salesken/Hindi-FastConformer-Streaming-ASR

Salesken-Streaming-FastConformer-Hindi-ASR

This model is a fine-tuned version of STT En FastConformer Hybrid Large Streaming from NVIDIA's NeMo framework, specifically optimized for Hindi automatic speech recognition (ASR). The model has been fine-tuned on the Aibharath Hindi dataset to enhance transcription accuracy for Hindi speech, including real-time streaming applications.

Model Details

Model Description

• Model type: Hybrid ASR model (CTC + Attention)
• Model Sample Rate: 16000 hz
• Language(s) (NLP): Hindi (hi)
• License: Apache-2.0
• Finetuned from model: nvidia/stt_en_fastconformer_hybrid_large_streaming_multi

How to Get Started with the Model

You can load and use the model with the following code:

import nemo.collections.asr as nemo_asr
asr_model = nemo_asr.models.EncDecHybridRNNTCTCBPEModel.from_pretrained(model_name="salesken/Hindi-FastConformer-Streaming-ASR")

# Optional: change the default latency. Default latency is 1040ms. Supported latencies: {0: 0ms, 1: 80ms, 16: 480ms, 33: 1040ms}.
# Note: These are the worst latency and average latency would be half of these numbers.
asr_model.encoder.set_default_att_context_size([70,13]) 

#Optional: change the default decoder. Default decoder is Transducer (RNNT). Supported decoders: {ctc, rnnt}.
asr_model.change_decoding_strategy(decoder_type='rnnt')

asr_model.transcribe(['2086-149220-0033.wav'])