Skip to content

NVIDIA NeMo's stt_en_fastconformer_ctc_large finetuned on open-source telugu data for Automatic Speech Recognition

License

Notifications You must be signed in to change notification settings

bharathraj-v/fastconformer-ctc-telugu

Repository files navigation

ASR using NVIDIA NeMo for Telugu

NVIDIA NeMo stt_en_fastconformer_ctc_large finetuned on open-source Telugu data.

Data sources:

(plus white noise, background noise, pitch-shifted augmentations)

Performance: 0.21 CER on Google FLEURS Test data

You can find the model and the inference interface at https://huggingface.co/spaces/bharathraj-v/fastconformer_ctc_telugu

About

NVIDIA NeMo's stt_en_fastconformer_ctc_large finetuned on open-source telugu data for Automatic Speech Recognition

Topics

Resources

License

Stars

Watchers

Forks