt2
batch_size = 8
delimiter = ,
Screenshot of example usage: