You can deploy this model and start earning money today!

alinerodrigues/
wav2vec2-large-xlsr-mecita-coraa-portuguese-2-all-clean-09

Name: wav2vec2-large-xlsr-mecita-coraa-portuguese-2-all-clean-09
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-clean-09

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: 0.1435
Wer: 0.0935
Cer: 0.0274

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
28.9568	1.0	67	4.5342	1.0	1.0
9.238	2.0	134	3.3773	0.9615	0.9278
3.8131	3.0	201	3.2552	0.9713	0.9672
3.8131	4.0	268	2.9818	0.9825	0.9845
3.7154	5.0	335	2.9048	1.0	1.0
3.2775	6.0	402	2.9244	1.0	1.0
3.2775	7.0	469	2.8494	1.0	1.0
3.295	8.0	536	2.8737	0.9829	0.9853
3.4359	9.0	603	2.8125	0.9648	0.9515
3.4359	10.0	670	2.8296	0.9704	0.9416
3.1631	11.0	737	2.7952	0.9690	0.9126
3.4425	12.0	804	2.7936	0.9746	0.9
3.4425	13.0	871	2.8312	0.9783	0.8831
3.1405	14.0	938	2.7702	0.9717	0.8519
3.1256	15.0	1005	2.7649	0.9769	0.8511
3.1256	16.0	1072	2.7452	0.9763	0.8156
3.3974	17.0	1139	2.7520	0.9769	0.8255
3.378	18.0	1206	2.7433	0.9835	0.7887
3.378	19.0	1273	2.7587	0.9845	0.7957
3.0588	20.0	1340	2.6809	0.9773	0.7508
3.2909	21.0	1407	2.6814	0.9812	0.7634
3.2909	22.0	1474	2.6797	0.9806	0.7662
3.2279	23.0	1541	2.6761	0.9710	0.7498
3.0068	24.0	1608	2.6769	0.9727	0.7580
3.0068	25.0	1675	2.6543	0.9743	0.7394
3.1098	26.0	1742	2.6652	0.9806	0.7654
3.2015	27.0	1809	2.6263	0.9707	0.7314
3.2015	28.0	1876	2.6353	0.9694	0.7274
2.9707	29.0	1943	2.5955	0.9727	0.7129
3.116	30.0	2010	2.5964	0.9677	0.7270
3.116	31.0	2077	2.5873	0.9704	0.7185
3.0265	32.0	2144	2.5629	0.9713	0.7012
3.0762	33.0	2211	2.5520	0.9717	0.6991
3.0762	34.0	2278	2.5075	0.9733	0.6937
3.0228	35.0	2345	2.5200	0.9707	0.7011
2.8856	36.0	2412	2.4967	0.9700	0.6913
2.8856	37.0	2479	2.5047	0.9717	0.7040
2.8122	38.0	2546	2.4714	0.9717	0.6900
2.9134	39.0	2613	2.4488	0.9667	0.6871
2.9134	40.0	2680	2.4372	0.9621	0.6826
2.7965	41.0	2747	2.4174	0.9661	0.6831
2.8698	42.0	2814	2.3973	0.9727	0.6733
2.8698	43.0	2881	2.3632	0.9681	0.6703
2.6243	44.0	2948	2.3325	0.9641	0.6784
2.7087	45.0	3015	2.3252	0.9730	0.6596
2.7087	46.0	3082	2.2806	0.9651	0.6643
2.6011	47.0	3149	2.2519	0.9592	0.6458
2.6838	48.0	3216	2.2211	0.9532	0.6363
2.6838	49.0	3283	2.1794	0.9433	0.6468
2.5863	50.0	3350	2.1214	0.9397	0.6294
2.5353	51.0	3417	2.0826	0.9499	0.6668
2.5353	52.0	3484	1.9885	0.9351	0.6165
2.3509	53.0	3551	1.8180	0.9104	0.5729
2.1853	54.0	3618	1.6297	0.8762	0.5067
2.1853	55.0	3685	1.3421	0.8603	0.4185
1.9876	56.0	3752	1.0128	0.7777	0.2953
1.5594	57.0	3819	0.7699	0.6663	0.2079
1.5594	58.0	3886	0.5548	0.4792	0.1294
1.1199	59.0	3953	0.4366	0.3768	0.0975
0.8571	60.0	4020	0.3568	0.3202	0.0805
0.8571	61.0	4087	0.3107	0.2556	0.0672
0.7263	62.0	4154	0.2792	0.2118	0.0573
0.5792	63.0	4221	0.2560	0.1838	0.0511
0.5792	64.0	4288	0.2452	0.1742	0.0494
0.5726	65.0	4355	0.2277	0.1624	0.0442
0.4892	66.0	4422	0.2126	0.1522	0.0416
0.4892	67.0	4489	0.2075	0.1456	0.0396
0.4639	68.0	4556	0.1982	0.1406	0.0382
0.435	69.0	4623	0.1916	0.1321	0.0365
0.435	70.0	4690	0.1888	0.1281	0.0354
0.4061	71.0	4757	0.1857	0.1215	0.0338
0.383	72.0	4824	0.1800	0.1179	0.0324
0.383	73.0	4891	0.1741	0.1182	0.0322
0.403	74.0	4958	0.1727	0.1159	0.0310
0.3688	75.0	5025	0.1711	0.1140	0.0312
0.3688	76.0	5092	0.1676	0.1126	0.0312
0.3522	77.0	5159	0.1645	0.1097	0.0305
0.3235	78.0	5226	0.1606	0.1094	0.0300
0.3235	79.0	5293	0.1615	0.1061	0.0302
0.3303	80.0	5360	0.1585	0.1041	0.0295
0.319	81.0	5427	0.1567	0.1061	0.0291
0.319	82.0	5494	0.1554	0.1038	0.0294
0.3257	83.0	5561	0.1557	0.1054	0.0296
0.3079	84.0	5628	0.1548	0.1005	0.0286
0.3079	85.0	5695	0.1512	0.1018	0.0289
0.302	86.0	5762	0.1513	0.1001	0.0286
0.2919	87.0	5829	0.1493	0.0965	0.0281
0.2919	88.0	5896	0.1479	0.0968	0.0283
0.2783	89.0	5963	0.1460	0.0968	0.0278
0.2862	90.0	6030	0.1453	0.0975	0.0280
0.2862	91.0	6097	0.1466	0.0942	0.0279
0.276	92.0	6164	0.1458	0.0958	0.0275
0.2854	93.0	6231	0.1462	0.0945	0.0273
0.2854	94.0	6298	0.1446	0.0935	0.0275
0.2887	95.0	6365	0.1445	0.0935	0.0275
0.2907	96.0	6432	0.1435	0.0935	0.0274
0.2907	97.0	6499	0.1450	0.0949	0.0272
0.298	98.0	6566	0.1454	0.0935	0.0272
0.2974	99.0	6633	0.1450	0.0939	0.0273
0.304	100.0	6700	0.1447	0.0935	0.0271

Framework versions

Transformers 4.28.0
Pytorch 2.2.1+cu121
Datasets 2.17.0
Tokenizers 0.13.3

alinerodrigues/wav2vec2-large-xlsr-mecita-coraa-portuguese-2-all-clean-09

wav2vec2-large-xlsr-mecita-coraa-portuguese-2-all-clean-09

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-clean-09