add: mean and std
Browse files- musdb18.py +46 -21
musdb18.py
CHANGED
@@ -36,50 +36,73 @@ class Musdb18Dataset(datasets.GeneratorBasedBuilder):
|
|
36 |
def _info(self):
|
37 |
return datasets.DatasetInfo(
|
38 |
description=_DESCRIPTION,
|
39 |
-
features=datasets.Features(
|
40 |
-
|
41 |
-
|
42 |
-
|
43 |
-
|
44 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
45 |
)
|
46 |
|
47 |
def _split_generators(self, dl_manager):
|
48 |
#! you must have your folder locally!
|
49 |
archive_path = dl_manager.download_and_extract(
|
50 |
-
"https://zenodo.org/record/1117372/files/musdb18.zip?download=1"
|
|
|
51 |
|
52 |
return [
|
53 |
datasets.SplitGenerator(
|
54 |
name=datasets.Split.TRAIN,
|
55 |
-
gen_kwargs={
|
56 |
-
"audio_path": f"{archive_path}/train"}
|
57 |
),
|
58 |
datasets.SplitGenerator(
|
59 |
name=datasets.Split.TEST,
|
60 |
-
gen_kwargs={
|
61 |
-
|
62 |
-
}
|
63 |
-
)
|
64 |
]
|
65 |
|
66 |
def _generate_stem_dict(self, S, song_name, end):
|
67 |
-
return {
|
68 |
-
|
69 |
-
|
|
|
|
|
|
|
|
|
|
|
70 |
|
71 |
def _generate_examples(self, audio_path):
|
72 |
id_ = 0
|
73 |
for stems_path in Path(audio_path).iterdir():
|
74 |
song_name = stems_path.stem
|
75 |
S, sr = stempeg.read_stems(
|
76 |
-
str(stems_path), dtype=np.float32, multiprocess=False
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
77 |
|
78 |
-
for idx, end in enumerate(
|
|
|
|
|
79 |
yield id_, {
|
80 |
"name": song_name,
|
81 |
"n_window": idx,
|
82 |
-
**self._generate_stem_dict(S, song_name, end)
|
|
|
|
|
83 |
}
|
84 |
|
85 |
id_ += 1
|
@@ -87,8 +110,10 @@ class Musdb18Dataset(datasets.GeneratorBasedBuilder):
|
|
87 |
# It's very rare for song to have exactly 3 minutes
|
88 |
yield id_, {
|
89 |
"name": song_name,
|
90 |
-
"n_window": idx+1,
|
91 |
-
**self._generate_stem_dict(S, song_name, end=S.shape[1])
|
|
|
|
|
92 |
}
|
93 |
|
94 |
id_ += 1
|
|
|
36 |
def _info(self):
|
37 |
return datasets.DatasetInfo(
|
38 |
description=_DESCRIPTION,
|
39 |
+
features=datasets.Features(
|
40 |
+
{
|
41 |
+
"name": datasets.Value("string"),
|
42 |
+
"n_window": datasets.Value("int16"),
|
43 |
+
**{
|
44 |
+
name: datasets.Audio(
|
45 |
+
sampling_rate=self.SAMPLING_RATE, mono=False
|
46 |
+
)
|
47 |
+
for name in self.INSTRUMENT_NAMES
|
48 |
+
},
|
49 |
+
"mean": datasets.Value("float"),
|
50 |
+
"std": datasets.Value("float"),
|
51 |
+
}
|
52 |
+
),
|
53 |
)
|
54 |
|
55 |
def _split_generators(self, dl_manager):
|
56 |
#! you must have your folder locally!
|
57 |
archive_path = dl_manager.download_and_extract(
|
58 |
+
"https://zenodo.org/record/1117372/files/musdb18.zip?download=1"
|
59 |
+
)
|
60 |
|
61 |
return [
|
62 |
datasets.SplitGenerator(
|
63 |
name=datasets.Split.TRAIN,
|
64 |
+
gen_kwargs={"audio_path": f"{archive_path}/train"},
|
|
|
65 |
),
|
66 |
datasets.SplitGenerator(
|
67 |
name=datasets.Split.TEST,
|
68 |
+
gen_kwargs={"audio_path": f"{archive_path}/test"},
|
69 |
+
),
|
|
|
|
|
70 |
]
|
71 |
|
72 |
def _generate_stem_dict(self, S, song_name, end):
|
73 |
+
return {
|
74 |
+
name: {
|
75 |
+
"path": f"{song_name}/{name}",
|
76 |
+
"array": S[i, end - self.WINDOW_SIZE : end, :],
|
77 |
+
"sampling_rate": self.SAMPLING_RATE,
|
78 |
+
}
|
79 |
+
for i, name in enumerate(self.INSTRUMENT_NAMES)
|
80 |
+
}
|
81 |
|
82 |
def _generate_examples(self, audio_path):
|
83 |
id_ = 0
|
84 |
for stems_path in Path(audio_path).iterdir():
|
85 |
song_name = stems_path.stem
|
86 |
S, sr = stempeg.read_stems(
|
87 |
+
str(stems_path), dtype=np.float32, multiprocess=False
|
88 |
+
)
|
89 |
+
|
90 |
+
mixture = S.sum(axis=0).T
|
91 |
+
assert mixture.shape[0] == 2
|
92 |
+
# from (n_instr, n_chann, n_samp) -> (n_chann, n_samp)
|
93 |
+
mixture = mixture.mean(0) # channel_wise mean -> (n_samples,)
|
94 |
+
mean = mixture.mean().item()
|
95 |
+
std = mixture.std().item()
|
96 |
|
97 |
+
for idx, end in enumerate(
|
98 |
+
range(self.WINDOW_SIZE, S.shape[1], self.WINDOW_SIZE)
|
99 |
+
):
|
100 |
yield id_, {
|
101 |
"name": song_name,
|
102 |
"n_window": idx,
|
103 |
+
**self._generate_stem_dict(S, song_name, end),
|
104 |
+
"mean": mean,
|
105 |
+
"std": std,
|
106 |
}
|
107 |
|
108 |
id_ += 1
|
|
|
110 |
# It's very rare for song to have exactly 3 minutes
|
111 |
yield id_, {
|
112 |
"name": song_name,
|
113 |
+
"n_window": idx + 1,
|
114 |
+
**self._generate_stem_dict(S, song_name, end=S.shape[1]),
|
115 |
+
"mean": mean,
|
116 |
+
"std": std,
|
117 |
}
|
118 |
|
119 |
id_ += 1
|