-
Notifications
You must be signed in to change notification settings - Fork 0
/
hyperparams.py
53 lines (49 loc) · 2.63 KB
/
hyperparams.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
class hyperparams:
def __init__(self):
#################################################################################
# #
# Preprocess Hyperparams #
# #
#################################################################################
# ------------------------ Paths And Directory ------------------------ #
self.CSV_PATH = './data/dataset.csv'
self.SOURCE_CSV_PATH = './paralleldata/source.csv' # 平行中的发起者:source
self.TOTAL_CSV_PATH = './paralleldata/total.csv' # 保存着所有的训练数据(source 和 target)
# Format: spk|fpath
# Example: xyb|0001.wav
self.TRAIN_DATASET_PATH = './paralleldata/train_data'
self.PARALLEL_TRAIN_DATASET_PATH = './paralleldata/train_data'
self.TEST_DATASET_PATH = './paralleldata/test_data'
self.WAVS_PATH = './data/wavs'
self.TOTAL_WAVS_PATH = './paralleldata/TOTAL'
self.SOURCE_WAVS_PATH = './paralleldata/VCC2SF1' # 新增加的:source语音路径
self.TARGET_WAVS_PATH = './paralleldata/VCC2TM1' # 新增加的:target语音路径
self.TRAIN_RATE = 1
# ------------------------ Setting And Hyperparams -------------------- #
self.MULTI_PROCESS = True
self.CPU_RATE = 1
self.SR = 16000
self.N_FFT = 1024
self.CODED_DIM = 80 # 压缩成80维 mel
self.SPK_NUM = 14
self.PARALLEL_SPK_NUM = 2
#################################################################################
# #
# Train Hyperparams #
# #
#################################################################################
# ------------------------ Paths And Directory ------------------------ #
self.LOG_DIR = './logs'
self.MODEL_DIR = './models'
# ------------------------ Setting And Hyperparams -------------------- #
self.NUM_EPOCHS = 150
self.BATCH_SIZE = 16
self.EMBED_SIZE = 256
self.G_LR = 0.001
self.D_LR = 0.001
# VAE LOSS WEIGHT
self.X0 = 2500
self.K = 0.0025
self.DROPOUT_RATE = 0.5
self.VAE_GAUSSION_UNITS = 16
self.PER_STEPS = 100