Skip to main content

DESCRIPTION

Project description

ICDR

Contrastive Data Retrieval with Inverted Indexes

Efficient Approximate/Precise retrieval of similar documents for fine-tuning language models. The library can be used to quickly create contrastive pairs/triplets from large document collections.

ICDR builds an inverted index structure and several fast look-up tables with the aim of retrieving similar texts from a corpus. The library is ideal for efficient entity matching, entity resolution, record linkage, and deduplication applications in the NLP realm. ICDR allows for very fast retrieval of similar, positive (i.e. matching), and negative (i.e. non-matching) text samples which can be used either directly, or to fine-tune LLMs and other models.

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

icdr-0.0.11.tar.gz (997.2 kB view details)

Uploaded Source

Built Distribution

If you're not sure about the file name format, learn more about wheel file names.

icdr-0.0.11-py3-none-any.whl (1.0 MB view details)

Uploaded Python 3

File details

Details for the file icdr-0.0.11.tar.gz.

File metadata

  • Download URL: icdr-0.0.11.tar.gz
  • Upload date:
  • Size: 997.2 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/6.2.0 CPython/3.12.2

File hashes

Hashes for icdr-0.0.11.tar.gz
Algorithm Hash digest
SHA256 5612cca1de8df1e4ec16073490f2885b8fae48718f02bba301dff57214b6b83c
MD5 ae706a82c0c091c4de0dc85681bc8389
BLAKE2b-256 30664263dbc9a753fe0d1375a7963164afa68a0a97108f7358ea61af70c7166a

See more details on using hashes here.

File details

Details for the file icdr-0.0.11-py3-none-any.whl.

File metadata

  • Download URL: icdr-0.0.11-py3-none-any.whl
  • Upload date:
  • Size: 1.0 MB
  • Tags: Python 3
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/6.2.0 CPython/3.12.2

File hashes

Hashes for icdr-0.0.11-py3-none-any.whl
Algorithm Hash digest
SHA256 04f2c6cbdfad670870179ee8d828736e8cf91535c5f72b17a0936d8842c0e1d6
MD5 1bd4c0257fd4abe0772352401d467db8
BLAKE2b-256 b6f145ab3a5e81be3e9aeafb314b51983992a3c93a448e8d29c981a0c576905f

See more details on using hashes here.

Supported by

AWS Cloud computing and Security Sponsor Datadog Monitoring Depot Continuous Integration Fastly CDN Google Download Analytics Pingdom Monitoring Sentry Error logging StatusPage Status page