Fix deduplication for high quality papers and tables

Fix bug with deduplication for medium quality tables.

Add more information to README

Switch to downloading remote files

Added custom data loader and added arxiv ids to data

Upload medium quality tables and papers (more lenient filtering and no full texts)

Update README.md

Upload tables_dataset_viewer_only.jsonl

Delete data

Upload README.md with huggingface_hub

Upload data/train-00000-of-00001-c41f83ec70fcae47.parquet with huggingface_hub

Update README to configure dataset viewer

Upload 3 files

initial commit