w2v2-base-pretrained_lr5e-5_at0.8_da0.9 – AI Model by MelanieKoe | AlphaNeural AI

You can deploy this model and start earning money today!

MelanieKoe/
w2v2-base-pretrained_lr5e-5_at0.8_da0.9

like

0

automatic-speech-recognition

generated_from_trainer

facebook/wav2vec2-base

endpoints_compatible

Views

No views yet

w2v2-base-pretrained_lr5e-5_at0.8_da0.9

This model is a fine-tuned version of facebook/wav2vec2-base on the None dataset. It achieves the following results on the evaluation set:

Loss: 2.3151
Wer: 0.1807

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: 5e-05
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: 1000
training_steps: 4000
mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
18.3155	6.1	250	3.8850	1.0
3.2773	12.2	500	3.1874	1.0
3.0747	18.29	750	3.0519	1.0
1.8747	24.39	1000	1.0181	0.5229
0.2774	30.49	1250	1.5081	0.2191
0.1428	36.59	1500	1.5603	0.2067
0.0989	42.68	1750	1.6293	0.1999
0.0748	48.78	2000	1.9993	0.1978
0.0584	54.88	2250	2.0658	0.1914
0.0481	60.98	2500	2.1470	0.1845
0.0433	67.07	2750	2.2501	0.1833
0.038	73.17	3000	2.1957	0.1892
0.0331	79.27	3250	2.2906	0.1845
0.0314	85.37	3500	2.3224	0.1841
0.0279	91.46	3750	2.3041	0.1828
0.0267	97.56	4000	2.3151	0.1807

Framework versions

Transformers 4.35.0
Pytorch 2.0.0
Datasets 2.14.6
Tokenizers 0.14.1