tuna2134/sbv2-v2
1
0
Fork 0
Code Issues Pull Requests 1 Actions Packages Projects Releases Wiki Activity

Fix: failed to start training

Browse Source
This commit is contained in:
tsukumi
2024-03-10 19:15:11 +00:00
parent be265d42ed
commit 859d940916
1 changed files with 16 additions and 15 deletions
Download Patch File Download Diff File Expand all files Collapse all files

31
style_bert_vits2/models/hyper_parameters.py
View File

@@ -39,8 +39,8 @@ class HyperParametersTrain(BaseModel):
class HyperParametersData(BaseModel): class HyperParametersData(BaseModel):
use_jp_extra: bool = True use_jp_extra: bool = True
training_files: str = "Data/dummy/train.list" training_files: str = "Data/Dummy/train.list"
validation_files: str = "Data/dummy/val.list" validation_files: str = "Data/Dummy/val.list"
max_wav_value: float = 32768.0 max_wav_value: float = 32768.0
sampling_rate: int = 44100 sampling_rate: int = 44100
filter_length: int = 2048 filter_length: int = 2048
@@ -53,7 +53,7 @@ class HyperParametersData(BaseModel):
n_speakers: int = 512 n_speakers: int = 512
cleaned_text: bool = True cleaned_text: bool = True
spk2id: dict[str, int] = { spk2id: dict[str, int] = {
"dummy": 0 "Dummy": 0,
} }
num_styles: int = 1 num_styles: int = 1
style2id: dict[str, int] = { style2id: dict[str, int] = {
@@ -61,6 +61,13 @@ class HyperParametersData(BaseModel):
} }
class HyperParametersModelSLM(BaseModel):
model: str = "./slm/wavlm-base-plus"
sr: int = 16000
hidden: int = 768
nlayers: int = 13
initial_channel: int = 64
class HyperParametersModel(BaseModel): class HyperParametersModel(BaseModel):
use_spk_conditioned_encoder: bool = True use_spk_conditioned_encoder: bool = True
use_noise_scaled_mas: bool = True use_noise_scaled_mas: bool = True
@@ -79,7 +86,7 @@ class HyperParametersModel(BaseModel):
resblock_dilation_sizes: list[list[int]] = [ resblock_dilation_sizes: list[list[int]] = [
[1, 3, 5], [1, 3, 5],
[1, 3, 5], [1, 3, 5],
[1, 3, 5] [1, 3, 5],
] ]
upsample_rates: list[int] = [8, 8, 2, 2, 2] upsample_rates: list[int] = [8, 8, 2, 2, 2]
upsample_initial_channel: int = 512 upsample_initial_channel: int = 512
@@ -87,21 +94,15 @@ class HyperParametersModel(BaseModel):
n_layers_q: int = 3 n_layers_q: int = 3
use_spectral_norm: bool = False use_spectral_norm: bool = False
gin_channels: int = 512 gin_channels: int = 512
slm: dict[str, Union[int, str]] = { slm: HyperParametersModelSLM = HyperParametersModelSLM()
"model": "./slm/wavlm-base-plus",
"sr": 16000,
"hidden": 768,
"nlayers": 13,
"initial_channel": 64
}
class HyperParameters(BaseModel): class HyperParameters(BaseModel):
model_name: str = 'dummy' model_name: str = 'Dummy'
version: str = "2.0-JP-Extra" version: str = "2.0-JP-Extra"
train: HyperParametersTrain train: HyperParametersTrain = HyperParametersTrain()
data: HyperParametersData data: HyperParametersData = HyperParametersData()
model: HyperParametersModel model: HyperParametersModel = HyperParametersModel()
# 以下は学習時にのみ動的に設定されるパラメータ (通常 config.json には存在しない) # 以下は学習時にのみ動的に設定されるパラメータ (通常 config.json には存在しない)
model_dir: Optional[str] = None model_dir: Optional[str] = None