asr_wav2vec2 – AI Model by Myriam123 | AlphaNeural AI

You can deploy this model and start earning money today!

Myriam123/
asr_wav2vec2

like

0

automatic-speech-recognition

generated_from_trainer

facebook/wav2vec2-base

endpoints_compatible

Views

No views yet

asr_wav2vec2

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

Loss: 0.1653
Wer: 0.0776
Cer: 0.0273

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: 32
eval_batch_size: 8
seed: 42
optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
lr_scheduler_type: linear
lr_scheduler_warmup_steps: 800
num_epochs: 7

Training results

Training Loss	Epoch	Step	Validation Loss	Wer	Cer
3.6591	0.32	500	3.1034	1.0	1.0
2.0669	0.64	1000	0.5677	0.4081	0.1235
0.5095	0.96	1500	0.3356	0.2305	0.0715
0.3491	1.28	2000	0.2896	0.1828	0.0587
0.2895	1.6	2500	0.2421	0.1621	0.0529
0.2588	1.92	3000	0.2242	0.1473	0.0487
0.2101	2.24	3500	0.2079	0.1336	0.0439
0.1845	2.56	4000	0.2095	0.1260	0.0419
0.1696	2.88	4500	0.2032	0.1204	0.0399
0.1476	3.2	5000	0.1995	0.1104	0.0378
0.1268	3.52	5500	0.1914	0.1051	0.0359
0.1237	3.84	6000	0.1778	0.0991	0.0331
0.1092	4.16	6500	0.1888	0.0966	0.0328
0.0974	4.48	7000	0.1774	0.0928	0.0315
0.0939	4.8	7500	0.1734	0.0898	0.0305
0.0856	5.12	8000	0.1689	0.0878	0.0297
0.0762	5.44	8500	0.1711	0.0858	0.0297
0.0733	5.76	9000	0.1692	0.0828	0.0287
0.0671	6.08	9500	0.1697	0.0798	0.0278
0.0605	6.4	10000	0.1696	0.0816	0.0283
0.0584	6.72	10500	0.1653	0.0776	0.0273

Framework versions

Transformers 4.38.1
Pytorch 2.1.2
Datasets 2.1.0
Tokenizers 0.15.2