GitHub - Conscht/DCASE-CNN-Transformer: DCASE 2017 Challange CNN-Transformer for the audio classification task. Model reached an accurarcy of 57% but huge improvement could be achieved by changing the MEL-Spec preprocessing.

Branches Tags

Name		Name	Last commit message	Last commit date
Latest commit History 1 Commit
audiotest		audiotest
src		src
.gitignore		.gitignore
audiotest.py		audiotest.py
changes.txt		changes.txt
config_data.yml		config_data.yml
config_gan.py		config_gan.py
config_mlp.yml		config_mlp.yml
config_run_mlp.yml		config_run_mlp.yml
config_transformer.yml		config_transformer.yml
dataset_stats.py		dataset_stats.py
predictions.csv		predictions.csv
run_mlp.py		run_mlp.py
run_transformer.py		run_transformer.py
tensorboard.txt		tensorboard.txt
test_transformer.py		test_transformer.py
train_gan.py		train_gan.py
train_mlp.py		train_mlp.py
train_transformer.py		train_transformer.py

About

DCASE 2017 Challange CNN-Transformer for the audio classification task. Model reached an accurarcy of 57% but huge improvement could be achieved by changing the MEL-Spec preprocessing.