Spaces:
Running
Running
Update
Browse files- app.py +1 -1
- style_gen.py +66 -0
app.py
CHANGED
|
@@ -469,7 +469,7 @@ if __name__ == "__main__":
|
|
| 469 |
style_weight = gr.Slider(
|
| 470 |
minimum=0,
|
| 471 |
maximum=50,
|
| 472 |
-
value=
|
| 473 |
step=0.1,
|
| 474 |
label="スタイルの強さ",
|
| 475 |
)
|
|
|
|
| 469 |
style_weight = gr.Slider(
|
| 470 |
minimum=0,
|
| 471 |
maximum=50,
|
| 472 |
+
value=5,
|
| 473 |
step=0.1,
|
| 474 |
label="スタイルの強さ",
|
| 475 |
)
|
style_gen.py
ADDED
|
@@ -0,0 +1,66 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
import argparse
|
| 2 |
+
import concurrent.futures
|
| 3 |
+
import sys
|
| 4 |
+
import warnings
|
| 5 |
+
|
| 6 |
+
import numpy as np
|
| 7 |
+
import torch
|
| 8 |
+
from tqdm import tqdm
|
| 9 |
+
|
| 10 |
+
import utils
|
| 11 |
+
from config import config
|
| 12 |
+
|
| 13 |
+
warnings.filterwarnings("ignore", category=UserWarning)
|
| 14 |
+
from pyannote.audio import Inference, Model
|
| 15 |
+
|
| 16 |
+
model = Model.from_pretrained("pyannote/wespeaker-voxceleb-resnet34-LM")
|
| 17 |
+
inference = Inference(model, window="whole")
|
| 18 |
+
device = torch.device(config.style_gen_config.device)
|
| 19 |
+
inference.to(device)
|
| 20 |
+
|
| 21 |
+
|
| 22 |
+
def extract_style_vector(wav_path):
|
| 23 |
+
return inference(wav_path)
|
| 24 |
+
|
| 25 |
+
|
| 26 |
+
def save_style_vector(wav_path):
|
| 27 |
+
style_vec = extract_style_vector(wav_path)
|
| 28 |
+
# `test.wav` -> `test.wav.npy`
|
| 29 |
+
np.save(f"{wav_path}.npy", style_vec)
|
| 30 |
+
|
| 31 |
+
|
| 32 |
+
if __name__ == "__main__":
|
| 33 |
+
parser = argparse.ArgumentParser()
|
| 34 |
+
parser.add_argument(
|
| 35 |
+
"-c", "--config", type=str, default=config.style_gen_config.config_path
|
| 36 |
+
)
|
| 37 |
+
parser.add_argument(
|
| 38 |
+
"--num_processes", type=int, default=config.style_gen_config.num_processes
|
| 39 |
+
)
|
| 40 |
+
args, _ = parser.parse_known_args()
|
| 41 |
+
config_path = args.config
|
| 42 |
+
num_processes = args.num_processes
|
| 43 |
+
|
| 44 |
+
hps = utils.get_hparams_from_file(config_path)
|
| 45 |
+
|
| 46 |
+
device = config.style_gen_config.device
|
| 47 |
+
|
| 48 |
+
lines = []
|
| 49 |
+
with open(hps.data.training_files, encoding="utf-8") as f:
|
| 50 |
+
lines.extend(f.readlines())
|
| 51 |
+
|
| 52 |
+
with open(hps.data.validation_files, encoding="utf-8") as f:
|
| 53 |
+
lines.extend(f.readlines())
|
| 54 |
+
|
| 55 |
+
wavnames = [line.split("|")[0] for line in lines]
|
| 56 |
+
|
| 57 |
+
with concurrent.futures.ThreadPoolExecutor(max_workers=num_processes) as executor:
|
| 58 |
+
list(
|
| 59 |
+
tqdm(
|
| 60 |
+
executor.map(save_style_vector, wavnames),
|
| 61 |
+
total=len(wavnames),
|
| 62 |
+
file=sys.stdout,
|
| 63 |
+
)
|
| 64 |
+
)
|
| 65 |
+
|
| 66 |
+
print(f"Finished generating style vectors! total: {len(wavnames)} npy files.")
|