Vendor page
Nemotron Speech Streaming EN 0.6B model card
NVIDIA, Hugging Face, n.d.
Read the originalRead on October 17, 2026
- What it is
- NVIDIA's model card for a 0.6-billion-parameter English speech recognition model built for live, low-latency transcription. It describes a cache-aware streaming encoder with an RNN-T decoder.
- How much weight it carries
- Written by the model's publisher, so it is authoritative for architecture and intended use. The card's license is listed only as 'other' in its metadata, so read the full terms before commercial use. The accuracy numbers on third-party sites do not match the card and are not used here.
Facts taken from this source
- NVIDIA's Nemotron Speech Streaming EN 0.6B model uses a cache-aware FastConformer encoder with 24 layers and an RNN-T decoder, which lets it stream audio as it arrives.
“architecture with 24 encoder layers and an RNNT (Recurrent Neural Network Transducer) decoder” (Model architecture)