FineTome-100k/README.md

dataset_info	configs
features splits download_size dataset_size name list conversations name dtype from string name dtype value string name dtype source string name dtype score float64 name num_bytes num_examples train 239650960.7474458 100000 116531415 239650960.7474458	config_name data_files default split path train data/train-*

FineTome-100k

The FineTome dataset is a subset of arcee-ai/The-Tome (without arcee-ai/qwen2-72b-magpie-en), re-filtered using HuggingFaceFW/fineweb-edu-classifier.

It was made for my article "Fine-tune Llama 3.1 Ultra-Efficiently with Unsloth".