How paperswithcode/paperswithcode-data works
This repository is a daily regenerated collection of Papers with Code datasets, not an application runtime. Its documented product is downloadable data dumps covering papers, paper-to-code links, evaluation tables, methods, and datasets. The README identifies the datasets and their format, but provides no implementation source or internal processing relationships, so the graph is limited to the documented data products and their external publisher. The datasets JSON is noted as compatible with the sota-extractor format; no loader implementation is present in this repository tree.
Main entry points: README.md Papers and abstracts
Components 7
Dataset products
- Papers and abstracts
README.md - Paper-code links
README.md - Evaluation tables
README.md - Methods
README.md - Datasets JSON
README.md
Other components
- Data user
- Hugging Face datasets
Connections 10
- Hugging Face datasets → Papers and abstracts: hosts
README.md - Hugging Face datasets → Paper-code links: hosts
README.md - Hugging Face datasets → Evaluation tables: hosts
README.md - Hugging Face datasets → Methods: hosts
README.md - Hugging Face datasets → Datasets JSON: hosts
README.md - Data user → Papers and abstracts: downloads
README.md - Data user → Paper-code links: downloads
README.md - Data user → Evaluation tables: downloads
README.md - Data user → Methods: downloads
README.md - Data user → Datasets JSON: downloads
README.md