OSError: [Errno 22] Invalid argumentの対処方法について

＜やりたいこと＞
OSError: [Errno 22] Invalid argumentの対処をしたいです。raw文字列リテラルを使う方法やファイルパスが間違っていないか確認しましたが解決できませんでした。対処法をご教授いただければと思います。

＜エラー内容＞
Traceback (most recent call last):
File "convertor.py", line 122, in <module>
cupy_mode(gpu)
File "convertor.py", line 73, in cupy_mode
ds = dataset.WaveDataset(wave_path, -1, True)
File "C:\Users\n\Desktop\Deep_VoiceChanger-master1\Deep_VoiceChanger-master\dataset.py", line 50, in init
self.wave = np.array(load(wave)[1], dtype=float)
File "C:\Users\n\Desktop\Deep_VoiceChanger-master1\Deep_VoiceChanger-master\dataset.py", line 22, in load
bps, data = wav.read(path)
File "C:\Users\n\AppData\Local\Programs\Python\Python37\lib\site-packages\scipy\io\wavfile.py", line 264, in read
fid = open(filename, 'rb')
OSError: [Errno 22] Invalid argument: '"C:\Users\n\Desktop\Deep_VoiceChanger-master1\Deep_VoiceChanger-master\results\test.wav"'

windowsコマンドプロンプトからファイルを指定して実行
C:\Users\n>cd C:\Users\n\Desktop\Deep_VoiceChanger-master1\Deep_VoiceChanger-master

C:\Users\n\Desktop\Deep_VoiceChanger-master1\Deep_VoiceChanger-master>python convertor.py
enter gpu number(if you have no gpu, enter -1)...0
enter netA path...\n\Desktop\Deep_VoiceChanger-master1\Deep_VoiceChanger-master\results\generator_ab.npz"
enter netB path..."C:\Users\n\Desktop\Deep_VoiceChanger-master1\Deep_VoiceChanger-master\results\generator_ba.npz"
enter wave path..."C:\Users\n\Desktop\Deep_VoiceChanger-master1\Deep_VoiceChanger-master\results\test.wav"
enter batch size...16

python3.7.0

python3
1コード
2
3import chainer
4import dataset
5import tqdm
6import queue
7import numpy as np
8from gla.gla_util import GLA
9from nets.models import Generator
10
11def numpy_mode():
12    with chainer.using_config('train', False):
13        with chainer.no_backprop_mode():
14            netA_path = input('enter netA path...')
15            netB_path = input('enter netB path...')
16            wave_path = input('enter wave path...')
17
18            ds = dataset.WaveDataset(wave_path, -1, True)
19
20            netA = Generator()
21            netB = Generator()
22            chainer.serializers.load_npz(netA_path, netA)
23            chainer.serializers.load_npz(netB_path, netB)
24
25            que_a = queue.deque()
26            que_ab = queue.deque()
27            que_aba = queue.deque()
28
29            gla = GLA()
30
31            print('converting...')
32            for i in tqdm.tqdm(range(ds.max//dataset.dif)):
33                x_a = ds.get_example(i)
34                x_a = chainer.dataset.convert.concat_examples([x_a], -1)
35                x_a = chainer.Variable(x_a)
36
37                x_ab = netA(x_a)
38                x_aba = netB(x_ab)
39
40                que_a  .append(x_a  .data[0])
41                que_ab .append(x_ab .data[0])
42                que_aba.append(x_aba.data[0])
43            print('done')
44
45            print('phase estimating...')
46            for i, que, name in zip(range(3), [que_a, que_ab, que_aba], ['a.wav', 'ab.wav', 'aba.wav']):
47                print()
48                print(i+1, '/ 3')
49                wave   = np.concatenate([gla.inverse(c_f) for i_f in tqdm.tqdm(que) for c_f in dataset.reverse(i_f)])
50                print('done...')
51                dataset.save(wave_path + name, 16000, wave)
52                print('wave-file saved at', wave_path + name)
53
54            print('all done')
55
56def cupy_mode(gpu):
57    from gla.gla_gpu import GLA_GPU
58    cp = chainer.cuda.cupy
59
60    with chainer.using_config('train', False):
61        with chainer.no_backprop_mode():
62            netA_path = input('enter netA path...')
63            netB_path = input('enter netB path...')
64            wave_path = input('enter wave path...')
65            batchsize = int(input('enter batch size...'))
66            chainer.cuda.get_device_from_id(gpu).use()
67
68            ds = dataset.WaveDataset(wave_path, -1, True)
69
70            netA = Generator()
71            netB = Generator()
72            chainer.serializers.load_npz(netA_path, netA)
73            chainer.serializers.load_npz(netB_path, netB)
74            netA.to_gpu()
75            netB.to_gpu()
76
77            que_a = queue.deque()
78            que_ab = queue.deque()
79            que_aba = queue.deque()
80
81            gla = GLA_GPU(batchsize*4)
82
83            print('converting...')
84            l = ds.max//dataset.dif
85            for i in tqdm.tqdm(range(0, l, batchsize)):
86                x_a = [ds.get_example(_i) for _i in range(i, min([i+batchsize, l]))]
87                x_a = chainer.dataset.convert.concat_examples(x_a, gpu)
88                x_a = chainer.Variable(x_a)
89
90                x_ab = netA(x_a)
91                x_aba = netB(x_ab)
92
93                que_a  .extend([dataset.reverse(_x) for _x in cp.asnumpy(x_a  .data)])
94                que_ab .extend([dataset.reverse(_x) for _x in cp.asnumpy(x_ab .data)])
95                que_aba.extend([dataset.reverse(_x) for _x in cp.asnumpy(x_aba.data)])
96            img_a   = np.concatenate(que_a,   axis=0)
97            img_ab  = np.concatenate(que_ab,  axis=0)
98            img_aba = np.concatenate(que_aba, axis=0)
99            print('done')
100
101            print('phase estimating...')
102            for i, img, name in zip(range(3), [img_a, img_ab, img_aba], ['a.wav', 'ab.wav', 'aba.wav']):
103                print()
104                print(i+1, '/ 3')
105                wave = gla.auto_inverse(img)
106                print('done...')
107                dataset.save(wave_path + name, 16000, wave)
108                print('wave-file saved at', wave_path + name)
109
110            print('all done')
111
112if __name__ == "__main__":
113    gpu = int(input('enter gpu number(if you have no gpu, enter -1)...'))
114    if gpu < 0:
115        numpy_mode()
116    else:
117        cupy_mode(gpu)
118