-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathconfig_SPEECHCOMMANDS.yml
183 lines (171 loc) · 3.4 KB
/
config_SPEECHCOMMANDS.yml
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
# parameters configuration
mode: train
root: data/
predefined_dataset: SPEECHCOMMANDS
classes:
[
"backward",
"bed",
"bird",
"cat",
"dog",
"down",
"eight",
"five",
"follow",
"forward",
"four",
"go",
"happy",
"house",
"learn",
"left",
"marvin",
"nine",
"no",
"off",
"on",
"one",
"right",
"seven",
"sheila",
"six",
"stop",
"three",
"tree",
"two",
"up",
"visual",
"wow",
"yes",
"zero",
]
max_samples: null
batch_size: 32
num_workers: 0
device: cuda
sample_rate: 16000
lr: 1e-3
model_name: tf_mobilenetv3_small_minimal_100
in_chans: 1
loss_function_name: BCEWithLogitsLoss
data_balance: False
checkpoint_path: null
seed: 0
early_stopping: True
patience: 3
default_root_dir: save/
gpus: 1
precision: 32
max_epochs: 100
web_interface: True
examples:
[
"examples/SPEECHCOMMANDS/32ad5b65_nohash_2.wav",
"examples/SPEECHCOMMANDS/3a33d3a4_nohash_2.wav",
"examples/SPEECHCOMMANDS/3ca784ec_nohash_0.wav",
"examples/SPEECHCOMMANDS/ab5d7179_nohash_0.wav",
"examples/SPEECHCOMMANDS/cae62f38_nohash_0.wav",
]
tuning_test: False
cpu_resources_per_trial: 1
gpu_resources_per_trial: 1
num_samples: 100
# transforms configuration
transforms_config:
train:
selfdefined.DigitalFilter:
filter_type: bandpass
sample_rate: 16000
cutoff_freq: [300, 3000]
selfdefined.PadWaveform:
max_waveform_length: 16000
torchaudio.MelSpectrogram:
sample_rate: 16000
n_mels: 64
torchaudio.AmplitudeToDB: null
torchvision.Resize:
- 64
- 81
torchvision.RandomErasing:
value: -80
val:
selfdefined.DigitalFilter:
filter_type: bandpass
sample_rate: 16000
cutoff_freq: [300, 3000]
selfdefined.PadWaveform:
max_waveform_length: 16000
torchaudio.MelSpectrogram:
sample_rate: 16000
n_mels: 64
torchaudio.AmplitudeToDB: null
torchvision.Resize:
- 64
- 81
test:
selfdefined.DigitalFilter:
filter_type: bandpass
sample_rate: 16000
cutoff_freq: [300, 3000]
selfdefined.PadWaveform:
max_waveform_length: 16000
torchaudio.MelSpectrogram:
sample_rate: 16000
n_mels: 64
torchaudio.AmplitudeToDB: null
torchvision.Resize:
- 64
- 81
predict:
selfdefined.DigitalFilter:
filter_type: bandpass
sample_rate: 16000
cutoff_freq: [300, 3000]
selfdefined.PadWaveform:
max_waveform_length: 16000
torchaudio.MelSpectrogram:
sample_rate: 16000
n_mels: 64
torchaudio.AmplitudeToDB: null
torchvision.Resize:
- 64
- 81
# target transforms configuration
target_transforms_config:
train:
selfdefined.LabelSmoothing:
alpha: 0.2
num_classes: null
val:
selfdefined.OneHotEncoder:
num_classes: null
test:
selfdefined.OneHotEncoder:
num_classes: null
predict:
selfdefined.OneHotEncoder:
num_classes: null
# optimizers configuration
optimizers_config:
Adam:
betas:
- 0.9
- 0.999
eps: 1e-08
weight_decay: 0
amsgrad: False
# learning rate schedulers configuration
lr_schedulers_config:
CosineAnnealingLR:
T_max: 10
# hyperparameter space configuration
hyperparameter_space_config:
lr:
uniform:
lower: 1e-4
upper: 1e-1
max_epochs:
randint:
lower: 10
upper: 200