Multi-lingual model supporting speech-to-text recognition and translation.
RIVA Canary-1b-Flash (around 1B parameters) [1][2] is a multi-lingual multi-tasking model that supports automatic speech-to-text recognition (ASR) and automatic speech-to-text translation (AST) between all non-English and English in both directions for the following languages: Arabic (ar-AR), Danish (da-DK), Czech (cs-CZ), English (en-US), Spanish (es-US, es-ES), German (de-DE), French (fr-FR), Hindi (hi-IN), Italian (it-IT), Portuguese (pt-BR, pt-PT), Japanese (ja-JP), Korean (ko-KR), Dutch (nl-NL), Norwegian (nb-NO), Polish (pl-PO), Russian (ru-RU), Swedish (sv-SE), Turkish (tr-TR), and Mandarin (zh-CN) with punctuation and capitalization (PnC). Bulgarian (bg-BG), Greek (el-GR), Estonian (et-EE), Finnish (fi-FI), Croatian (hr-HR), Hungarian (hu-HU), Indonesian (id-ID), Lithuanian (lt-LT), Latvian (lv-LV), Romanian (ro-RO), Slovak (sk-SK), Slovenian (sl-SI), Ukrainian (uk-UA) and Vietnamese (vi-VN) are also supported as target languages for translation from English. British English (en-GB), Canadian French (fr-CA), and Norwegian Nynorsk (nn-NO) are supported for ASR and translation to English. Hebrew (he-IL) is supported for ASR. Thai (th-TH) is supported for ASR and translation target from English.
This model is ready for commercial use.
GOVERNING TERMS: The NIM container is governed by the NVIDIA Software License Agreement and Product-Specific Terms for AI Products. Use of this model is governed by the NVIDIA Community Model License.
Global
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: Build.Nvidia.com [10/15/2025] via [https://build.nvidia.com/nvidia/canary-1b-asr] Hugging Face [10/15/2025] via [URL] NGC [10/15/2025] via [https://build.stg.ngc.nvidia.com/nvidia/canary-1b-asr]
[1] Less is More: Accurate Speech Recognition & Translation without Web-Scale Data
[2] Training and Inference Efficiency of Encoder-Decoder Speech Models
[3] New Standard for Speech Recognition and Translation from the NVIDIA NeMo Canary Model
[4] Fast Conformer with Linearly Scalable Attention for Efficient Speech Recognition
[6] Attention Is All You Need
Architecture Type: This model was developed based on the Canary Flash Architecture [2]. An encoder-decoder model with FastConformer [4] encoder and Transformer decoder [5]
Network Architecture: 42 layer encoder, 8 layer decoder, Number of model parameters: 918M
Input Type(s): Audio
Input Format: wav
Input Parameters: One-Dimensional (1D)
Other Properties Related to Input: Mono channel is required
Output Type(s): Text
Output Format: String
Output Parameters: One-Dimensional (1D)
Other Properties Related to Output: No Maximum Character Length, Does not handle special characters
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.
Runtime Engine(s):
Supported Hardware Microarchitecture Compatibility:
[Preferred/Supported] Operating System(s):
The integration of foundation and fine-tuned models into AI systems requires additional testing using use-case-specific data to ensure safe and effective deployment. Following the V-model methodology, iterative testing and validation at both unit and system levels are essential to mitigate risks, meet technical and functional requirements, and ensure compliance with safety and ethical standards before deployment.
Canary-1B-Flash 2.0
Data Modality
[Audio]
[Text]
Audio Training Data Size
[10,000 to 1 Million Hours]
Data Collection Method by dataset
Labeling Method by dataset
Properties:
Mixture of organic ASR data aligned with human voices and machine generated translations to create AST pairings.
Data Collection Method by dataset
Labeling Method by dataset
Properties:
A dynamic blend of public and internal proprietary and customer datasets aligning text with human audio data.
Acceleration Engine: Triton
Test Hardware:
Get access to knowledge base articles and support cases or submit a ticket.
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.
Please report security vulnerabilities or NVIDIA AI Concerns here.