wav2vec2-vi5

qbaro/wav2vec2-vi5

wav2vec2-vi5 is a automatic speech recognition model from qbaro. Use it when you need speech turned into text. It is set up for transformers. The card lists the license as cc-by-nc-4.0.

should probably proofread and complete it, then remove this comment. --

Downloads · 30 days

2% of all-time downloads

All-time downloads

380

Public

Repo size

52.6 GB

Likes

Public

Hugging Face

Repo makeup

Click a slice to open those files.

.bin1.3 GB · 100%

At a glance

Task: Automatic Speech Recognition
Library: transformers
License: cc-by-nc-4.0
Model type: wav2vec2
Access: Public
Created: Jan 3, 2023
Updated: Jan 31, 2023
SHA: b16c4bb4

Model card

From the Hugging Face model README

wav2vec2-vi5

This model is a fine-tuned version of nguyenvulebinh/wav2vec2-large-vi-vlsp2020 on an unknown dataset. It achieves the following results on the evaluation set:

Loss: 0.1632
Wer: 0.0714

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

learning_rate: 0.0001
train_batch_size: 1
eval_batch_size: 8
seed: 42
gradient_accumulation_steps: 4
total_train_batch_size: 4
optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
lr_scheduler_type: linear
lr_scheduler_warmup_steps: 500
num_epochs: 5

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
0.1279	0.13	100	0.1495	0.0691
0.1067	0.27	200	0.1407	0.0656
0.1013	0.4	300	0.1352	0.0654
0.1282	0.53	400	0.1468	0.0705
0.1117	0.67	500	0.1658	0.0804
0.1173	0.8	600	0.1485	0.0753
0.1	0.94	700	0.1418	0.0699
0.1029	1.07	800	0.1523	0.0715
0.0823	1.2	900	0.1597	0.0731
0.0893	1.34	1000	0.1441	0.0688
0.0822	1.47	1100	0.1580	0.0711
0.1077	1.6	1200	0.1664	0.0832
0.0988	1.74	1300	0.1626	0.0744
0.0777	1.87	1400	0.1498	0.0705
0.0849	2.01	1500	0.1581	0.0711
0.0647	2.14	1600	0.1636	0.0729
0.0638	2.27	1700	0.1590	0.0750
0.0674	2.41	1800	0.1627	0.0755
0.0644	2.54	1900	0.1576	0.0746
0.0666	2.67	2000	0.1569	0.0704
0.0647	2.81	2100	0.1601	0.0693
0.0769	2.94	2200	0.1595	0.0680
0.0585	3.07	2300	0.1591	0.0718
0.0515	3.21	2400	0.1628	0.0714
0.068	3.34	2500	0.1549	0.0704
0.0473	3.48	2600	0.1618	0.0694
0.0545	3.61	2700	0.1539	0.0685
0.049	3.74	2800	0.1546	0.0698
0.0538	3.88	2900	0.1507	0.0698
0.0452	4.01	3000	0.1516	0.0702
0.0398	4.14	3100	0.1604	0.0723
0.0539	4.28	3200	0.1587	0.0713
0.0443	4.41	3300	0.1607	0.0718
0.0405	4.55	3400	0.1626	0.0709
0.038	4.68	3500	0.1623	0.0707
0.0388	4.81	3600	0.1621	0.0726
0.0418	4.95	3700	0.1632	0.0714

Framework versions

Transformers 4.20.0
Pytorch 1.10.0+cu113
Datasets 1.18.3
Tokenizers 0.12.1

Task: Automatic Speech Recognition
Library: transformers
Type: wav2vec2
License: cc-by-nc-4.0
Created: Jan 3, 2023
Updated: Jan 31, 2023