2020-12-16 16:42:50 +00:00
|
|
|
import os.path as osp
|
|
|
|
import logging
|
2021-10-27 20:56:16 +00:00
|
|
|
import shutil
|
2020-12-16 16:42:50 +00:00
|
|
|
import time
|
|
|
|
import argparse
|
|
|
|
|
|
|
|
import os
|
|
|
|
|
|
|
|
import utils
|
|
|
|
import utils.options as option
|
|
|
|
import utils.util as util
|
2020-12-18 16:18:34 +00:00
|
|
|
from trainer.ExtensibleTrainer import ExtensibleTrainer
|
2020-12-16 16:42:50 +00:00
|
|
|
from data import create_dataset, create_dataloader
|
|
|
|
from tqdm import tqdm
|
|
|
|
import torch
|
|
|
|
import torchvision
|
|
|
|
|
|
|
|
|
|
|
|
if __name__ == "__main__":
|
|
|
|
#### options
|
|
|
|
torch.backends.cudnn.benchmark = True
|
|
|
|
want_metrics = False
|
|
|
|
parser = argparse.ArgumentParser()
|
2021-10-27 20:56:16 +00:00
|
|
|
parser.add_argument('-opt', type=str, help='Path to options YAML file.', default='../options/test_noisy_audio_clips_classifier.yml')
|
2020-12-16 16:42:50 +00:00
|
|
|
opt = option.parse(parser.parse_args().opt, is_train=False)
|
|
|
|
opt = option.dict_to_nonedict(opt)
|
|
|
|
utils.util.loaded_options = opt
|
|
|
|
|
|
|
|
util.mkdirs(
|
|
|
|
(path for key, path in opt['path'].items()
|
|
|
|
if not key == 'experiments_root' and 'pretrain_model' not in key and 'resume' not in key))
|
|
|
|
util.setup_logger('base', opt['path']['log'], 'test_' + opt['name'], level=logging.INFO,
|
|
|
|
screen=True, tofile=True)
|
|
|
|
logger = logging.getLogger('base')
|
|
|
|
logger.info(option.dict2str(opt))
|
|
|
|
|
|
|
|
#### Create test dataset and dataloader
|
|
|
|
test_loaders = []
|
|
|
|
for phase, dataset_opt in sorted(opt['datasets'].items()):
|
2021-10-27 20:56:16 +00:00
|
|
|
if 'test' in phase:
|
|
|
|
test_set = create_dataset(dataset_opt)
|
|
|
|
if hasattr(test_set, 'wrapped_dataset'):
|
|
|
|
test_set = test_set.wrapped_dataset
|
|
|
|
test_loader = create_dataloader(test_set, dataset_opt, opt)
|
|
|
|
logger.info('Number of test images: {:d}'.format(len(test_set)))
|
|
|
|
test_loaders.append((dataset_opt['name'], test_loader))
|
2020-12-16 16:42:50 +00:00
|
|
|
|
|
|
|
model = ExtensibleTrainer(opt)
|
2021-10-27 20:56:16 +00:00
|
|
|
# Remove all losses, since often labels will not be provided and this is not implied in test.
|
|
|
|
for s in model.steps:
|
|
|
|
s.losses = {}
|
2020-12-16 16:42:50 +00:00
|
|
|
|
2021-10-27 20:56:16 +00:00
|
|
|
output_key = opt['eval']['classifier_logits_key']
|
|
|
|
output_base_dir = opt['eval']['output_dir']
|
|
|
|
labels = opt['eval']['output_labels']
|
|
|
|
path_key = opt['eval']['path_key']
|
|
|
|
|
|
|
|
step = 0
|
|
|
|
for test_set_name, test_loader in test_loaders:
|
2020-12-16 16:42:50 +00:00
|
|
|
test_start_time = time.time()
|
|
|
|
dataset_dir = osp.join(opt['path']['results_root'], opt['name'])
|
|
|
|
util.mkdir(dataset_dir)
|
|
|
|
|
2021-10-27 20:56:16 +00:00
|
|
|
for data in tqdm(test_loader):
|
|
|
|
with torch.no_grad():
|
|
|
|
model.feed_data(data, 0)
|
|
|
|
model.test()
|
|
|
|
|
|
|
|
lbls = torch.nn.functional.softmax(model.eval_state[output_key][0].cpu(), dim=-1)
|
|
|
|
for k, lbl in enumerate(lbls):
|
|
|
|
lbl = torch.argmax(lbl, dim=0)
|
|
|
|
src_path = data[path_key][k]
|
|
|
|
dest = os.path.join(output_base_dir, labels[lbl])
|
|
|
|
os.makedirs(dest, exist_ok=True)
|
|
|
|
shutil.copy(str(src_path), os.path.join(dest, f'{step}_{os.path.basename(str(src_path))}'))
|
|
|
|
step += 1
|