nvidia/parakeet-ctc-1.1b · Hugging Face

|

|

parakeet-ctc-1.1b 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 XXL version of FastConformer CTC [1] (around 1.1B parameters) model. See the

model architecture

section and

NeMo documentation

for complete architecture details.

NVIDIA NeMo: Training

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'] How to Use this Model

There are several ways to use this model. Choose the one that fits your needs.

Run locally with NeMo-Speech.cpp

NeMo-Speech.cpp

provides a lightweight native C++ runtime for local inference with this model. After

installing the runtime

:

hf download nvidia/parakeet-ctc-1.1b \ parakeet-ctc-1.1b.q8_0.gguf \ --local-dir models nemo-speech transcribe audio.wav \ --model models/parakeet-ctc-1.1b.q8_0.gguf See the

NeMo-Speech.cpp documentation

for more details.

NVIDIA NeMo

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. Moreover, you can now run Parakeet CTC natively with

Transformers

🤗.

Automatically instantiate the model

import nemo.collections.asr as nemo_asr asr_model = nemo_asr.models.EncDecCTCModelBPE.from_pretrained(model_name="nvidia/parakeet-ctc-1.1b") Transcribing using NeMo

First, let's get a sample

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

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

Transformers

🤗

Make sure to install transformers from source.

pip install git+https://github.com/huggingface/transformers ➡️ Pipeline usagefrom transformers import pipeline pipe = pipeline("automatic-speech-recognition", model="nvidia/parakeet-ctc-1.1b") out = pipe("https://huggingface.co/datasets/hf-internal-testing/dummy-audio-samples/resolve/main/bcn_weather.mp3") print(out) ➡️ AutoModelfrom transformers import AutoModelForCTC, AutoProcessor from datasets import load_dataset, Audio import torch device = "cuda"if torch.cuda.is_available() else"cpu" processor = AutoProcessor.from_pretrained("nvidia/parakeet-ctc-1.1b") model = AutoModelForCTC.from_pretrained("nvidia/parakeet-ctc-1.1b", dtype="auto", device_map=device) 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"][:5]] inputs = processor(speech_samples, sampling_rate=processor.feature_extractor.sampling_rate) inputs.to(model.device, dtype=model.dtype) outputs = model.generate(**inputs) print(processor.batch_decode(outputs)) ➡️ Trainingfrom transformers import AutoModelForCTC, AutoProcessor from datasets import load_dataset, Audio import torch device = "cuda"if torch.cuda.is_available() else"cpu" processor = AutoProcessor.from_pretrained("nvidia/parakeet-ctc-1.1b") model = AutoModelForCTC.from_pretrained("nvidia/parakeet-ctc-1.1b", dtype="auto", device_map=device) 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"][:5]] text_samples = [el for el in ds["text"][:5]] # 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, dtype=model.dtype) outputs = model(**inputs) outputs.loss.backward() For more details about usage, the refer to

Transformers' documentation

.

Transcribing many audio files

python [NEMO_GIT_FOLDER]/examples/asr/transcribe_speech.py pretrained_name="nvidia/parakeet-ctc-1.1b" audio_dir="<DIRECTORY CONTAINING AUDIO FILES>" 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 using CTC 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 CTC decoder. Performances of the ASR models are reported in terms of Word Error Rate (WER%) with greedy decoding.

VersionTokenizerVocabulary SizeAMIEarnings-22Giga SpeechLS test-cleanSPGI SpeechTEDLIUM-v3Vox PopuliCommon Voice1.22.0SentencePiece Unigram102415.6213.6910.271.833.544.203.546.53These 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

Licence

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.