Model reference · open weights

indicconformer_stt_ta_hybrid_ctc_rnnt_large

indicconformer_stt_ta_hybrid_ctc_rnnt_large is an open-weight audio or speech model from ai4bharat.

• IndicConformer is a large-scale automatic speech recognition model developed by ai4bharat for transcribing Tamil audio.
• The model features a conformer encoder with 120 million parameters and utilizes a hybrid CTC-RNNT decoder to process 16 kHz mono-channel audio inputs.
• It is distributed under the MIT license and requires the AI4Bharat NeMo framework for loading and inference.

Summary of the ai4bharat/indicconformer_stt_ta_hybrid_ctc_rnnt_large model card, 2026-10-01

Hugging Face ↗