You can deploy this model and start earning money today!

alinerodrigues/
wav2vec2-large-xlsr-mecita-coraa-portuguese-2-all-05

Name: wav2vec2-large-xlsr-mecita-coraa-portuguese-2-all-05
Availability: InStock

transformers

pytorch

wav2vec2

automatic-speech-recognition

generated_from_trainer

apache-2.0

endpoints_compatible

Views

No views yet

wav2vec2-large-xlsr-mecita-coraa-portuguese-2-all-05

This model is a fine-tuned version of Edresson/wav2vec2-large-xlsr-coraa-portuguese on an unknown dataset. It achieves the following results on the evaluation set:

Loss: 0.1720
Wer: 0.0834
Cer: 0.0306

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: 3e-05
train_batch_size: 16
eval_batch_size: 8
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
num_epochs: 100
mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer	Cer
28.6104	1.0	86	8.8298	1.0	0.9860
13.0564	2.0	172	3.2780	1.0	1.0
5.8701	3.0	258	2.9901	1.0	1.0
2.9873	4.0	344	2.9260	1.0	1.0
2.917	5.0	430	2.8919	1.0	1.0
2.8987	6.0	516	2.8910	1.0	1.0
2.8562	7.0	602	2.6200	0.9995	0.9783
2.8562	8.0	688	1.0176	0.9842	0.2639
2.0957	9.0	774	0.4470	0.2097	0.0614
0.912	10.0	860	0.3430	0.1726	0.0529
0.6362	11.0	946	0.2901	0.1516	0.0472
0.5136	12.0	1032	0.2610	0.1331	0.0434
0.4453	13.0	1118	0.2458	0.1266	0.0413
0.3797	14.0	1204	0.2293	0.1188	0.0393
0.3797	15.0	1290	0.2183	0.1081	0.0370
0.3492	16.0	1376	0.2064	0.1009	0.0352
0.313	17.0	1462	0.2002	0.1048	0.0355
0.3066	18.0	1548	0.1981	0.0960	0.0335
0.2891	19.0	1634	0.1971	0.1001	0.0348
0.2731	20.0	1720	0.1954	0.0941	0.0336
0.2452	21.0	1806	0.1911	0.0982	0.0337
0.2452	22.0	1892	0.1884	0.0938	0.0332
0.2641	23.0	1978	0.1865	0.0921	0.0325
0.2558	24.0	2064	0.1820	0.0909	0.0321
0.2471	25.0	2150	0.1823	0.0919	0.0321
0.2213	26.0	2236	0.1784	0.0900	0.0318
0.2083	27.0	2322	0.1809	0.0941	0.0323
0.2063	28.0	2408	0.1789	0.0883	0.0316
0.2063	29.0	2494	0.1795	0.0912	0.0331
0.205	30.0	2580	0.1783	0.0883	0.0325
0.1887	31.0	2666	0.1758	0.0880	0.0321
0.1842	32.0	2752	0.1759	0.0912	0.0322
0.1817	33.0	2838	0.1751	0.0856	0.0308
0.1802	34.0	2924	0.1784	0.0854	0.0310
0.1721	35.0	3010	0.1750	0.0907	0.0315
0.1721	36.0	3096	0.1745	0.0824	0.0305
0.1801	37.0	3182	0.1748	0.0844	0.0312
0.1734	38.0	3268	0.1751	0.0849	0.0308
0.1784	39.0	3354	0.1736	0.0829	0.0309
0.1758	40.0	3440	0.1753	0.0822	0.0305
0.1603	41.0	3526	0.1757	0.0856	0.0308
0.1621	42.0	3612	0.1759	0.0871	0.0310
0.1621	43.0	3698	0.1747	0.0856	0.0311
0.1486	44.0	3784	0.1720	0.0834	0.0306
0.1475	45.0	3870	0.1742	0.0824	0.0300
0.154	46.0	3956	0.1749	0.0820	0.0303
0.1488	47.0	4042	0.1770	0.0851	0.0308
0.1461	48.0	4128	0.1746	0.0841	0.0306
0.1472	49.0	4214	0.1772	0.0863	0.0307
0.1373	50.0	4300	0.1777	0.0868	0.0307
0.1373	51.0	4386	0.1768	0.0849	0.0302
0.1321	52.0	4472	0.1752	0.0839	0.0304
0.1443	53.0	4558	0.1763	0.0844	0.0303
0.1316	54.0	4644	0.1780	0.0861	0.0307
0.1399	55.0	4730	0.1788	0.0834	0.0297
0.1285	56.0	4816	0.1774	0.0815	0.0298
0.1235	57.0	4902	0.1780	0.0846	0.0295
0.1235	58.0	4988	0.1764	0.0854	0.0301
0.1264	59.0	5074	0.1773	0.0846	0.0299
0.1313	60.0	5160	0.1740	0.0832	0.0297
0.1243	61.0	5246	0.1766	0.0820	0.0293
0.1285	62.0	5332	0.1753	0.0839	0.0296
0.1424	63.0	5418	0.1775	0.0844	0.0300
0.1268	64.0	5504	0.1745	0.0805	0.0288

Framework versions

Transformers 4.28.0
Pytorch 2.1.1+cu121
Datasets 2.15.0
Tokenizers 0.13.3

alinerodrigues/wav2vec2-large-xlsr-mecita-coraa-portuguese-2-all-05

wav2vec2-large-xlsr-mecita-coraa-portuguese-2-all-05

Model description

Intended uses & limitations

Training and evaluation data

Training procedure

Training hyperparameters

Training results

Framework versions

alinerodrigues/
wav2vec2-large-xlsr-mecita-coraa-portuguese-2-all-05