HF Model & Benchmark Matrix

24,078 variants · 14,984 scores · 120 benchmarks

parakeet-tdt_ctc-1.1b

nvidia/parakeet-tdt_ctc-1.1b

kind
original
quant
bf16 (16-bit)
family
misc
modality
automatic-speech-recognition
root model
nvidia/parakeet-tdt_ctc-1.1b
provider
nvidia
license
cc-by-4.0
language
en
context
—
library
nemo
vRAM (GB)
—
size on disk
—
downloads
879
likes
22
created
—
last modified
2025-02-18T13:41:32+00:00 (2025)
huggingface
https://huggingface.co/nvidia/parakeet-tdt_ctc-1.1b
tags
nemoautomatic-speech-recognitionspeechaudioTransducerTDTFastConformerConformerpytorchNeMohf-asr-leaderboardendataset:librispeech_asrdataset:fisher_corpusdataset:mozilla-foundation/common_voice_8_0dataset:National-Singapore-Corpus-Part-1dataset:vctkdataset:voxpopulidataset:europarldataset:multilingual_librispeecharxiv:2305.05084arxiv:2304.06795arxiv:2104.02821license:cc-by-4.0

Benchmark scores (8)

benchmarkcategoryscorecontextconfevidence
WER AMIspeech-asr15.94low| 1.23.0 | SentencePiece Unigram | 1024 | 15.94 | 11.86 | 10.19 | 1.82 | 3.67 | 2.24 | 3.87 | 6.19 | 8.69 |
WERspeech-asr12.81% WERmedium| % WER | 12.81 | 10.49 | 13.88 | 23.12 |
WER Earnings-22speech-asr11.86low| 1.23.0 | SentencePiece Unigram | 1024 | 15.94 | 11.86 | 10.19 | 1.82 | 3.67 | 2.24 | 3.87 | 6.19 | 8.69 |
WER GigaSpeechspeech-asr10.19low| 1.23.0 | SentencePiece Unigram | 1024 | 15.94 | 11.86 | 10.19 | 1.82 | 3.67 | 2.24 | 3.87 | 6.19 | 8.69 |
WER Common Voicespeech-asr8.69low| 1.23.0 | SentencePiece Unigram | 1024 | 15.94 | 11.86 | 10.19 | 1.82 | 3.67 | 2.24 | 3.87 | 6.19 | 8.69 |
WER VoxPopulispeech-asr6.19low| 1.23.0 | SentencePiece Unigram | 1024 | 15.94 | 11.86 | 10.19 | 1.82 | 3.67 | 2.24 | 3.87 | 6.19 | 8.69 |
WER TED-LIUMspeech-asr3.87low| 1.23.0 | SentencePiece Unigram | 1024 | 15.94 | 11.86 | 10.19 | 1.82 | 3.67 | 2.24 | 3.87 | 6.19 | 8.69 |
WER SPGISpeechspeech-asr2.24low| 1.23.0 | SentencePiece Unigram | 1024 | 15.94 | 11.86 | 10.19 | 1.82 | 3.67 | 2.24 | 3.87 | 6.19 | 8.69 |

Discussions (0)

none scraped.

Hugging Face model & benchmark data scraped & matrix-ified. 24,078 variants · 14,984 scores · 120 benchmarks.