whatmodel.sh / stt / STT En FastConformer-Transducer Large

STT En FastConformer-Transducer Large

+ contribute
Update model informationAdd a place to run itAdd a test resultPropose a benchmarkCorrect a benchmark scoreEdit data on GitHub

STT En FastConformer-Transducer Large is an open-weights speech-to-text model from NVIDIA, released June 2023. Its published size is 0.114B parameters. Licence: cc-by-4.0.

Catalogue identity notes

Exact checkpoint nvidia/stt_en_fastconformer_transducer_large (NeMo). Params: card (~114M). released: HF repo creation month.

45.6±3.2#118 of 162 speech-to-text modelsfrom 4 benchmark results across 2 sources

Score within the Open ASR English short-form WER (March 2026) comparison set only.

$ how it compares

The speech-to-text models ranked either side of STT En FastConformer-Transducer Large, out of 176 with a score, then more from NVIDIA.

#modelscorewhereRAMreleased
129ESPnet E-Branchformer (no LM) ESPnet46±4local~0.8GB2023-01
130ESPnet Multiconvformer (no LM) ESPnet46±4local~0.8GB2024-03
131STT En FastConformer-Transducer Large NVIDIA46±3local~1.1GB2023-06
132Moonshine Streaming Small Useful Sensors (Moonshine AI)46±3local~1.2GB2026-01
133SenseVoice Small Alibaba (FunAudioLLM)46±3local~0.3GB2024-07
more from NVIDIA
28Canary-Qwen-2.5B NVIDIA61±2local~6.6GB2025-07
35Parakeet TDT 0.6B v2 NVIDIA60±1local~1.9GB2025-05
37Parakeet Unified EN 0.6B NVIDIA60±4local~0.7GB2026-04

see all speech-to-text models · find the ones that fit your machine

$ what it's good at

long-form transcription▁▃▄▁·▁▂█▄▁▁▁ #26 of 27 * · 35% of its score · 1 test

Trails STT En Conformer-CTC Small, STT En Conformer-Transducer Small, STT En FastConformer-CTC Large here, though they rank lower overall.

testresultvs the others in its rundifficultycountswhy
Open ASR English long-form WER16.3% ✎▁▃▅▁·▃▅█▆▁▂▂ beats 1/30avg 12%▇▇▇▇▇▇▇▇▇▇▇▇▇▇ 35%typical
other tests3 not in a task · 61% of its score
testresultvs the others in its rundifficultycountswhy
Open ASR English short-form WER (March 2026)9.1% ✎▁▁▁▃··▁▁▂██▆ beats 22/86avg 8.2%▇▇▇▇▇▇▇▇▇▇▇▇▇▇ 35%blunt
LibriSpeech test-other WER lab3.8% ✎▁▁▁▂▅▃▅▄█▂▂▁ beats 42/54avg 3.5%▇▇▇▇▇▇▇▇ 21%typical
LibriSpeech test-clean WER lab1.8% ✎▁·▁·▁▁▁▁▃▇█▆ beats 74/100avg 1.8%▇▇ 6.0%blunt

$ where to run it

Other ways to run it

  1. Run it as published · needs ~1.05 GB, NVIDIA only · from huggingface.co/nvidia (official)
    uvx --with 'nemo_toolkit[asr]' python -c "import nemo.collections.asr as a; print(a.models.ASRModel.from_pretrained('nvidia/stt_en_fastconformer_transducer_large').transcribe(['audio.wav']))"

Official weights and documentation: huggingface.co/nvidia ↗

$ how the score was worked out · Open ASR Leaderboard 69%, NVIDIA NeMo docs: ASR scores… 27%

Each result implies a score on its own, after adjusting for how hard its test is and who ran it. The score is their weighted average, with a small pull toward the middle for thinly tested models. Each test's weight is its counts column above; here it's added up by who ran the tests. How it works

sourceresultsshare of the scorelicence
Open ASR Leaderboard2▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇ 69%Apache-2.0
NVIDIA NeMo docs: ASR scores tables (English Conformer / FastConformer) · lab-reported2▇▇▇▇▇▇▇▇ 27%Apache-2.0
prior · pull toward the middle▇ 4%
weighted average= 45.6

$ improve this page

Update model informationAdd a place to run itAdd a test resultPropose a benchmarkCorrect a benchmark scoreEdit data on GitHub

Use a short GitHub form or edit the data through a pull request. Every number needs a source; here's which ones we can use.

data updated · sources and licences · json