metadata

license: apache-2.0
base_model: facebook/wav2vec2-base
tags:
  - generated_from_trainer
metrics:
  - wer
model-index:
  - name: asr_wav2vec2_base_1.4
    results: []

asr_wav2vec2_base_1.4

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: 0.2687
Wer: 0.2410
Cer: 0.0769

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: 16
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: 500
num_epochs: 5

Training results

Training Loss	Epoch	Step	Validation Loss	Wer	Cer
4.4108	1.12	700	1.2046	0.7513	0.3423
0.9111	2.24	1400	0.4008	0.3362	0.1116
0.5359	3.36	2100	0.2947	0.2700	0.0857
0.397	4.48	2800	0.2687	0.2410	0.0769

Framework versions

Transformers 4.39.3
Pytorch 2.1.2
Datasets 2.18.0
Tokenizers 0.15.2