import from zenodo
Browse files
exp/tts_stats_raw_phn_jaconv_pyopenjtalk/train/feats_stats.npz
ADDED
Binary file (1.4 kB). View file
|
|
exp/tts_train_tacotron2_raw_phn_jaconv_pyopenjtalk/199epoch.pth
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:f0340ab7d52b1898cacfb8815134a83ddc2318b8f26fe6eed78b106c48673c67
|
3 |
+
size 106939103
|
exp/tts_train_tacotron2_raw_phn_jaconv_pyopenjtalk/config.yaml
ADDED
@@ -0,0 +1,197 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
config: conf/tuning/train_tacotron2.yaml
|
2 |
+
print_config: false
|
3 |
+
log_level: INFO
|
4 |
+
dry_run: false
|
5 |
+
iterator_type: sequence
|
6 |
+
output_dir: exp/tts_train_tacotron2_raw_batch_bins3750000
|
7 |
+
ngpu: 1
|
8 |
+
seed: 0
|
9 |
+
num_workers: 1
|
10 |
+
num_att_plot: 3
|
11 |
+
dist_backend: nccl
|
12 |
+
dist_init_method: env://
|
13 |
+
dist_world_size: null
|
14 |
+
dist_rank: null
|
15 |
+
local_rank: 0
|
16 |
+
dist_master_addr: null
|
17 |
+
dist_master_port: null
|
18 |
+
dist_launcher: null
|
19 |
+
multiprocessing_distributed: false
|
20 |
+
cudnn_enabled: true
|
21 |
+
cudnn_benchmark: false
|
22 |
+
cudnn_deterministic: true
|
23 |
+
collect_stats: false
|
24 |
+
write_collected_feats: false
|
25 |
+
max_epoch: 200
|
26 |
+
patience: null
|
27 |
+
val_scheduler_criterion:
|
28 |
+
- valid
|
29 |
+
- loss
|
30 |
+
early_stopping_criterion:
|
31 |
+
- valid
|
32 |
+
- loss
|
33 |
+
- min
|
34 |
+
best_model_criterion:
|
35 |
+
- - valid
|
36 |
+
- loss
|
37 |
+
- min
|
38 |
+
- - train
|
39 |
+
- loss
|
40 |
+
- min
|
41 |
+
keep_nbest_models: 5
|
42 |
+
grad_clip: 1.0
|
43 |
+
grad_noise: false
|
44 |
+
accum_grad: 1
|
45 |
+
no_forward_run: false
|
46 |
+
resume: true
|
47 |
+
train_dtype: float32
|
48 |
+
log_interval: null
|
49 |
+
pretrain_path: []
|
50 |
+
pretrain_key: []
|
51 |
+
num_iters_per_epoch: null
|
52 |
+
batch_size: 20
|
53 |
+
valid_batch_size: null
|
54 |
+
batch_bins: 3750000
|
55 |
+
valid_batch_bins: null
|
56 |
+
train_shape_file:
|
57 |
+
- exp/tts_stats_raw_phn_jaconv_pyopenjtalk/train/text_shape.phn
|
58 |
+
- exp/tts_stats_raw_phn_jaconv_pyopenjtalk/train/speech_shape
|
59 |
+
valid_shape_file:
|
60 |
+
- exp/tts_stats_raw_phn_jaconv_pyopenjtalk/valid/text_shape.phn
|
61 |
+
- exp/tts_stats_raw_phn_jaconv_pyopenjtalk/valid/speech_shape
|
62 |
+
batch_type: numel
|
63 |
+
valid_batch_type: null
|
64 |
+
fold_length:
|
65 |
+
- 150
|
66 |
+
- 240000
|
67 |
+
sort_in_batch: descending
|
68 |
+
sort_batch: descending
|
69 |
+
multiple_iterator: false
|
70 |
+
chunk_length: 500
|
71 |
+
chunk_shift_ratio: 0.5
|
72 |
+
num_cache_chunks: 1024
|
73 |
+
train_data_path_and_name_and_type:
|
74 |
+
- - dump/raw/tr_no_dev/text
|
75 |
+
- text
|
76 |
+
- text
|
77 |
+
- - dump/raw/tr_no_dev/wav.scp
|
78 |
+
- speech
|
79 |
+
- sound
|
80 |
+
valid_data_path_and_name_and_type:
|
81 |
+
- - dump/raw/dev/text
|
82 |
+
- text
|
83 |
+
- text
|
84 |
+
- - dump/raw/dev/wav.scp
|
85 |
+
- speech
|
86 |
+
- sound
|
87 |
+
allow_variable_data_keys: false
|
88 |
+
max_cache_size: 0.0
|
89 |
+
valid_max_cache_size: null
|
90 |
+
optim: adam
|
91 |
+
optim_conf:
|
92 |
+
lr: 0.001
|
93 |
+
eps: 1.0e-06
|
94 |
+
weight_decay: 0.0
|
95 |
+
scheduler: null
|
96 |
+
scheduler_conf: {}
|
97 |
+
token_list:
|
98 |
+
- <blank>
|
99 |
+
- <unk>
|
100 |
+
- ty
|
101 |
+
- dy
|
102 |
+
- v
|
103 |
+
- py
|
104 |
+
- my
|
105 |
+
- by
|
106 |
+
- ny
|
107 |
+
- hy
|
108 |
+
- gy
|
109 |
+
- ry
|
110 |
+
- ky
|
111 |
+
- f
|
112 |
+
- p
|
113 |
+
- z
|
114 |
+
- ch
|
115 |
+
- ts
|
116 |
+
- j
|
117 |
+
- b
|
118 |
+
- y
|
119 |
+
- h
|
120 |
+
- cl
|
121 |
+
- I
|
122 |
+
- U
|
123 |
+
- w
|
124 |
+
- g
|
125 |
+
- d
|
126 |
+
- sh
|
127 |
+
- pau
|
128 |
+
- m
|
129 |
+
- N
|
130 |
+
- s
|
131 |
+
- r
|
132 |
+
- t
|
133 |
+
- n
|
134 |
+
- k
|
135 |
+
- e
|
136 |
+
- u
|
137 |
+
- i
|
138 |
+
- o
|
139 |
+
- a
|
140 |
+
- <sos/eos>
|
141 |
+
odim: null
|
142 |
+
model_conf: {}
|
143 |
+
use_preprocessor: true
|
144 |
+
token_type: phn
|
145 |
+
bpemodel: null
|
146 |
+
non_linguistic_symbols: null
|
147 |
+
cleaner: jaconv
|
148 |
+
g2p: pyopenjtalk
|
149 |
+
feats_extract: fbank
|
150 |
+
feats_extract_conf:
|
151 |
+
fs: 24000
|
152 |
+
fmin: 80
|
153 |
+
fmax: 7600
|
154 |
+
n_mels: 80
|
155 |
+
hop_length: 300
|
156 |
+
n_fft: 2048
|
157 |
+
win_length: 1200
|
158 |
+
normalize: global_mvn
|
159 |
+
normalize_conf:
|
160 |
+
stats_file: exp/tts_stats_raw_phn_jaconv_pyopenjtalk/train/feats_stats.npz
|
161 |
+
tts: tacotron2
|
162 |
+
tts_conf:
|
163 |
+
embed_dim: 512
|
164 |
+
elayers: 1
|
165 |
+
eunits: 512
|
166 |
+
econv_layers: 3
|
167 |
+
econv_chans: 512
|
168 |
+
econv_filts: 5
|
169 |
+
atype: location
|
170 |
+
adim: 512
|
171 |
+
aconv_chans: 32
|
172 |
+
aconv_filts: 15
|
173 |
+
cumulate_att_w: true
|
174 |
+
dlayers: 2
|
175 |
+
dunits: 1024
|
176 |
+
prenet_layers: 2
|
177 |
+
prenet_units: 256
|
178 |
+
postnet_layers: 5
|
179 |
+
postnet_chans: 512
|
180 |
+
postnet_filts: 5
|
181 |
+
output_activation: null
|
182 |
+
use_batch_norm: true
|
183 |
+
use_concate: true
|
184 |
+
use_residual: false
|
185 |
+
dropout_rate: 0.5
|
186 |
+
zoneout_rate: 0.1
|
187 |
+
reduction_factor: 1
|
188 |
+
spk_embed_dim: null
|
189 |
+
use_masking: true
|
190 |
+
bce_pos_weight: 5.0
|
191 |
+
use_guided_attn_loss: true
|
192 |
+
guided_attn_loss_sigma: 0.4
|
193 |
+
guided_attn_loss_lambda: 1.0
|
194 |
+
required:
|
195 |
+
- output_dir
|
196 |
+
- token_list
|
197 |
+
distributed: false
|
meta.yaml
ADDED
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
espnet: 0.8.0
|
2 |
+
files:
|
3 |
+
model_file: exp/tts_train_tacotron2_raw_phn_jaconv_pyopenjtalk/199epoch.pth
|
4 |
+
python: "3.7.3 (default, Mar 27 2019, 22:11:17) \n[GCC 7.3.0]"
|
5 |
+
timestamp: 1595923764.402962
|
6 |
+
torch: 1.5.1
|
7 |
+
yaml_files:
|
8 |
+
train_config: exp/tts_train_tacotron2_raw_phn_jaconv_pyopenjtalk/config.yaml
|