viet_tones_model Open-source Vietnamese Tone Recognition Model - Free Deployment for High-Accuracy Tone Recognition

Home

Viet Tones Model

Developed by StevenLe456

Vietnamese tone recognition model fine-tuned on wav2vec2-base-vietnamese-250h, accuracy 59.72%

Speech Recognition

Transformers

#Vietnamese tone recognition #wav2vec2 fine-tuning #low-resource speech processing

Downloads 22

Release Time : 8/10/2023

Model Overview

This model focuses on tone recognition tasks in Vietnamese speech, suitable for speech processing and analysis scenarios

Model Features

Vietnamese Tone Recognition

Specialized classification capability for six Vietnamese tones

Fine-tuned on Pre-trained Model

Optimized based on wav2vec2 pre-trained with 250 hours of Vietnamese speech data

Model Capabilities

Vietnamese speech analysis

Tone feature extraction

Speech classification

Use Cases

Speech Processing

Vietnamese Pronunciation Assessment

Used for pronunciation accuracy evaluation in language learning applications

Can recognize 59.72% of tones

Speech-to-Text Preprocessing

Serves as a pre-processing tone module for Vietnamese ASR systems

🚀 viet_tones_model

This model is a fine - tuned version of [nguyenvulebinh/wav2vec2 - base - vietnamese - 250h](https://huggingface.co/nguyenvulebinh/wav2vec2 - base - vietnamese - 250h), aiming to achieve better performance on specific tasks.

🚀 Quick Start

This model is a fine - tuned version of [nguyenvulebinh/wav2vec2 - base - vietnamese - 250h](https://huggingface.co/nguyenvulebinh/wav2vec2 - base - vietnamese - 250h) on an unknown dataset. It achieves the following results on the evaluation set:

Loss: 0.9783
Accuracy: 0.5972

📚 Documentation

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

🔧 Technical Details

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

learning_rate: 5e - 05
train_batch_size: 32
eval_batch_size: 32
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_ratio: 0.1
num_epochs: 110

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
No log	0.89	6	1.7955	0.1296
1.7924	1.93	13	1.7938	0.1343
1.7919	2.96	20	1.7916	0.2037
1.7919	4.0	27	1.7907	0.1713
1.7903	4.89	33	1.7886	0.1852
1.7883	5.93	40	1.7798	0.2269
1.7883	6.96	47	1.7487	0.25
1.7717	8.0	54	1.7104	0.2407
1.726	8.89	60	1.6488	0.2685
1.726	9.93	67	1.5835	0.2731
1.6651	10.96	74	1.6020	0.2778
1.6332	12.0	81	1.5351	0.2778
1.6332	12.89	87	1.4977	0.2963
1.5708	13.93	94	1.4903	0.2870
1.5543	14.96	101	1.4671	0.2731
1.5543	16.0	108	1.3992	0.3194
1.4872	16.89	114	1.3854	0.3009
1.4861	17.93	121	1.3411	0.3426
1.4861	18.96	128	1.3142	0.3472
1.4281	20.0	135	1.3021	0.4259
1.38	20.89	141	1.2657	0.4028
1.38	21.93	148	1.2372	0.4352
1.3472	22.96	155	1.2341	0.4815
1.3029	24.0	162	1.1815	0.4306
1.3029	24.89	168	1.1797	0.4954
1.3042	25.93	175	1.1403	0.4583
1.281	26.96	182	1.1349	0.4722
1.281	28.0	189	1.1369	0.4907
1.2614	28.89	195	1.0999	0.4954
1.2133	29.93	202	1.1677	0.4676
1.2133	30.96	209	1.0785	0.5
1.2527	32.0	216	1.1092	0.4861
1.1722	32.89	222	1.0424	0.5185
1.1722	33.93	229	1.0791	0.4907
1.1225	34.96	236	1.0447	0.4907
1.1447	36.0	243	1.0777	0.4583
1.1447	36.89	249	1.0141	0.4954
1.1484	37.93	256	1.0196	0.5324
1.11	38.96	263	0.9791	0.5417
1.046	40.0	270	0.9798	0.5231
1.046	40.89	276	0.9366	0.5694
1.0582	41.93	283	0.9645	0.5602
1.0569	42.96	290	0.9764	0.5694
1.0569	44.0	297	1.0340	0.5324
1.028	44.89	303	0.9969	0.5463
1.04	45.93	310	1.0251	0.5185
1.04	46.96	317	1.0447	0.5417
0.9889	48.0	324	0.9487	0.5324
1.0055	48.89	330	1.0147	0.5
1.0055	49.93	337	1.0015	0.5046
0.9955	50.96	344	0.9763	0.5278
0.9382	52.0	351	1.0306	0.5278
0.9382	52.89	357	0.9970	0.5463
0.9601	53.93	364	0.9487	0.5741
0.9736	54.96	371	0.9658	0.5463
0.9736	56.0	378	0.9789	0.5602
0.9237	56.89	384	0.9940	0.5463
0.9588	57.93	391	0.9778	0.5463
0.9588	58.96	398	0.9789	0.5648
0.9393	60.0	405	0.9612	0.5602
0.9291	60.89	411	0.9141	0.5556
0.9291	61.93	418	0.9770	0.5463
0.929	62.96	425	0.9385	0.5556
0.9448	64.0	432	0.9504	0.5463
0.9448	64.89	438	0.9984	0.5463
0.9426	65.93	445	0.9228	0.5602
0.8949	66.96	452	0.9729	0.5509
0.8949	68.0	459	0.9825	0.5602
0.9041	68.89	465	0.9769	0.5509
0.8828	69.93	472	0.9914	0.5648
0.8828	70.96	479	0.9838	0.5509
0.8874	72.0	486	0.9646	0.5741
0.8723	72.89	492	1.0682	0.5324
0.8723	73.93	499	1.0629	0.5417
0.8953	74.96	506	0.9770	0.5648
0.879	76.0	513	1.0038	0.5787
0.879	76.89	519	1.0529	0.5648
0.896	77.93	526	1.0300	0.5602
0.8519	78.96	533	1.0451	0.5463
0.8414	80.0	540	1.0755	0.5509
0.8414	80.89	546	1.0287	0.5556
0.8342	81.93	553	1.0140	0.5602
0.8653	82.96	560	1.0787	0.5463
0.8653	84.0	567	1.0762	0.5509
0.8357	84.89	573	1.0307	0.5741
0.8455	85.93	580	1.0171	0.5648
0.8455	86.96	587	0.9886	0.5880
0.8238	88.0	594	0.9806	0.5741
0.8613	88.89	600	1.0177	0.5833
0.8613	89.93	607	1.0273	0.5602
0.8265	90.96	614	0.9857	0.5926
0.831	92.0	621	0.9701	0.5972
0.831	92.89	627	0.9726	0.5972
0.8247	93.93	634	0.9765	0.5880
0.8041	94.96	641	0.9801	0.5926
0.8041	96.0	648	0.9796	0.5926
0.8387	96.89	654	0.9790	0.5972
0.7906	97.78	660	0.9783	0.5972

Framework versions

Transformers 4.31.0
Pytorch 2.0.1+cu117
Datasets 2.14.4
Tokenizers 0.13.3

📄 License

This model is licensed under [CC - BY - NC 4.0](https://creativecommons.org/licenses/by - nc/4.0/).

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