wav2vec2-5 – AI Model by chrisvinsen | AlphaNeural AI

You can deploy this model and start earning money today!

chrisvinsen/
wav2vec2-5

like

0

automatic-speech-recognition

generated_from_trainer

endpoints_compatible

Views

No views yet

wav2vec2-5

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: 3.0700
Wer: 1.0

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: 0.003
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: 400
num_epochs: 30

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
3.4082	1.37	200	3.3181	1.0
2.8798	2.74	400	2.9921	1.0
2.8703	4.11	600	3.1937	1.0
2.8643	5.48	800	3.0304	1.0
2.8655	6.85	1000	3.0321	1.0
2.8655	8.22	1200	3.0716	1.0
2.863	9.59	1400	3.1764	1.0
2.8567	10.96	1600	3.0600	1.0
2.861	12.33	1800	3.1761	1.0
2.8606	13.7	2000	3.1028	1.0
2.8613	15.07	2200	3.2119	1.0
2.8612	16.44	2400	3.1158	1.0
2.8603	17.81	2600	3.1230	1.0
2.8601	19.18	2800	3.0380	1.0
2.856	20.55	3000	3.0729	1.0
2.8557	21.92	3200	3.0511	1.0
2.8556	23.29	3400	3.0710	1.0
2.8552	24.66	3600	3.1364	1.0
2.8574	26.03	3800	3.0104	1.0
2.8543	27.4	4000	3.1068	1.0
2.8558	28.77	4200	3.0700	1.0

Framework versions

Transformers 4.19.2
Pytorch 1.11.0+cu113
Datasets 2.2.2
Tokenizers 0.12.1