def main():
# get config
config = load_cfg(None)
# get vars from config
dataset_name = config["experiment"]["dataset"]
dataset_subject_count = config["data"][dataset_name]["n_subjects"]
experiment_type = config["experiment"]["type"]
experiment_n_folds = config["experiment"]["n_folds"]
model_name = config["model"]["name"]
# path to save csv
now = (
str(datetime.datetime.now())
.replace("-", "_")
.replace(" ", "_")
.replace(".", "_")
.replace(":", "_")
)
results_path = f"/home/no316758/results/{now}_{experiment_type}_{model_name}_{dataset_name}_{experiment_n_folds}/"
if not os.path.exists(results_path):
os.makedirs(results_path)
# copy config
exp_cfg_path = f"{results_path}config.yaml"
with open(exp_cfg_path, "w") as outfile:
yaml.dump(config, outfile, default_flow_style=False)
script_name = "_single_slurm_run.sh"
script_path = os.path.abspath(os.path.join(os.path.dirname(__file__), script_name))
p = subprocess.Popen(
f"sbatch -o {results_path}sbatch_stdout_%j.txt {script_path} {results_path}",
shell=True,
)
p.wait()
def main(args):
# Load config file
config = load_cfg(args)
# Set subject id and valid fold
subject_id = config["experiment"]["subject_id"]
i_valid_fold = config["experiment"]["i_valid_fold"]
if config["server"]["full_cv"]:
full_cv_all_subjects(config)
else:
# Run single subject single fold:
single_subject_single_fold(subject_id, i_valid_fold, config)
if len(sys.argv) > 1:
results_dir_path = sys.argv[1]
logging.basicConfig(
format="%(asctime)s %(levelname)s : %(message)s",
level=logging.DEBUG,
stream=sys.stdout,
)
# Should contain both .gdf files and .mat-labelfiles from competition
# data_folder = "/home/no316758/data/BCICIV_2a_gdf/"
# data_folder = '/Users/sebas/code/_eeg_data/BCICIV_2a_gdf/'
# low_cut_hz = 4 # 0 or 4
cuda = True
# get config
config = load_cfg(None, cfg_path=f"{results_dir_path}config.yaml")
experiment_type = config["experiment"]["type"]
# get vars from config
dataset_name = config["experiment"]["dataset"]
dataset_subject_count = config["data"][dataset_name]["n_subjects"]
dataset_path = config["data"][dataset_name]["proc_path"]
dataset_n_classes = config["data"][dataset_name]["n_classes"]
experiment_type = config["experiment"]["type"]
experiment_n_folds = config["experiment"]["n_folds"]
experiment_i_valid_fold = config["experiment"]["i_valid_fold"]
model_name = config["model"]["name"] # 'eegnet' or 'shallow' or 'deep'
experiment_max_epochs = config["train"]["max_epochs"]
experiment_max_increase_epochs = config["train"]["early_stop_patience"]
experiment_batch_size = config["train"]["batch_size"]