xls-r-1B-te is free and open-source! Achieve automatic speech recognition in Telugu, a practical choice!

Xls R 1B Te

Developed by chmanoj

A Telugu speech recognition model fine-tuned on the OPENSLR_SLR66-NA dataset based on facebook/wav2vec2-xls-r-1b, supporting automatic speech recognition tasks.

Speech Recognition

Transformers

OtherOpen Source License:Apache-2.0 #Telugu speech recognition #Low word error rate #Large model fine-tuning

Downloads 18

Release Time : 3/2/2022

Model Overview

This model is an automatic speech recognition system optimized for Telugu, based on the XLS-R architecture and fine-tuned on the SLR66 dataset from OpenSLR, suitable for Telugu speech-to-text tasks.

Model Features

High-precision Telugu recognition

Achieves a word error rate (WER) of 20.624% and a character error rate (CER) of 3.979% on the test set

Large-scale pre-training foundation

Fine-tuned based on the facebook/wav2vec2-xls-r-1b model, inheriting powerful feature extraction capabilities from large-scale pre-trained models

Supports language model decoding

Can further improve recognition accuracy by integrating language models, reducing WER from 26.14% to 20.69%

Model Capabilities

Telugu speech recognition

Speech-to-text

Supports language model integration

Use Cases

Speech transcription

Telugu speech transcription

Convert Telugu speech content into text

Achieves 20.624% WER on the test set

Voice assistants

Telugu voice interaction

Provides voice interaction capabilities for Telugu users

🚀 xls-r-1B-te

This model is a fine - tuned version of facebook/wav2vec2-xls-r-1b on the OPENSLR_SLR66 - NA dataset, which is used for automatic speech recognition and achieves good results in evaluation.

🚀 Quick Start

This model is a fine - tuned version of facebook/wav2vec2-xls-r-1b on the OPENSLR_SLR66 - NA dataset. It achieves the following results on the evaluation set:

Loss: 0.3119
Wer: 0.2613

Evaluation metrics

Metric	Split	Decode with LM	Value
WER	Train	No	5.36
CER	Train	No	1.11
WER	Test	No	26.14
CER	Test	No	4.93
WER	Train	Yes	5.04
CER	Train	Yes	1.07
WER	Test	Yes	20.69
CER	Test	Yes	3.986

✨ Features

Automatic Speech Recognition: Specifically designed for automatic speech recognition tasks.
Fine - Tuned: Based on the pre - trained [facebook/wav2vec2-xls-r-1b] model, fine - tuned on the OPENSLR_SLR66 - NA dataset.
Multiple Metrics Evaluation: Evaluated using metrics such as WER and CER, providing comprehensive performance assessment.

📚 Documentation

Model Index

Property	Details
Name	xls - r - 1B - te
Task Type	Automatic Speech Recognition
Dataset	Open SLR (SLR66)
Metrics	Test WER: 20.624 Test CER: 3.979 Test WER (without LM): 26.14777618364419 Test CER (without LM): 4.932543184970369

Training and Evaluation

Training hyperparameters

The following hyperparameters were used during training:

learning_rate: 2e - 05
train_batch_size: 16
eval_batch_size: 4
seed: 42
gradient_accumulation_steps: 2
total_train_batch_size: 32
optimizer: Adam with betas=(0.9,0.999) and epsilon = 1e - 08
lr_scheduler_type: linear
lr_scheduler_warmup_steps: 2000
num_epochs: 150.0
mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
2.9038	4.8	500	3.0125	1.0
1.3777	9.61	1000	0.8681	0.8753
1.1436	14.42	1500	0.6256	0.7961
1.0997	19.23	2000	0.5244	0.6875
1.0363	24.04	2500	0.4585	0.6276
0.7996	28.84	3000	0.4072	0.5295
0.825	33.65	3500	0.3590	0.5222
0.8018	38.46	4000	0.3678	0.4671
0.7545	43.27	4500	0.3474	0.3962
0.7375	48.08	5000	0.3224	0.3869
0.6198	52.88	5500	0.3233	0.3630
0.6608	57.69	6000	0.3029	0.3308
0.645	62.5	6500	0.3195	0.3722
0.5249	67.31	7000	0.3004	0.3202
0.4875	72.11	7500	0.2826	0.2992
0.5171	76.92	8000	0.2962	0.2976
0.4974	81.73	8500	0.2990	0.2933
0.4387	86.54	9000	0.2834	0.2755
0.4511	91.34	9500	0.2886	0.2787
0.4112	96.15	10000	0.3093	0.2976
0.4064	100.96	10500	0.3123	0.2863
0.4047	105.77	11000	0.2968	0.2719
0.3519	110.57	11500	0.3106	0.2832
0.3719	115.38	12000	0.3030	0.2737
0.3669	120.19	12500	0.2964	0.2714
0.3386	125.0	13000	0.3101	0.2714
0.3137	129.8	13500	0.3063	0.2710
0.3008	134.61	14000	0.3082	0.2617
0.301	139.42	14500	0.3121	0.2628
0.3291	144.23	15000	0.3105	0.2612
0.3133	149.04	15500	0.3114	0.2624

Framework versions

Transformers 4.16.0.dev0
Pytorch 1.10.1+cu102
Datasets 1.17.1.dev0
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