explainability.md

Field	Response
Intended Task/Domain:	Speech Recognition
Model Type:	FastConformer-RNNT
Intended Users:	People who work with conversational AI models and need to transcribe speech to text with low latency in streaming scenarios.
Output:	Text tokens
Describe how the model works:	Raw audio is passed into the model, and the model outputs text.
Name the adversely impacted groups this has been tested to deliver comparable outcomes regardless of:	Not Applicable
Technical Limitations & Mitigation:	The model is trained on only a limited amount of English speech data; therefore, it may not work well for other languages, and its performance may degrade in noisy environments.
Verified to have met prescribed NVIDIA quality standards:	Yes
Performance Metrics:	Word Error Rate (WER)
Potential Known Risks:	The model may produce incorrect transcriptions if the audio is noisy or the speech is not clear, and predicted text may be inaccurate in domains that are not well-represented in the training data.
Licensing:	NVIDIA Open Model License