Parakeet RNNT 0.6B (en)

| |

parakeet-rnnt-0.6b is an ASR model that transcribes speech in lower case English alphabet. This model is jointly developed by NVIDIA NeMo and Suno.ai teams. It is an XL version of FastConformer Transducer [1] (around 600M parameters) model. See the model architecture section and NeMo documentation for complete architecture details.

Licence/Terms of Use

License to use this model is covered by the CC-BY-4.0. By downloading the public and release version of the model, you accept the terms and conditions of the CC-BY-4.0 license.

Discover more from NVIDIA:

For documentation, deployment guides, enterprise-ready APIs, and the latest open models—including Nemotron and other cutting-edge speech, translation, and generative AI—visit the NVIDIA Developer Portal at developer.nvidia.com. Join the community to access tools, support, and resources to accelerate your development with NVIDIA’s NeMo, Riva, NIM, and foundation models.

Explore more from NVIDIA:

What is Nemotron?
NVIDIA Developer Nemotron
NVIDIA Riva Speech
NeMo Documentation

How to Use this Model

The model is available for use in the NeMo toolkit [3], and can be used as a pre-trained checkpoint for inference or for fine-tuning on another dataset.

You can also run Parakeet RNNT with Transformers 🤗 (more below).

1) NeMo usage

To train, fine-tune or play with the model you will need to install NVIDIA NeMo. We recommend you install it after you've installed latest PyTorch version.

pip install nemo_toolkit['all']

Automatically instantiate the model

import nemo.collections.asr as nemo_asr
asr_model = nemo_asr.models.EncDecRNNTBPEModel.from_pretrained(model_name="nvidia/parakeet-rnnt-0.6b")

Transcribing using Python

First, let's get a sample

wget https://dldata-public.s3.us-east-2.amazonaws.com/2086-149220-0033.wav

Then simply do:

output = asr_model.transcribe(['2086-149220-0033.wav'])
print(output[0].text)

Transcribing many audio files

python [NEMO_GIT_FOLDER]/examples/asr/transcribe_speech.py 
 pretrained_name="nvidia/parakeet-rnnt-0.6b" 
 audio_dir="<DIRECTORY CONTAINING AUDIO FILES>"

2) Transformers 🤗 usage

Parakeet RNNT is available in 🤗 Transformers starting from v5.13.0.

pip install "transformers>=5.13.0"

from transformers import pipeline

pipe = pipeline("automatic-speech-recognition", model="nvidia/parakeet-rnnt-0.6b")
out = pipe("https://huggingface.co/datasets/hf-internal-testing/dummy-audio-samples/resolve/main/bcn_weather.mp3")
print(out)

from transformers import AutoModelForRNNT, AutoProcessor
from datasets import load_dataset, Audio

num_samples = 3

model_id = "nvidia/parakeet-rnnt-0.6b"
processor = AutoProcessor.from_pretrained(model_id)
model = AutoModelForRNNT.from_pretrained(model_id, dtype="auto", device_map="auto")

ds = load_dataset("hf-internal-testing/librispeech_asr_dummy", "clean", split="validation")
ds = ds.cast_column("audio", Audio(sampling_rate=processor.feature_extractor.sampling_rate))
speech_samples = [el["array"] for el in ds["audio"][:num_samples]]

inputs = processor(speech_samples, sampling_rate=processor.feature_extractor.sampling_rate)
inputs.to(model.device, dtype=model.dtype)
output = model.generate(**inputs, return_dict_in_generate=True)
print(processor.decode(output.sequences, skip_special_tokens=True))

from datasets import Audio, load_dataset
from transformers import AutoModelForRNNT, AutoProcessor

num_samples = 3

model_id = "nvidia/parakeet-rnnt-0.6b"
processor = AutoProcessor.from_pretrained(model_id)
model = AutoModelForRNNT.from_pretrained(model_id, dtype="auto", device_map="auto")

ds = load_dataset("hf-internal-testing/librispeech_asr_dummy", "clean", split="validation")
ds = ds.cast_column("audio", Audio(sampling_rate=processor.feature_extractor.sampling_rate))
speech_samples = [el["array"] for el in ds["audio"][:num_samples]]

inputs = processor(speech_samples, sampling_rate=processor.feature_extractor.sampling_rate)
inputs.to(model.device, dtype=model.dtype)
output = model.generate(**inputs, return_dict_in_generate=True)
decoded_output, decoded_timestamps = processor.decode(
    output.sequences,
    durations=output.durations,
    skip_special_tokens=True,
)
print("Transcription:", decoded_output)
print("Timestamped tokens:", decoded_timestamps)

from transformers import AutoModelForRNNT, AutoProcessor
from datasets import load_dataset, Audio
import torch

model_id = "nvidia/parakeet-rnnt-0.6b"
NUM_SAMPLES = 4

processor = AutoProcessor.from_pretrained(model_id)
model = AutoModelForRNNT.from_pretrained(model_id, dtype=torch.bfloat16, device_map="auto")
model.train()

ds = load_dataset("hf-internal-testing/librispeech_asr_dummy", "clean", split="validation")
ds = ds.cast_column("audio", Audio(sampling_rate=processor.feature_extractor.sampling_rate))
speech_samples = [el["array"] for el in ds["audio"][:NUM_SAMPLES]]
text_samples = ds["text"][:NUM_SAMPLES]

# passing `text` to the processor will prepare inputs' `labels` key
inputs = processor(audio=speech_samples, text=text_samples, sampling_rate=processor.feature_extractor.sampling_rate)
inputs.to(device=model.device, dtype=model.dtype)

outputs = model(**inputs)
print("Loss:", outputs.loss.item())
outputs.loss.backward()

For more details about usage, please refer to the Transformers' documentation.

Input

This model accepts 16000 Hz mono-channel audio (wav files) as input.

Output

This model provides transcribed speech as a string for a given audio sample.

Model Architecture

FastConformer [1] is an optimized version of the Conformer model with 8x depthwise-separable convolutional downsampling. The model is trained in a multitask setup with a Transducer decoder (RNNT) loss. You may find more information on the details of FastConformer here: Fast-Conformer Model.

Training

The NeMo toolkit [3] was used for training the models for over several hundred epochs. These model are trained with this example script and this base config.

The tokenizers for these models were built using the text transcripts of the train set with this script.

Datasets

The model was trained on 64K hours of English speech collected and prepared by NVIDIA NeMo and Suno teams.

The training dataset consists of private subset with 40K hours of English speech plus 24K hours from the following public datasets:

• Librispeech 960 hours of English speech
• Fisher Corpus
• Switchboard-1 Dataset
• WSJ-0 and WSJ-1
• National Speech Corpus (Part 1, Part 6)
• VCTK
• VoxPopuli (EN)
• Europarl-ASR (EN)
• Multilingual Librispeech (MLS EN) - 2,000 hour subset
• Mozilla Common Voice (v7.0)
• People's Speech - 12,000 hour subset

Performance

The performance of Automatic Speech Recognition models is measuring using Word Error Rate. Since this dataset is trained on multiple domains and a much larger corpus, it will generally perform better at transcribing audio in general.

The following tables summarizes the performance of the available models in this collection with the Transducer decoder. Performances of the ASR models are reported in terms of Word Error Rate (WER%) with greedy decoding.

These are greedy WER numbers without external LM. More details on evaluation can be found at HuggingFace ASR Leaderboard

NVIDIA Riva: Deployment

NVIDIA Riva, is an accelerated speech AI SDK deployable on-prem, in all clouds, multi-cloud, hybrid, on edge, and embedded. Additionally, Riva provides:

• World-class out-of-the-box accuracy for the most common languages with model checkpoints trained on proprietary data with hundreds of thousands of GPU-compute hours
• Best in class accuracy with run-time word boosting (e.g., brand and product names) and customization of acoustic model, language model, and inverse text normalization
• Streaming speech recognition, Kubernetes compatible scaling, and enterprise-grade support.

Although this model isn’t supported yet by Riva, the list of supported models is here.
Check out Riva live demo.

References

[1] Fast Conformer with Linearly Scalable Attention for Efficient Speech Recognition

[2] Google Sentencepiece Tokenizer

[3] NVIDIA NeMo Toolkit

[4] Suno.ai

[5] HuggingFace ASR Leaderboard