Dataset Viewer
Auto-converted to Parquet Duplicate
The dataset viewer is not available for this split.
Parquet error: Scan size limit exceeded: attempted to read 416920696 bytes, limit is 300000000 bytes Make sure that 1. the Parquet files contain a page index to enable random access without loading entire row groups2. otherwise use smaller row-group sizes when serializing the Parquet files
Error code:   TooBigContentError

Need help to make the dataset viewer work? Make sure to review how to configure the dataset viewer, and open a discussion for direct support.

YAML Metadata Warning:empty or missing yaml metadata in repo card

Check out the documentation for more information.

AgriSLM Deduped Pre-Training Corpus

High-quality, deduplicated, India-focused corpus for 300M parameter agriculture SLM.

Built from agri-slm-curated with MD5 near-deduplication (first 500 words).

Stats

  • Input: 1,719,470 chunks from agri-slm-curated
  • Duplicates removed: 372,047 (21.6%)
  • Output: 1,347,423 unique chunks
  • Estimated tokens: ~3.07B
  • Batches: 27

Pipeline

agri-slm-corpus (×4 repos) → agri-slm-final (chunked) → agri-slm-curated (deep clean) → agri-slm-deduped

Downloads last month
51