nvidia / automatic-speech-recognition updated 1 month ago

nemotron-speech-streaming-en-0.6b

Name: nemotron-speech-streaming-en-0.6b
Author: nvidia

/ Badge alignment consistency / img { display: inline; vertical-align: middle; }

Params

618 M

Context

—

Downloads 30d

419 K

Likes

624

Licence: other Not gated GGUF · SAFETENSORS View on Hugging Face ↗

Download history

tracking started — chart appears after 7 days of snapshots (4 recorded)

419 K downloads in the last 30 days

Estimated VRAM at 8K context unless noted. Pick your hardware to see the verdict per quantization.

File	Quant	Size	Est. VRAM	Verdict on RTX 4090 · 24 GB
nemotron-speech-streaming-en-0.6b.q8_0.gguf	Q8_0	0.7 GB	1.4 GB	✅ Runs comfortably
model.safetensors	f32	2.5 GB	3.3 GB	✅ Runs comfortably

Estimate: file size × 1.1 + KV cache at 8K + 0.5 GB overhead. Not a benchmark — how we calculate this.

Architecture: NemotronAsrStreamingForRNNT
Parameters: 618 M
Tensor type: F32
Vocabulary: 1,025
Licence: other
First seen on the Hub: 2025-12-17
Training datasets: nvidia/Granary, YTC, Yodas2, LibriLight, librispeech_asr, fisher_corpus, Switchboard-1, WSJ-0, WSJ-1, National-Singapore-Corpus-Part-1, National-Singapore-Corpus-Part-6, vctk, voxpopuli, europarl, multilingual_librispeech, fleurs, mozilla-foundation/common_voice_8_0, MLCommons/peoples_speech, google/speech_commands
AMI (reported): 11.73
TEDLIUM (reported): 3.5
VoxPopuli (reported): 7.97
Earnings22 (reported): 12.52
Gigaspeech (reported): 9.66
SPGI Speech (reported): 2.97
LibriSpeech test-clean (reported): 2.32
LibriSpeech test-other (reported): 4.84
Added to our catalog: 2026-09-18

Compare with any automatic-speech-recognition model

Popular comparisons