-
Notifications
You must be signed in to change notification settings - Fork 2
/
baseline_eval_task2.py
231 lines (185 loc) · 8.11 KB
/
baseline_eval_task2.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
import os
from argparse import ArgumentParser
from concurrent import futures
from pathlib import Path
import musdb
import museval
import numpy as np
import soundfile
import wandb
from tqdm import tqdm
from src.data.musdb_amss_dataset.amss_task2_datasets import task2_config
from src.data.musdb_wrapper import MusdbUnmixedTestSet
from task2_eval import multi_channel_dist
def get_model(argument):
model_name = argument['model']
if model_name == 'demucs':
from separators.demucs_wrapper import Demucs_separator
model_param = argument['model_param']
if model_param is None:
model_param = 'demucs'
if isinstance(model_param, str) and model_param.lower() == "none":
model_param = 'demucs'
return Demucs_separator(model_param)
elif model_name == 'spleeter':
from separators.spleeter_wrapper import Spleeter_separator
return Spleeter_separator(json_path='4stems-16kHz.json')
elif model_name == 'x_umx':
from separators.x_umx_wrapper import X_umx_wrapper
import nnabla as nn
from nnabla.ext_utils import get_extension_context
ctx = get_extension_context('cudnn')
nn.set_default_context(ctx)
nn.set_auto_forward(True)
return X_umx_wrapper()
elif model_name == 'lasaftnet':
model_param = argument['model_param']
from separators.lasaftnet_wrapper import LaSAFT_separator
if model_param is None:
model_param = 'lasaft_large_2020'
return LaSAFT_separator(model_param)
else:
raise ModuleNotFoundError
def eval_amss_track(separated, unmixed_track, amss):
result_dict = {}
_before, _after, _tar_before, _tar_after, _acc = amss.edit_for_test(separated)
result_dict['amss_hat'] = _after
before, after, tar_before, tar_after, acc = amss.edit_for_test(unmixed_track)
result_dict['amss'] = after
return result_dict
def task2_evaluation(_model, _test_set, _unmixed_test_set, _logger, _wandb_logger, _cached=False, _wav_cache=False, _cache_dir=None):
# eval
lr_mode = False
# _cache_dir = Path(_cache_dir)
# separated_results = []
# for i in tqdm(range(len(_test_set))):
# # https://github.com/facebookresearch/demucs/blob/fa7480d5822945e17bf8a16e4baad9f2b631dffc/demucs/test.py#L53
# track = _test_set.tracks[i]
# separated = separate_all(_cache_dir, _cached, _model, _wav_cache, track)
# separated = [separated[s] for s in ['vocals', 'drums', 'bass', 'other']]
# separated = numpy.stack(separated)
# separated_results.append(separated)
for amss in task2_config.evaluation_amss_set:
desc = amss.gen_desc_default()
skip_keyword = ['separate', 'mute', 'reverb']
skip = False
for keyword in skip_keyword:
if keyword in desc:
skip = True
break
if skip:
continue
print(amss)
a_prime_results = []
tar_results = []
acc_results = []
for i in tqdm(range(len(_test_set))):
# https://github.com/facebookresearch/demucs/blob/fa7480d5822945e17bf8a16e4baad9f2b631dffc/demucs/test.py#L53
unmixed_track = _unmixed_test_set[i]
before, after, tar_before, tar_after, acc = amss.edit_for_test(unmixed_track)
separated = separate_all(_cache_dir, _cached, _model, _wav_cache, before)
separated = [separated[s] for s in ['vocals', 'drums', 'bass', 'other']]
separated = np.stack(separated)
_before, _after, _tar_before, _tar_after, _acc = amss.edit_for_test(separated)
result_dict = {'amss_hat':_after, 'amss':after}
a_prime_result = multi_channel_dist(result_dict['amss'], result_dict['amss_hat'])
if logger == 'wandb':
for key in a_prime_result.keys():
if ('left' in key or 'right' in key) and not lr_mode:
continue
wandb_logger.log({'a_prime/{}_{}'.format(desc, key): a_prime_result[key]})
a_prime_results.append(
np.array([a_prime_result['mae'],
a_prime_result['mae_left'],
a_prime_result['mae_right'],
a_prime_result['mfcc_rmse'],
a_prime_result['mfcc_rmse_left'],
a_prime_result['mfcc_rmse_right']])
)
if logger == 'wandb':
scores = np.mean(np.stack(a_prime_results), axis=0)
wandb_logger.log({'agg_mid/a_prime_mae_{}'.format(desc): scores[0]})
wandb_logger.log({'agg_left/a_prime_mae_{}'.format(desc): scores[1]})
wandb_logger.log({'agg_right/a_prime_mae_{}'.format(desc): scores[2]})
wandb_logger.log({'agg_mid/a_prime_mfccrmse_{}'.format(desc): scores[3]})
wandb_logger.log({'agg_left/a_prime_mfccrmse_{}'.format(desc): scores[4]})
wandb_logger.log({'agg_right/a_prime_mfccrmse_{}'.format(desc): scores[5]})
else:
scores = np.mean(np.stack(a_prime_results), axis=0)
print({'agg_mid/a_prime_mae_{}'.format(desc): scores[0]})
print({'agg_left/a_prime_mae_{}'.format(desc): scores[1]})
print({'agg_right/a_prime_mae_{}'.format(desc): scores[2]})
print({'agg_mid/a_prime_mfccrmse_{}'.format(desc): scores[3]})
print({'agg_left/a_prime_mfccrmse_{}'.format(desc): scores[4]})
print({'agg_right/a_prime_mfccrmse_{}'.format(desc): scores[5]})
def separate_all(_cache_dir, _cached, _model, _wav_cache, track):
if _cached:
estimated = {source: soundfile.read(_cache_dir.joinpath('{}_{}.wav'.format(track.name, source)))[0]
for source
in ['vocals', 'drums', 'bass', 'other']}
else:
estimated = _model.separate(track)
if _wav_cache and not _cached:
for source in estimated.keys():
soundfile.write(
_cache_dir.joinpath('{}_{}.wav'.format(track.name, source)),
estimated[source],
44100
)
return estimated
if __name__ == '__main__':
parser = ArgumentParser()
parser.add_argument('--model', type=str)
parser.add_argument('--model_param', type=str)
parser.add_argument('--musdb_root', type=str)
parser.add_argument('--log', type=str)
parser.add_argument('--wav_cache', type=bool, default=False)
args = parser.parse_args().__dict__
model_param = args['model_param']
# # wav_cache
# wav_cache = args['wav_cache']
#
# # check cached.
# cache_dir = 'etc/result_cached/{}_{}'.format(args['model'], model_param)
#
# if os.path.exists(cache_dir):
# cached = True
# if cached and wav_cache:
# print('please remove the existing directory for cache.')
# raise RuntimeError
#
# print('result cached')
# else:
# cached = False
# if wav_cache:
# if not os.path.exists('etc'):
# os.mkdir('etc')
# if not os.path.exists('etc/result_cached'):
# os.mkdir('etc/result_cached')
# if not os.path.exists(cache_dir):
# os.mkdir(cache_dir)
# model
# model = None if cached else get_model(args)
model = get_model(args)
if model_param is None:
model_param = "None"
# dataset
musdb_path = args['musdb_root']
unmixed_test_set = MusdbUnmixedTestSet(musdb_path)
test_set = musdb.DB(musdb_path, is_wav=True, subsets=["test"])
assert len(unmixed_test_set) == 50
assert len(test_set) == 50
# logger
logger = args['log']
assert logger in ['wandb', None]
if logger == 'wandb':
wandb_logger = wandb.init(job_type='eval', config=args, project='task2_eval_mm', tags=[args['model']],
name='{}_{}'.format(args['model'], args['model_param']))
else:
wandb_logger = None
# try:
task2_evaluation(model, test_set, unmixed_test_set, logger, wandb_logger)
# except Exception as ex:
# print(ex)
# if wav_cache:
# os.rmdir(cache_dir)