Vietnamese Court Judgments with PDF, JSON, and Embeddings

Congbobanan Toaan Gov Vn is a document-level mirror of the Vietnamese Cổng công bản án portal from the Supreme People's Court of Vietnam. Each case is provided as a raw PDF, parsed markdown, a structured JSON record with metadata and NER, a 2,048-dimension dense embedding, and a 2-D projection with cluster IDs. The corpus was produced end-to-end by author tmquan and was last updated on HuggingFace in May 2026.

Use Cases

• Train legal document classification models based on the raw PDF and markdown text.
• Develop named entity recognition (NER) systems for Vietnamese legal text based on the structured JSON records.
• Perform semantic search and clustering of legal cases using the provided 2,048-dimension dense embeddings.
• Analyze patterns in statute references extracted from the court judgments.
• Create visualizations of the legal corpus using the provided 2-D projections (PCA/t-SNE/UMAP).

Strengths

• Provides multiple representations of each case, including raw PDFs, parsed markdown, and structured JSON.
• Includes machine learning-ready features like 2,048-dimension dense embeddings and 2-D projections with cluster IDs.
• Contains structured metadata, generic NER, and statute references within the JSON records.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment.
• Freshness should be verified as the last update timestamp is from the future (2026-05-20).

Provenance

Source

Supreme People's Court of Vietnam (Tòa án nhân dân tối cao) via congbobanan.toaan.gov.vn.

Collection Method

Document-level mirror produced end-to-end by the dataset author.

Freshness

Last updated 2026-05-20 06:24:31.

Geography

Vietnam