XLS-R Kyrgiz CV8 Open-Source Speech Recognition Model - Free Deployment for Accurate Kyrgyz Speech Recognition

Xls R Kyrgiz Cv8

Developed by lucio

This model is a fine-tuned automatic speech recognition model based on facebook/wav2vec2-xls-r-300m on the Common Voice 8.0 Kyrgyz dataset

Speech Recognition

Transformers

OtherOpen Source License:Apache-2.0 #Kyrgyz speech recognition #Low-resource language processing #WER optimization

Downloads 16

Release Time : 3/2/2022

Model Overview

Speech recognition model optimized for Kyrgyz language, suitable for speech-to-text conversion tasks

Model Features

Low Word Error Rate

Achieves 19.01% WER on test set (with language model)

Multi-scenario applicability

Optimized for low-fidelity speech scenarios, suitable for various practical applications

Progressive learning

Adopts progressive learning rate scheduling strategy to optimize training results

Model Capabilities

Kyrgyz speech recognition

Speech-to-text

Audio content indexing

Use Cases

Media processing

Video caption generation

Automatically generates draft subtitles for Kyrgyz video content

WER 19.01% (with language model)

Broadcast content indexing

Indexes content of recorded Kyrgyz radio programs

CER 5.38% (with language model)

🚀 XLS-R-300M Kyrgiz CV8

This model is a fine - tuned version of facebook/wav2vec2 - xls - r - 300m on the MOZILLA - FOUNDATION/COMMON_VOICE_8_0 - KY dataset, aiming at automatic speech recognition tasks.

✨ Features

Fine - tuned Model: Based on facebook/wav2vec2 - xls - r - 300m, fine - tuned on the MOZILLA - FOUNDATION/COMMON_VOICE_8_0 - KY dataset.
Good Performance: Achieves certain results on the validation set, including Loss: 0.5497, Wer: 0.2945, Cer: 0.0791.

📚 Documentation

Model description

For a description of the model architecture, see facebook/wav2vec2 - xls - r - 300m.

Vocabulary: The model vocabulary consists of the cyrillic alphabet with punctuation removed.
Language Model: The kenlm language model is built using the text of the train and invalidated corpus splits.

Intended uses & limitations

This model is expected to be of some utility for low - fidelity use cases such as:

Draft video captions
Indexing of recorded broadcasts

The model is not reliable enough to use as a substitute for live captions for accessibility purposes, and it should not be used in a manner that would infringe the privacy of any of the contributors to the Common Voice dataset nor any other speakers.

Training and evaluation data

The combination of train, dev and other of common voice official splits were used as training data. The half of the official test split was used as validation data, as and the full test set was used for final evaluation.

Training procedure

The featurization layers of the XLS - R model are frozen while tuning a final CTC/LM layer on the Kyrgiz CV8 example sentences. A ramped learning rate is used with an initial warmup phase of 500 steps, a max of 0.0001, and cooling back towards 0 for the remainder of the 8100 steps (300 epochs).

Training hyperparameters

The following hyperparameters were used during training:

Property	Details
learning_rate	0.0001
train_batch_size	32
eval_batch_size	8
seed	42
gradient_accumulation_steps	4
total_train_batch_size	128
optimizer	Adam with betas=(0.9,0.999) and epsilon = 1e - 08
lr_scheduler_type	linear
lr_scheduler_warmup_steps	500
num_epochs	300.0
mixed_precision_training	Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer	Cer
3.1079	18.51	500	2.6795	0.9996	0.9825
0.8506	37.04	1000	0.4323	0.3718	0.0961
0.6821	55.55	1500	0.4105	0.3311	0.0878
0.6091	74.07	2000	0.4281	0.3168	0.0851
0.5429	92.58	2500	0.4525	0.3147	0.0842
0.5063	111.11	3000	0.4619	0.3144	0.0839
0.4661	129.62	3500	0.4660	0.3039	0.0818
0.4353	148.15	4000	0.4695	0.3083	0.0820
0.4048	166.65	4500	0.4909	0.3085	0.0824
0.3852	185.18	5000	0.5074	0.3048	0.0812
0.3567	203.69	5500	0.5111	0.3012	0.0810
0.3451	222.22	6000	0.5225	0.2982	0.0804
0.325	240.73	6500	0.5270	0.2955	0.0796
0.3089	259.25	7000	0.5381	0.2929	0.0793
0.2941	277.76	7500	0.5565	0.2923	0.0794
0.2945	296.29	8000	0.5495	0.2951	0.0789

Framework versions

Transformers 4.17.0.dev0
Pytorch 1.10.2+cu102
Datasets 1.18.3
Tokenizers 0.11.0

📄 License

This model is licensed under the apache - 2.0 license.

Featured Recommended AI Models

Empowering the Future, Your AI Solution Knowledge Base

English 简体中文繁體中文にほんご