w2v2-base-pretrained_lr5e-5_at0.8_da0.1 – AI Model by MelanieKoe | AlphaNeural AI

You can deploy this model and start earning money today!

MelanieKoe/
w2v2-base-pretrained_lr5e-5_at0.8_da0.1

like

0

automatic-speech-recognition

generated_from_trainer

facebook/wav2vec2-base

endpoints_compatible

Views

No views yet

w2v2-base-pretrained_lr5e-5_at0.8_da0.1

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: 4.7430
Wer: 0.8847

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
32.6282	50.0	250	4.0044	1.0
3.5474	100.0	500	3.1597	1.0
3.0565	150.0	750	3.0680	1.0
2.8909	200.0	1000	3.0005	1.0
2.292	250.0	1250	2.6442	0.9991
1.2329	300.0	1500	2.9696	0.9577
0.6684	350.0	1750	3.4034	0.9453
0.4377	400.0	2000	3.7834	0.9389
0.3292	450.0	2250	3.9480	0.9000
0.2554	500.0	2500	4.3241	0.8906
0.2073	550.0	2750	4.4139	0.9047
0.1762	600.0	3000	4.5423	0.9000
0.1551	650.0	3250	4.5978	0.8817
0.1403	700.0	3500	4.6304	0.8757
0.1286	750.0	3750	4.6839	0.8855
0.1201	800.0	4000	4.7430	0.8847

Framework versions

Transformers 4.35.0
Pytorch 2.0.0
Datasets 2.14.6
Tokenizers 0.14.1