AI4BharatIndicBERT
IndicBERT v2 Back-Translation TLM
This 278M-parameter IndicBERT v2 release uses a distinct training configuration: Translation-language modeling uses IndicCorp v2 text translated into English with IndicTrans.
The objective distinguishes this checkpoint
The model card links separate weights for the MLM, Samanantar, back-translation and shared-script experiments. These are training variants, not different service tiers.IndicBERT v2 Back-Translation TLM official model card
Control the downstream comparison
Practical guidance. Compare synthetic parallel data with the Samanantar variant. Fluency of a generated translation does not establish the quality of the representation it teaches.
Specifications & access
| Detail | Value and conditions |
|---|---|
| Access | Downloadable |
| Model identifiers | IndicBERTv2-MLM-Back-TLM |
| License | MIT.IndicBERT v2 Back-Translation TLM official model card |
| Evaluation | IndicXTREME; the release supports downstream classification, entity extraction, question answering and retrieval research.IndicBERT v2 Back-Translation TLM official model card |
Related models
- IndicBERT v2 Masked Language Model
AI4Bharat ยท IndicBERT
Multilingual IndicBERT v2 encoder trained with the masked-language-model objective.
Launch date not confirmedDownloadable