Ported from upstream commit
1b149a3,
pinned 2026-04-15.
Validated against the NeMo reference at transcribe.cpp commit
bf0d0b7
on 2026-04-18.
Offline English speech-to-text. A 0.6B-parameter Conformer encoder with a
TDT/RNNT transducer decoder. Takes a 16 kHz mono WAV and produces a transcript
with optional token-level timestamps. Not a streaming model; no multilingual
capability (see v3 for that).
WER on the full LibriSpeech test-clean split (2,620 utterances), batch size 1, timestamps none. Figures without a commit were published before provenance was recorded.
Greedy transducer decoding, no external LM. F32 reference baseline: 1.68%. NVIDIA's
self-reported number on the same split is 1.69%, so the F32 and Q8_0 ports match the
upstream reference within rounding.
If your audio isn't already 16 kHz mono WAV, convert it first:
ffmpeg -i input.mp3 -ar 16000 -ac 1 output.wav
See the transcribe.cpp model page for performance
numbers, numerical validation, and reproduction steps.
License
Inherited from the base model: CC-BY-4.0. See the
upstream model card for full terms.
Original Model Card
The section below is reproduced from
nvidia/parakeet-tdt-0.6b-v2 at commit
1b149a3 for offline reference. The upstream card is the
authoritative source.
parakeet-tdt-0.6b-v2 is a 600-million-parameter automatic speech recognition (ASR) model designed for high-quality English transcription, featuring support for punctuation, capitalization, and accurate timestamp prediction. Try Demo here: https://huggingface.co/spaces/nvidia/parakeet-tdt-0.6b-v2
This XL variant of the FastConformer [1] architecture integrates the TDT [2] decoder and is trained with full attention, enabling efficient transcription of audio segments up to 24 minutes in a single pass. The model achieves an RTFx of 3380 on the HF-Open-ASR leaderboard with a batch size of 128. Note: RTFx Performance may vary depending on dataset audio duration and batch size.
Key Features
Accurate word-level timestamp predictions
Automatic punctuation and capitalization
Robust performance on spoken numbers, and song lyrics transcription
This model is ready for commercial/non-commercial use.
License/Terms of Use:
GOVERNING TERMS: Use of this model is governed by 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.
This model serves developers, researchers, academics, and industries building applications that require speech-to-text capabilities, including but not limited to: conversational AI, voice assistants, transcription services, subtitle generation, and voice analytics platforms.
Release Date:
05/01/2025
Model Architecture:
Architecture Type:
FastConformer-TDT
Network Architecture:
This model was developed based on FastConformer encoder architecture[1] and TDT decoder[2]
This model has 600 million model parameters.
Input:
Input Type(s): 16kHz Audio
Input Format(s):.wav and .flac audio formats
Input Parameters: 1D (audio signal)
Other Properties Related to Input: Monochannel audio
Output:
Output Type(s): Text
Output Format: String
Output Parameters: 1D (text)
Other Properties Related to Output: Punctuations and Capitalizations included.
Our AI models are designed and/or optimized to run on NVIDIA GPU-accelerated systems. By leveraging NVIDIA's hardware (e.g. GPU cores) and software frameworks (e.g., CUDA libraries), the model achieves faster training and inference times compared to CPU-only solutions.
How to Use this Model:
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 -U nemo_toolkit["asr"]
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.
Automatically instantiate the model
python
1import nemo.collections.asr as nemo_asr
2asr_model = nemo_asr.models.ASRModel.from_pretrained(model_name="nvidia/parakeet-tdt-0.6b-v2")
1output = asr_model.transcribe(['2086-149220-0033.wav'], timestamps=True)2# by default, timestamps are enabled for char, word and segment level3word_timestamps = output[0].timestamp['word']# word level timestamps for first sample4segment_timestamps = output[0].timestamp['segment']# segment level timestamps5char_timestamps = output[0].timestamp['char']# char level timestamps67for stamp in segment_timestamps:8print(f"{stamp['start']}s - {stamp['end']}s : {stamp['segment']}")
Try via API — No Setup Required
Transcribe audio instantly using the free hosted API on build.nvidia.com — no GPU, no Docker, no model download needed.
Atleast 2GB RAM for model to load. The bigger the RAM, the larger audio input it supports.
Model Version
Current version: parakeet-tdt-0.6b-v2. Previous versions can be accessed here.
Training and Evaluation Datasets:
Training
This model was trained using the NeMo toolkit [3], following the strategies below:
Initialized from a FastConformer SSL checkpoint that was pretrained with a wav2vec method on the LibriLight dataset[7].
Trained for 150,000 steps on 64 A100 GPUs.
Dataset corpora were balanced using a temperature sampling value of 0.5.
Stage 2 fine-tuning was performed for 2,500 steps on 4 A100 GPUs using approximately 500 hours of high-quality, human-transcribed data of NeMo ASR Set 3.0.
All transcriptions preserve punctuation and capitalization. The Granary dataset[8] will be made publicly available after presentation at Interspeech 2025.
Data Collection Method by dataset
Hybrid: Automated, Human
Labeling Method by dataset
Hybrid: Synthetic, Human
Properties:
Noise robust data from various sources
Single channel, 16kHz sampled data
Evaluation Dataset
Huggingface Open ASR Leaderboard datasets are used to evaluate the performance of this model.
Data Collection Method by dataset
Human
Labeling Method by dataset
Human
Properties:
All are commonly used for benchmarking English ASR systems.
Audio data is typically processed into a 16kHz mono channel format for ASR evaluation, consistent with benchmarks like the Open ASR Leaderboard.
Performance
Huggingface Open-ASR-Leaderboard Performance
The performance of Automatic Speech Recognition (ASR) models is measured using Word Error Rate (WER). Given that this model is trained on a large and diverse dataset spanning multiple domains, it is generally more robust and accurate across various types of audio.
Base Performance
The table below summarizes the WER (%) using a Transducer decoder with greedy decoding (without an external language model):
Model
Avg WER
AMI
Earnings-22
GigaSpeech
LS test-clean
LS test-other
SPGI Speech
TEDLIUM-v3
VoxPopuli
parakeet-tdt-0.6b-v2
6.05
11.16
11.15
9.74
1.69
3.19
2.17
3.38
5.95
Noise Robustness
Performance across different Signal-to-Noise Ratios (SNR) using MUSAN music and noise samples:
SNR Level
Avg WER
AMI
Earnings
GigaSpeech
LS test-clean
LS test-other
SPGI
Tedlium
VoxPopuli
Relative Change
Clean
6.05
11.16
11.15
9.74
1.69
3.19
2.17
3.38
5.95
-
SNR 10
6.95
14.38
12.04
10.24
1.92
4.13
2.84
3.63
6.38
-14.75%
SNR 5
8.23
18.07
13.82
11.18
2.33
5.58
3.81
4.24
6.81
-35.97%
SNR 0
11.88
25.43
18.59
14.32
4.40
10.07
7.27
6.42
8.54
-96.28%
SNR -5
20.26
36.57
28.06
22.27
11.82
19.91
16.14
13.07
14.23
-234.66%
Telephony Audio Performance
Performance comparison between standard 16kHz audio and telephony-style audio (using μ-law encoding with 16kHz→8kHz→16kHz conversion):
Audio Format
Avg WER
AMI
Earnings
GigaSpeech
LS test-clean
LS test-other
SPGI
Tedlium
VoxPopuli
Relative Change
Standard 16kHz
6.05
11.16
11.15
9.74
1.69
3.19
2.17
3.38
5.95
-
μ-law 8kHz
6.32
11.98
11.16
10.02
1.78
3.52
2.20
3.38
6.52
-4.10%
These WER scores were obtained using greedy decoding without an external language model. Additional evaluation details are available on the Hugging Face ASR Leaderboard.[6]
NVIDIA believes Trustworthy AI is a shared responsibility and we have established policies and practices to enable development for a wide array of AI applications. When downloaded or used in accordance with our terms of service, developers should work with their supporting model team to ensure this model meets requirements for the relevant industry and use case and addresses unforeseen product misuse.
For more detailed information on ethical considerations for this model, please see the Model Card++ Explainability, Bias, Safety & Security, and Privacy Subcards here.
Please report security vulnerabilities or NVIDIA AI Concerns here.
Bias:
Field
Response
Participation considerations from adversely impacted groups protected classes in model design and testing
None
Measures taken to mitigate against unwanted bias
None
Explainability:
Field
Response
Intended Domain
Speech to Text Transcription
Model Type
FastConformer
Intended Users
This model is intended for developers, researchers, academics, and industries building conversational based applications.
Output
Text
Describe how the model works
Speech input is encoded into embeddings and passed into conformer-based model and output a text response.
Name the adversely impacted groups this has been tested to deliver comparable outcomes regardless of
Not Applicable
Technical Limitations & Mitigation
Transcripts may be not 100% accurate. Accuracy varies based on language and characteristics of input audio (Domain, Use Case, Accent, Noise, Speech Type, Context of speech, etc.)
Verified to have met prescribed NVIDIA quality standards
Yes
Performance Metrics
Word Error Rate
Potential Known Risks
If a word is not trained in the language model and not presented in vocabulary, the word is not likely to be recognized. Not recommended for word-for-word/incomplete sentences as accuracy varies based on the context of input text
Licensing
GOVERNING TERMS: Use of this model is governed by the CC-BY-4.0 license.
Privacy:
Field
Response
Generatable or reverse engineerable personal data?
None
Personal data used to create this model?
None
Is there provenance for all datasets used in training?
Yes
Does data labeling (annotation, metadata) comply with privacy laws?
Yes
Is data compliant with data subject requests for data correction or removal, if such a request was made?
The Principle of least privilege (PoLP) is applied limiting access for dataset generation and model development. Restrictions enforce dataset access during training, and dataset license constraints adhered to.
