marathi_openslr64_wav2vec2_asrconformer5

espnet/marathi_openslr64_wav2vec2_asrconformer5

marathi_openslr64_wav2vec2_asrconformer5 is a automatic speech recognition model from espnet. Use it when you need speech turned into text.

- date: Wed Mar 23 05:58:21 UTC 2022 - python version: 3.9.10 | packaged by conda-forge | (main, Feb 1 2022, 21:24:11) [GCC 9.4.0] - espnet version: espnet 0.10.7a1 - pytorch version: pytorch 1.10.1 - Git hash: 1991a2…

Downloads · 30 days

Access

Public

Updated Sep 19, 2026

Repo size

18.3 GB

Likes

Public

Hugging Face

Repo makeup

Click a slice to open those files.

.pth35.8 GB · 98%

At a glance

Task: Automatic Speech Recognition
Access: Public
Created: Mar 23, 2022
Updated: Sep 19, 2026
SHA: a6ced28a

Model card

From the Hugging Face model README

Usage

import librosa
from espnet2.bin.asr_inference import Speech2Text

speech2text = Speech2Text.from_pretrained(model_tag="espnet/marathi_openslr64_wav2vec2_asrconformer5")
# librosa resamples and mixes to one channel, so any file works; 16000 is
# what nearly every espnet recogniser is trained on - check this model's
# config if its audio is not 16 kHz
speech, rate = librosa.load("audio.wav", sr=16000, mono=True)
text, *_ = speech2text(speech)[0]
print(text)

RESULTS

Environments

date: Wed Mar 23 05:58:21 UTC 2022
python version: 3.9.10 | packaged by conda-forge | (main, Feb 1 2022, 21:24:11) [GCC 9.4.0]
espnet version: espnet 0.10.7a1
pytorch version: pytorch 1.10.1
Git hash: 1991a25855821b8b61d775681aa0cdfd6161bbc8
- Commit date: Mon Mar 21 22:19:19 2022 +0800

asr_train_asr_conformer5_raw_mr_bpe150_sp

WER

dataset	Snt	Wrd	Corr	Sub	Del	Ins	Err	S.Err
inference_asr_model_valid.acc.ave/dev_mr	137	1563	80.2	17.1	2.8	2.0	21.8	71.5
inference_asr_model_valid.acc.ave/test_mr	200	2536	73.9	20.8	5.4	1.1	27.2	82.0
inference_lm_config_mr_bpe150_valid.loss.ave_asr_model_valid.acc.ave/dev_mr	137	1563	81.3	15.6	3.1	2.0	20.7	72.3
inference_lm_config_mr_bpe150_valid.loss.ave_asr_model_valid.acc.ave/test_mr	200	2536	76.6	20.7	2.7	0.9	24.3	80.5

CER

dataset	Snt	Wrd	Corr	Sub	Del	Ins	Err	S.Err
inference_asr_model_valid.acc.ave/dev_mr	137	9369	93.7	2.8	3.5	2.3	8.6	71.5
inference_asr_model_valid.acc.ave/test_mr	200	14174	90.3	3.7	5.9	1.6	11.3	82.0
inference_lm_config_mr_bpe150_valid.loss.ave_asr_model_valid.acc.ave/dev_mr	137	9369	92.4	3.8	3.8	2.7	10.2	72.3
inference_lm_config_mr_bpe150_valid.loss.ave_asr_model_valid.acc.ave/test_mr	200	14174	88.3	7.6	4.1	2.7	14.4	80.5

TER

dataset	Snt	Wrd	Corr	Sub	Del	Ins	Err	S.Err
inference_asr_model_valid.acc.ave/dev_mr	137	6050	90.0	5.6	4.5	2.4	12.4	71.5
inference_asr_model_valid.acc.ave/test_mr	200	9254	85.6	7.6	6.8	1.6	16.0	82.0
inference_lm_config_mr_bpe150_valid.loss.ave_asr_model_valid.acc.ave/dev_mr	137	6050	88.8	7.0	4.2	2.7	13.9	72.3
inference_lm_config_mr_bpe150_valid.loss.ave_asr_model_valid.acc.ave/test_mr	200	9254	83.2	12.3	4.5	3.9	20.7	80.5

Task: Automatic Speech Recognition
Created: Mar 23, 2022
Updated: Sep 19, 2026