AI4BharatIndicLID

IndicLID BERT

IndicLID BERT is one of three separately trained components in IndicLID.

This component has a specific role

The slower BERT component is fine-tuned on Romanized text from IndicBERT. The full system covers native and Romanized scripts.IndicLID components, training and license

Review short and mixed-language inputs

Practical guidance. A single label may not capture a sentence that switches languages. Use longer text samples where the application permits, and inspect uncertain predictions.

Specifications & access

IndicLID BERT specifications
DetailValue and conditions
AccessDownloadable
Model identifiersIndicLID-BERT
LicenseMIT.IndicLID components, training and license

Related models

  • IndicLID fastText component for identifying languages written in their native scripts.

    Launch date not confirmedDownloadable
  • IndicLID fastText component for language identification in Romanized text.

    Launch date not confirmedDownloadable