SELD

Download Dataset from https://zenodo.org/record/4844825#.YNHbL2gzaUk
Preprocess Data with batch_feature_extracion.py in https://github.com/sharathadavanne/seld-dcase2021.
run pip install -r requirements.txt
run train.py with arguments. You should set 'abspath' to the directory preprocessed datasets are located.

we reported the scores in the paper by using the Neural Architecture Search (NAS) method. Thus, you can't get the same scores in the paper.

Name		Name	Last commit message	Last commit date
Latest commit History 422 Commits
config		config
model_config		model_config
.gitignore		.gitignore
README.md		README.md
SELD_evaluation_metrics.py		SELD_evaluation_metrics.py
analyzer.py		analyzer.py
complexity.py		complexity.py
complexity_test.py		complexity_test.py
config_manager.py		config_manager.py
config_sampler.py		config_sampler.py
config_sampler_test.py		config_sampler_test.py
data_loader.py		data_loader.py
data_loader_test.py		data_loader_test.py
data_utils.py		data_utils.py
evaluator.py		evaluator.py
feature_extractor.py		feature_extractor.py
feature_extractor_test.py		feature_extractor_test.py
layers.py		layers.py
layers_test.py		layers_test.py
losses.py		losses.py
make_answer.py		make_answer.py
metrics.py		metrics.py
model_complexity.py		model_complexity.py
model_complexity_test.py		model_complexity_test.py
models.py		models.py
models_test.py		models_test.py
modules.py		modules.py
modules_test.py		modules_test.py
nas_seldnet.py		nas_seldnet.py
nas_vad.py		nas_vad.py
params.py		params.py
plot_overall.py		plot_overall.py
plot_results.py		plot_results.py
requirements.txt		requirements.txt
result_merge.py		result_merge.py
search_best.py		search_best.py
single_class.py		single_class.py
stage_complexity.py		stage_complexity.py
stage_complexity_test.py		stage_complexity_test.py
swa.py		swa.py
train.py		train.py
train_vad_baseline.py		train_vad_baseline.py
trainv2.py		trainv2.py
transforms.py		transforms.py
transforms_test.py		transforms_test.py
utils.py		utils.py
utils_test.py		utils_test.py
vad_dataloader.py		vad_dataloader.py

Provide feedback