You can deploy this model and start earning money today!

alinerodrigues/
wav2vec2-large-xlsr-mecita-coraa-portuguese-2-all-clean-07

Name: wav2vec2-large-xlsr-mecita-coraa-portuguese-2-all-clean-07
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-07

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: 0.1209
Wer: 0.0903
Cer: 0.0256

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
27.4046	1.0	67	3.3438	1.0	1.0
6.9713	2.0	134	2.9841	1.0	1.0
3.0045	3.0	201	2.9144	1.0	1.0
3.0045	4.0	268	2.8956	1.0	1.0
2.9124	5.0	335	2.7397	1.0	1.0
2.5598	6.0	402	1.2712	0.8962	0.2337
2.5598	7.0	469	0.5554	0.2947	0.0751
1.1631	8.0	536	0.3916	0.1968	0.0531
0.6927	9.0	603	0.3124	0.1737	0.0463
0.6927	10.0	670	0.2679	0.1672	0.0433
0.5782	11.0	737	0.2395	0.1537	0.0405
0.4523	12.0	804	0.2139	0.1365	0.0372
0.4523	13.0	871	0.2009	0.1275	0.0354
0.3903	14.0	938	0.1852	0.1138	0.0317
0.3522	15.0	1005	0.1734	0.1165	0.0319
0.3522	16.0	1072	0.1669	0.1089	0.0294
0.3397	17.0	1139	0.1607	0.1058	0.0296
0.2904	18.0	1206	0.1536	0.0969	0.0277
0.2904	19.0	1273	0.1503	0.0979	0.0277
0.2733	20.0	1340	0.1472	0.0938	0.0275
0.2698	21.0	1407	0.1417	0.0948	0.0273
0.2698	22.0	1474	0.1403	0.0903	0.0266
0.2439	23.0	1541	0.1368	0.0965	0.0262
0.2348	24.0	1608	0.1364	0.0927	0.0263
0.2348	25.0	1675	0.1333	0.0920	0.0263
0.2241	26.0	1742	0.1342	0.0924	0.0267
0.2174	27.0	1809	0.1336	0.0931	0.0266
0.2174	28.0	1876	0.1352	0.0910	0.0266
0.2325	29.0	1943	0.1315	0.0927	0.0260
0.1873	30.0	2010	0.1298	0.0924	0.0260
0.1873	31.0	2077	0.1255	0.0865	0.0248
0.1926	32.0	2144	0.1251	0.0876	0.0253
0.1934	33.0	2211	0.1274	0.0879	0.0253
0.1934	34.0	2278	0.1209	0.0903	0.0256
0.2157	35.0	2345	0.1233	0.0879	0.0253
0.1855	36.0	2412	0.1242	0.0900	0.0252
0.1855	37.0	2479	0.1210	0.0851	0.0240
0.1644	38.0	2546	0.1260	0.0841	0.0241
0.1666	39.0	2613	0.1243	0.0862	0.0242
0.1666	40.0	2680	0.1254	0.0855	0.0247
0.1883	41.0	2747	0.1233	0.0834	0.0246
0.175	42.0	2814	0.1253	0.0831	0.0242
0.175	43.0	2881	0.1239	0.0827	0.0241
0.1763	44.0	2948	0.1250	0.0838	0.0242
0.1615	45.0	3015	0.1226	0.0872	0.0245
0.1615	46.0	3082	0.1235	0.0879	0.0247
0.1544	47.0	3149	0.1227	0.0858	0.0246
0.1554	48.0	3216	0.1217	0.0862	0.0246
0.1554	49.0	3283	0.1247	0.0876	0.0248
0.1476	50.0	3350	0.1221	0.0855	0.0246
0.1649	51.0	3417	0.1237	0.0855	0.0238
0.1649	52.0	3484	0.1244	0.0841	0.0242
0.144	53.0	3551	0.1246	0.0858	0.0244
0.1434	54.0	3618	0.1241	0.0831	0.0235

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-07

wav2vec2-large-xlsr-mecita-coraa-portuguese-2-all-clean-07

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-07