mirror of
https://github.com/Anjok07/ultimatevocalremovergui.git
synced 2024-11-24 07:20:10 +01:00
Delete lib_v5 directory
This commit is contained in:
Anjok07
GitHub
parent
e9c963d05d
commit
8ee30276b2
@ -1,170 +0,0 @@
|
||||
import os
|
||||
import random
|
||||
|
||||
import numpy as np
|
||||
import torch
|
||||
import torch.utils.data
|
||||
from tqdm import tqdm
|
||||
|
||||
from lib_v5 import spec_utils
|
||||
|
||||
|
||||
class VocalRemoverValidationSet(torch.utils.data.Dataset):
|
||||
|
||||
def __init__(self, patch_list):
|
||||
self.patch_list = patch_list
|
||||
|
||||
def __len__(self):
|
||||
return len(self.patch_list)
|
||||
|
||||
def __getitem__(self, idx):
|
||||
path = self.patch_list[idx]
|
||||
data = np.load(path)
|
||||
|
||||
X, y = data['X'], data['y']
|
||||
|
||||
X_mag = np.abs(X)
|
||||
y_mag = np.abs(y)
|
||||
|
||||
return X_mag, y_mag
|
||||
|
||||
|
||||
def make_pair(mix_dir, inst_dir):
|
||||
input_exts = ['.wav', '.m4a', '.mp3', '.mp4', '.flac']
|
||||
|
||||
X_list = sorted([
|
||||
os.path.join(mix_dir, fname)
|
||||
for fname in os.listdir(mix_dir)
|
||||
if os.path.splitext(fname)[1] in input_exts])
|
||||
y_list = sorted([
|
||||
os.path.join(inst_dir, fname)
|
||||
for fname in os.listdir(inst_dir)
|
||||
if os.path.splitext(fname)[1] in input_exts])
|
||||
|
||||
filelist = list(zip(X_list, y_list))
|
||||
|
||||
return filelist
|
||||
|
||||
|
||||
def train_val_split(dataset_dir, split_mode, val_rate, val_filelist):
|
||||
if split_mode == 'random':
|
||||
filelist = make_pair(
|
||||
os.path.join(dataset_dir, 'mixtures'),
|
||||
os.path.join(dataset_dir, 'instruments'))
|
||||
|
||||
random.shuffle(filelist)
|
||||
|
||||
if len(val_filelist) == 0:
|
||||
val_size = int(len(filelist) * val_rate)
|
||||
train_filelist = filelist[:-val_size]
|
||||
val_filelist = filelist[-val_size:]
|
||||
else:
|
||||
train_filelist = [
|
||||
pair for pair in filelist
|
||||
if list(pair) not in val_filelist]
|
||||
elif split_mode == 'subdirs':
|
||||
if len(val_filelist) != 0:
|
||||
raise ValueError('The `val_filelist` option is not available in `subdirs` mode')
|
||||
|
||||
train_filelist = make_pair(
|
||||
os.path.join(dataset_dir, 'training/mixtures'),
|
||||
os.path.join(dataset_dir, 'training/instruments'))
|
||||
|
||||
val_filelist = make_pair(
|
||||
os.path.join(dataset_dir, 'validation/mixtures'),
|
||||
os.path.join(dataset_dir, 'validation/instruments'))
|
||||
|
||||
return train_filelist, val_filelist
|
||||
|
||||
|
||||
def augment(X, y, reduction_rate, reduction_mask, mixup_rate, mixup_alpha):
|
||||
perm = np.random.permutation(len(X))
|
||||
for i, idx in enumerate(tqdm(perm)):
|
||||
if np.random.uniform() < reduction_rate:
|
||||
y[idx] = spec_utils.reduce_vocal_aggressively(X[idx], y[idx], reduction_mask)
|
||||
|
||||
if np.random.uniform() < 0.5:
|
||||
# swap channel
|
||||
X[idx] = X[idx, ::-1]
|
||||
y[idx] = y[idx, ::-1]
|
||||
if np.random.uniform() < 0.02:
|
||||
# mono
|
||||
X[idx] = X[idx].mean(axis=0, keepdims=True)
|
||||
y[idx] = y[idx].mean(axis=0, keepdims=True)
|
||||
if np.random.uniform() < 0.02:
|
||||
# inst
|
||||
X[idx] = y[idx]
|
||||
|
||||
if np.random.uniform() < mixup_rate and i < len(perm) - 1:
|
||||
lam = np.random.beta(mixup_alpha, mixup_alpha)
|
||||
X[idx] = lam * X[idx] + (1 - lam) * X[perm[i + 1]]
|
||||
y[idx] = lam * y[idx] + (1 - lam) * y[perm[i + 1]]
|
||||
|
||||
return X, y
|
||||
|
||||
|
||||
def make_padding(width, cropsize, offset):
|
||||
left = offset
|
||||
roi_size = cropsize - left * 2
|
||||
if roi_size == 0:
|
||||
roi_size = cropsize
|
||||
right = roi_size - (width % roi_size) + left
|
||||
|
||||
return left, right, roi_size
|
||||
|
||||
|
||||
def make_training_set(filelist, cropsize, patches, sr, hop_length, n_fft, offset):
|
||||
len_dataset = patches * len(filelist)
|
||||
|
||||
X_dataset = np.zeros(
|
||||
(len_dataset, 2, n_fft // 2 + 1, cropsize), dtype=np.complex64)
|
||||
y_dataset = np.zeros(
|
||||
(len_dataset, 2, n_fft // 2 + 1, cropsize), dtype=np.complex64)
|
||||
|
||||
for i, (X_path, y_path) in enumerate(tqdm(filelist)):
|
||||
X, y = spec_utils.cache_or_load(X_path, y_path, sr, hop_length, n_fft)
|
||||
coef = np.max([np.abs(X).max(), np.abs(y).max()])
|
||||
X, y = X / coef, y / coef
|
||||
|
||||
l, r, roi_size = make_padding(X.shape[2], cropsize, offset)
|
||||
X_pad = np.pad(X, ((0, 0), (0, 0), (l, r)), mode='constant')
|
||||
y_pad = np.pad(y, ((0, 0), (0, 0), (l, r)), mode='constant')
|
||||
|
||||
starts = np.random.randint(0, X_pad.shape[2] - cropsize, patches)
|
||||
ends = starts + cropsize
|
||||
for j in range(patches):
|
||||
idx = i * patches + j
|
||||
X_dataset[idx] = X_pad[:, :, starts[j]:ends[j]]
|
||||
y_dataset[idx] = y_pad[:, :, starts[j]:ends[j]]
|
||||
|
||||
return X_dataset, y_dataset
|
||||
|
||||
|
||||
def make_validation_set(filelist, cropsize, sr, hop_length, n_fft, offset):
|
||||
patch_list = []
|
||||
patch_dir = 'cs{}_sr{}_hl{}_nf{}_of{}'.format(cropsize, sr, hop_length, n_fft, offset)
|
||||
os.makedirs(patch_dir, exist_ok=True)
|
||||
|
||||
for i, (X_path, y_path) in enumerate(tqdm(filelist)):
|
||||
basename = os.path.splitext(os.path.basename(X_path))[0]
|
||||
|
||||
X, y = spec_utils.cache_or_load(X_path, y_path, sr, hop_length, n_fft)
|
||||
coef = np.max([np.abs(X).max(), np.abs(y).max()])
|
||||
X, y = X / coef, y / coef
|
||||
|
||||
l, r, roi_size = make_padding(X.shape[2], cropsize, offset)
|
||||
X_pad = np.pad(X, ((0, 0), (0, 0), (l, r)), mode='constant')
|
||||
y_pad = np.pad(y, ((0, 0), (0, 0), (l, r)), mode='constant')
|
||||
|
||||
len_dataset = int(np.ceil(X.shape[2] / roi_size))
|
||||
for j in range(len_dataset):
|
||||
outpath = os.path.join(patch_dir, '{}_p{}.npz'.format(basename, j))
|
||||
start = j * roi_size
|
||||
if not os.path.exists(outpath):
|
||||
np.savez(
|
||||
outpath,
|
||||
X=X_pad[:, :, start:start + cropsize],
|
||||
y=y_pad[:, :, start:start + cropsize])
|
||||
patch_list.append(outpath)
|
||||
|
||||
return VocalRemoverValidationSet(patch_list)
|
||||
@ -1,423 +0,0 @@
|
||||
import json
|
||||
|
||||
def get_vr_download_list(list):
|
||||
with open("lib_v5/filelists/download_lists/vr_download_list.txt", "r") as f:
|
||||
text=f.read().splitlines()
|
||||
|
||||
list = text
|
||||
|
||||
return list
|
||||
|
||||
def get_mdx_download_list(list):
|
||||
with open("lib_v5/filelists/download_lists/mdx_download_list.txt", "r") as f:
|
||||
text=f.read().splitlines()
|
||||
|
||||
list = text
|
||||
|
||||
return list
|
||||
|
||||
def get_demucs_download_list(list):
|
||||
with open("lib_v5/filelists/download_lists/demucs_download_list.txt", "r") as f:
|
||||
text=f.read().splitlines()
|
||||
|
||||
list = text
|
||||
|
||||
return list
|
||||
|
||||
def get_mdx_demucs_en_list(list):
|
||||
with open("lib_v5/filelists/ensemble_list/mdx_demuc_en_list.txt", "r") as f:
|
||||
text=f.read().splitlines()
|
||||
|
||||
list = text
|
||||
|
||||
return list
|
||||
|
||||
def get_vr_en_list(list):
|
||||
with open("lib_v5/filelists/ensemble_list/vr_en_list.txt", "r") as f:
|
||||
text=f.read().splitlines()
|
||||
|
||||
list = text
|
||||
|
||||
return list
|
||||
|
||||
def get_download_links(links, downloads=''):
|
||||
|
||||
f = open(f"lib_v5/filelists/download_lists/download_links.json")
|
||||
download_links = json.load(f)
|
||||
|
||||
if downloads == 'Demucs v3: mdx':
|
||||
url_1 = download_links['Demucs_v3_mdx_url_1']
|
||||
url_2 = download_links['Demucs_v3_mdx_url_2']
|
||||
url_3 = download_links['Demucs_v3_mdx_url_3']
|
||||
url_4 = download_links['Demucs_v3_mdx_url_4']
|
||||
url_5 = download_links['Demucs_v3_mdx_url_5']
|
||||
|
||||
links = url_1, url_2, url_3, url_4, url_5
|
||||
|
||||
|
||||
if downloads == 'Demucs v3: mdx_q':
|
||||
url_1 = download_links['Demucs_v3_mdx_q_url_1']
|
||||
url_2 = download_links['Demucs_v3_mdx_q_url_2']
|
||||
url_3 = download_links['Demucs_v3_mdx_q_url_3']
|
||||
url_4 = download_links['Demucs_v3_mdx_q_url_4']
|
||||
url_5 = download_links['Demucs_v3_mdx_q_url_5']
|
||||
|
||||
links = url_1, url_2, url_3, url_4, url_5
|
||||
|
||||
if downloads == 'Demucs v3: mdx_extra':
|
||||
url_1 = download_links['Demucs_v3_mdx_extra_url_1']
|
||||
url_2 = download_links['Demucs_v3_mdx_extra_url_2']
|
||||
url_3 = download_links['Demucs_v3_mdx_extra_url_3']
|
||||
url_4 = download_links['Demucs_v3_mdx_extra_url_4']
|
||||
url_5 = download_links['Demucs_v3_mdx_extra_url_5']
|
||||
|
||||
links = url_1, url_2, url_3, url_4, url_5
|
||||
|
||||
if downloads == 'Demucs v3: mdx_extra_q':
|
||||
url_1 = download_links['Demucs_v3_mdx_extra_q_url_1']
|
||||
url_2 = download_links['Demucs_v3_mdx_extra_q_url_2']
|
||||
url_3 = download_links['Demucs_v3_mdx_extra_q_url_3']
|
||||
url_4 = download_links['Demucs_v3_mdx_extra_q_url_4']
|
||||
url_5 = download_links['Demucs_v3_mdx_extra_q_url_5']
|
||||
|
||||
links = url_1, url_2, url_3, url_4, url_5
|
||||
|
||||
if downloads == 'Demucs v3: UVR Models':
|
||||
url_1 = download_links['Demucs_v3_UVR_url_1']
|
||||
url_2 = download_links['Demucs_v3_UVR_url_2']
|
||||
url_3 = download_links['Demucs_v3_UVR_url_3']
|
||||
url_4 = download_links['Demucs_v3_UVR_url_4']
|
||||
url_5 = download_links['Demucs_v3_UVR_url_5']
|
||||
|
||||
links = url_1, url_2, url_3, url_4, url_5
|
||||
|
||||
if downloads == 'Demucs v2: demucs':
|
||||
url_1 = download_links['Demucs_v2_demucs_url_1']
|
||||
links = url_1
|
||||
|
||||
if downloads == 'Demucs v2: demucs_extra':
|
||||
url_1 = download_links['Demucs_v2_demucs_extra_url_1']
|
||||
|
||||
links = url_1
|
||||
|
||||
if downloads == 'Demucs v2: demucs48_hq':
|
||||
url_1 = download_links['Demucs_v2_demucs48_hq_url_1']
|
||||
|
||||
links = url_1
|
||||
|
||||
if downloads == 'Demucs v2: tasnet':
|
||||
url_1 = download_links['Demucs_v2_tasnet_url_1']
|
||||
|
||||
links = url_1
|
||||
|
||||
if downloads == 'Demucs v2: tasnet_extra':
|
||||
url_1 = download_links['Demucs_v2_tasnet_extra_url_1']
|
||||
|
||||
links = url_1
|
||||
|
||||
if downloads == 'Demucs v2: demucs_unittest':
|
||||
url_1 = download_links['Demucs_v2_demucs_unittest_url_1']
|
||||
|
||||
links = url_1
|
||||
|
||||
if downloads == 'Demucs v1: demucs':
|
||||
url_1 = download_links['Demucs_v1_demucs_url_1']
|
||||
|
||||
links = url_1
|
||||
|
||||
if downloads == 'Demucs v1: demucs_extra':
|
||||
url_1 = download_links['Demucs_v1_demucs_extra_url_1']
|
||||
|
||||
links = url_1
|
||||
|
||||
if downloads == 'Demucs v1: light':
|
||||
url_1 = download_links['Demucs_v1_light_url_1']
|
||||
|
||||
links = url_1
|
||||
|
||||
if downloads == 'Demucs v1: light_extra':
|
||||
url_1 = download_links['Demucs_v1_light_extra_url_1']
|
||||
|
||||
links = url_1
|
||||
|
||||
if downloads == 'Demucs v1: tasnet':
|
||||
url_1 = download_links['Demucs_v1_tasnet_url_1']
|
||||
|
||||
links = url_1
|
||||
|
||||
if downloads == 'Demucs v1: tasnet_extra':
|
||||
url_1 = download_links['Demucs_v1_tasnet_extra_url_1']
|
||||
|
||||
links = url_1
|
||||
|
||||
if downloads == 'model_repo':
|
||||
url_1 = download_links['model_repo_url_1']
|
||||
|
||||
links = url_1
|
||||
|
||||
if downloads == 'single_model_repo':
|
||||
url_1 = download_links['single_model_repo_url_1']
|
||||
|
||||
links = url_1
|
||||
|
||||
if downloads == 'exclusive':
|
||||
url_1 = download_links['exclusive_url_1']
|
||||
url_2 = download_links['exclusive_url_2']
|
||||
|
||||
links = url_1, url_2, url_3
|
||||
|
||||
if downloads == 'refresh':
|
||||
url_1 = download_links['refresh_url_1']
|
||||
url_2 = download_links['refresh_url_2']
|
||||
url_3 = download_links['refresh_url_3']
|
||||
|
||||
links = url_1, url_2, url_3
|
||||
|
||||
if downloads == 'app_patch':
|
||||
url_1 = download_links['app_patch']
|
||||
|
||||
links = url_1
|
||||
|
||||
return links
|
||||
|
||||
def provide_model_param_hash(model_hash):
|
||||
#v5 Models
|
||||
if model_hash == '47939caf0cfe52a0e81442b85b971dfd':
|
||||
model_params_set=str('lib_v5/modelparams/4band_44100.json')
|
||||
param_name=str('4band_44100')
|
||||
elif model_hash == '4e4ecb9764c50a8c414fee6e10395bbe':
|
||||
model_params_set=str('lib_v5/modelparams/4band_v2.json')
|
||||
param_name=str('4band_v2')
|
||||
elif model_hash == 'e60a1e84803ce4efc0a6551206cc4b71':
|
||||
model_params_set=str('lib_v5/modelparams/4band_44100.json')
|
||||
param_name=str('4band_44100')
|
||||
elif model_hash == 'a82f14e75892e55e994376edbf0c8435':
|
||||
model_params_set=str('lib_v5/modelparams/4band_44100.json')
|
||||
param_name=str('4band_44100')
|
||||
elif model_hash == '6dd9eaa6f0420af9f1d403aaafa4cc06':
|
||||
model_params_set=str('lib_v5/modelparams/4band_v2_sn.json')
|
||||
param_name=str('4band_v2_sn')
|
||||
elif model_hash == '5c7bbca45a187e81abbbd351606164e5':
|
||||
model_params_set=str('lib_v5/modelparams/3band_44100_msb2.json')
|
||||
param_name=str('3band_44100_msb2')
|
||||
elif model_hash == 'd6b2cb685a058a091e5e7098192d3233':
|
||||
model_params_set=str('lib_v5/modelparams/3band_44100_msb2.json')
|
||||
param_name=str('3band_44100_msb2')
|
||||
elif model_hash == 'c1b9f38170a7c90e96f027992eb7c62b':
|
||||
model_params_set=str('lib_v5/modelparams/4band_44100.json')
|
||||
param_name=str('4band_44100')
|
||||
elif model_hash == 'c3448ec923fa0edf3d03a19e633faa53':
|
||||
model_params_set=str('lib_v5/modelparams/4band_44100.json')
|
||||
param_name=str('4band_44100')
|
||||
elif model_hash == '68aa2c8093d0080704b200d140f59e54':
|
||||
model_params_set=str('lib_v5/modelparams/3band_44100.json')
|
||||
param_name=str('3band_44100.json')
|
||||
elif model_hash == 'fdc83be5b798e4bd29fe00fe6600e147':
|
||||
model_params_set=str('lib_v5/modelparams/3band_44100_mid.json')
|
||||
param_name=str('3band_44100_mid.json')
|
||||
elif model_hash == '2ce34bc92fd57f55db16b7a4def3d745':
|
||||
model_params_set=str('lib_v5/modelparams/3band_44100_mid.json')
|
||||
param_name=str('3band_44100_mid.json')
|
||||
elif model_hash == '52fdca89576f06cf4340b74a4730ee5f':
|
||||
model_params_set=str('lib_v5/modelparams/4band_44100.json')
|
||||
param_name=str('4band_44100.json')
|
||||
elif model_hash == '41191165b05d38fc77f072fa9e8e8a30':
|
||||
model_params_set=str('lib_v5/modelparams/4band_44100.json')
|
||||
param_name=str('4band_44100.json')
|
||||
elif model_hash == '89e83b511ad474592689e562d5b1f80e':
|
||||
model_params_set=str('lib_v5/modelparams/2band_32000.json')
|
||||
param_name=str('2band_32000.json')
|
||||
elif model_hash == '0b954da81d453b716b114d6d7c95177f':
|
||||
model_params_set=str('lib_v5/modelparams/2band_32000.json')
|
||||
param_name=str('2band_32000.json')
|
||||
|
||||
#v4 Models
|
||||
|
||||
elif model_hash == '6a00461c51c2920fd68937d4609ed6c8':
|
||||
model_params_set=str('lib_v5/modelparams/1band_sr16000_hl512.json')
|
||||
param_name=str('1band_sr16000_hl512')
|
||||
elif model_hash == '0ab504864d20f1bd378fe9c81ef37140':
|
||||
model_params_set=str('lib_v5/modelparams/1band_sr32000_hl512.json')
|
||||
param_name=str('1band_sr32000_hl512')
|
||||
elif model_hash == '7dd21065bf91c10f7fccb57d7d83b07f':
|
||||
model_params_set=str('lib_v5/modelparams/1band_sr32000_hl512.json')
|
||||
param_name=str('1band_sr32000_hl512')
|
||||
elif model_hash == '80ab74d65e515caa3622728d2de07d23':
|
||||
model_params_set=str('lib_v5/modelparams/1band_sr32000_hl512.json')
|
||||
param_name=str('1band_sr32000_hl512')
|
||||
elif model_hash == 'edc115e7fc523245062200c00caa847f':
|
||||
model_params_set=str('lib_v5/modelparams/1band_sr33075_hl384.json')
|
||||
param_name=str('1band_sr33075_hl384')
|
||||
elif model_hash == '28063e9f6ab5b341c5f6d3c67f2045b7':
|
||||
model_params_set=str('lib_v5/modelparams/1band_sr33075_hl384.json')
|
||||
param_name=str('1band_sr33075_hl384')
|
||||
elif model_hash == 'b58090534c52cbc3e9b5104bad666ef2':
|
||||
model_params_set=str('lib_v5/modelparams/1band_sr44100_hl512.json')
|
||||
param_name=str('1band_sr44100_hl512')
|
||||
elif model_hash == '0cdab9947f1b0928705f518f3c78ea8f':
|
||||
model_params_set=str('lib_v5/modelparams/1band_sr44100_hl512.json')
|
||||
param_name=str('1band_sr44100_hl512')
|
||||
elif model_hash == 'ae702fed0238afb5346db8356fe25f13':
|
||||
model_params_set=str('lib_v5/modelparams/1band_sr44100_hl1024.json')
|
||||
param_name=str('1band_sr44100_hl1024')
|
||||
else:
|
||||
try:
|
||||
with open(f"lib_v5/filelists/model_cache/vr_param_cache/{model_hash}.txt", "r") as f:
|
||||
name = f.read()
|
||||
model_params_set=str(f'lib_v5/modelparams/{name}')
|
||||
param_name=str(name)
|
||||
('using text of hash worked')
|
||||
except:
|
||||
model_params_set=str('Not Found Using Hash')
|
||||
param_name=str('Not Found Using Hash')
|
||||
|
||||
model_params = model_params_set, param_name
|
||||
|
||||
return model_params
|
||||
|
||||
def provide_model_param_name(ModelName):
|
||||
#1 Band
|
||||
if '1band_sr16000_hl512' in ModelName:
|
||||
model_params_set=str('lib_v5/modelparams/1band_sr16000_hl512.json')
|
||||
param_name=str('1band_sr16000_hl512')
|
||||
elif '1band_sr32000_hl512' in ModelName:
|
||||
model_params_set=str('lib_v5/modelparams/1band_sr32000_hl512.json')
|
||||
param_name=str('1band_sr32000_hl512')
|
||||
elif '1band_sr33075_hl384' in ModelName:
|
||||
model_params_set=str('lib_v5/modelparams/1band_sr33075_hl384.json')
|
||||
param_name=str('1band_sr33075_hl384')
|
||||
elif '1band_sr44100_hl256' in ModelName:
|
||||
model_params_set=str('lib_v5/modelparams/1band_sr44100_hl256.json')
|
||||
param_name=str('1band_sr44100_hl256')
|
||||
elif '1band_sr44100_hl512' in ModelName:
|
||||
model_params_set=str('lib_v5/modelparams/1band_sr44100_hl512.json')
|
||||
param_name=str('1band_sr44100_hl512')
|
||||
elif '1band_sr44100_hl1024' in ModelName:
|
||||
model_params_set=str('lib_v5/modelparams/1band_sr44100_hl1024.json')
|
||||
param_name=str('1band_sr44100_hl1024')
|
||||
|
||||
#2 Band
|
||||
elif '2band_44100_lofi' in ModelName:
|
||||
model_params_set=str('lib_v5/modelparams/2band_44100_lofi.json')
|
||||
param_name=str('2band_44100_lofi')
|
||||
|
||||
#3 Band
|
||||
|
||||
elif '3band_44100_mid' in ModelName:
|
||||
model_params_set=str('lib_v5/modelparams/3band_44100_mid.json')
|
||||
param_name=str('3band_44100_mid')
|
||||
elif '3band_44100_msb2' in ModelName:
|
||||
model_params_set=str('lib_v5/modelparams/3band_44100_msb2.json')
|
||||
param_name=str('3band_44100_msb2')
|
||||
|
||||
#4 Band
|
||||
|
||||
elif '4band_44100_msb' in ModelName:
|
||||
model_params_set=str('lib_v5/modelparams/4band_44100_msb.json')
|
||||
param_name=str('4band_44100_msb')
|
||||
elif '4band_44100_msb2' in ModelName:
|
||||
model_params_set=str('lib_v5/modelparams/4band_44100_msb2.json')
|
||||
param_name=str('4band_44100_msb2')
|
||||
elif '4band_44100_reverse' in ModelName:
|
||||
model_params_set=str('lib_v5/modelparams/4band_44100_reverse.json')
|
||||
param_name=str('4band_44100_reverse')
|
||||
elif 'tmodelparam' in ModelName:
|
||||
model_params_set=str('lib_v5/modelparams/tmodelparam.json')
|
||||
param_name=str('User Model Param Set')
|
||||
else:
|
||||
model_params_set=str('Not Found Using Name')
|
||||
param_name=str('Not Found Using Name')
|
||||
|
||||
model_params = model_params_set, param_name
|
||||
|
||||
return model_params
|
||||
|
||||
def provide_mdx_model_param_name(modelhash):
|
||||
with open("lib_v5/filelists/hashes/mdx_original_hashes.txt", "r") as f:
|
||||
mdx_original=f.read()
|
||||
with open("lib_v5/filelists/hashes/mdx_new_hashes.txt", "r") as f:
|
||||
mdx_new=f.read()
|
||||
with open("lib_v5/filelists/hashes/mdx_new_inst_hashes.txt", "r") as f:
|
||||
mdx_new_inst=f.read()
|
||||
|
||||
if modelhash in mdx_original:
|
||||
MDX_modeltype = 'mdx_original'
|
||||
elif modelhash in mdx_new:
|
||||
MDX_modeltype = 'mdx_new'
|
||||
elif modelhash in mdx_new_inst:
|
||||
MDX_modeltype = 'mdx_new_inst'
|
||||
else:
|
||||
MDX_modeltype = 'None'
|
||||
|
||||
if MDX_modeltype == 'mdx_original':
|
||||
modeltype = 'v'
|
||||
noise_pro = 'MDX-NET_Noise_Profile_14_kHz'
|
||||
stemset_n = '(Vocals)'
|
||||
compensate = 1.03597672895
|
||||
source_val = 3
|
||||
n_fft_scale_set=6144
|
||||
dim_f_set=2048
|
||||
elif MDX_modeltype == 'mdx_new':
|
||||
modeltype = 'v'
|
||||
noise_pro = 'MDX-NET_Noise_Profile_17_kHz'
|
||||
stemset_n = '(Vocals)'
|
||||
compensate = 1.08
|
||||
source_val = 3
|
||||
n_fft_scale_set=7680
|
||||
dim_f_set=3072
|
||||
elif MDX_modeltype == 'mdx_new_inst':
|
||||
modeltype = 'v'
|
||||
noise_pro = 'MDX-NET_Noise_Profile_17_kHz'
|
||||
stemset_n = '(Instrumental)'
|
||||
compensate = 1.08
|
||||
source_val = 3
|
||||
n_fft_scale_set=7680
|
||||
dim_f_set=3072
|
||||
elif modelhash == '6f7eefc2e6b9d819ba88dc0578056ca5':
|
||||
modeltype = 'o'
|
||||
noise_pro = 'MDX-NET_Noise_Profile_Full_Band'
|
||||
stemset_n = '(Other)'
|
||||
compensate = 1.03597672895
|
||||
source_val = 2
|
||||
n_fft_scale_set=8192
|
||||
dim_f_set=2048
|
||||
elif modelhash == '72a27258a69b2381b60523a50982e9f1':
|
||||
modeltype = 'd'
|
||||
noise_pro = 'MDX-NET_Noise_Profile_Full_Band'
|
||||
stemset_n = '(Drums)'
|
||||
compensate = 1.03597672895
|
||||
source_val = 1
|
||||
n_fft_scale_set=4096
|
||||
dim_f_set=2048
|
||||
elif modelhash == '7051d7315c04285e94a97edcac3f2f76':
|
||||
modeltype = 'b'
|
||||
noise_pro = 'MDX-NET_Noise_Profile_Full_Band'
|
||||
stemset_n = '(Bass)'
|
||||
compensate = 1.03597672895
|
||||
source_val = 0
|
||||
n_fft_scale_set=16384
|
||||
dim_f_set=2048
|
||||
else:
|
||||
try:
|
||||
f = open(f"lib_v5/filelists/model_cache/mdx_model_cache/{modelhash}.json")
|
||||
mdx_model_de = json.load(f)
|
||||
modeltype = mdx_model_de["modeltype"]
|
||||
noise_pro = mdx_model_de["noise_pro"]
|
||||
stemset_n = mdx_model_de["stemset_n"]
|
||||
compensate = mdx_model_de["compensate"]
|
||||
source_val = mdx_model_de["source_val"]
|
||||
n_fft_scale_set = mdx_model_de["n_fft_scale_set"]
|
||||
dim_f_set = mdx_model_de["dim_f_set"]
|
||||
except:
|
||||
modeltype = 'Not Set'
|
||||
noise_pro = 'Not Set'
|
||||
stemset_n = 'Not Set'
|
||||
compensate = 'Not Set'
|
||||
source_val = 'Not Set'
|
||||
n_fft_scale_set='Not Set'
|
||||
dim_f_set='Not Set'
|
||||
|
||||
|
||||
model_params = modeltype, noise_pro, stemset_n, compensate, source_val, n_fft_scale_set, dim_f_set
|
||||
|
||||
return model_params
|
||||
@ -1 +0,0 @@
|
||||
temp
|
||||
@ -1 +0,0 @@
|
||||
|
||||
@ -1 +0,0 @@
|
||||
|
||||
@ -1,19 +0,0 @@
|
||||
No Model Selected
|
||||
No Model Selected
|
||||
Demucs v3: UVR Models
|
||||
Demucs v3: mdx
|
||||
Demucs v3: mdx_q
|
||||
Demucs v3: mdx_extra
|
||||
Demucs v3: mdx_extra_q
|
||||
Demucs v2: demucs
|
||||
Demucs v2: demucs_extra
|
||||
Demucs v2: demucs48_hq
|
||||
Demucs v2: tasnet
|
||||
Demucs v2: tasnet_extra
|
||||
Demucs v2: demucs_unittest
|
||||
Demucs v1: demucs
|
||||
Demucs v1: demucs_extra
|
||||
Demucs v1: light
|
||||
Demucs v1: light_extra
|
||||
Demucs v1: tasnet
|
||||
Demucs v1: tasnet_extra
|
||||
@ -1,42 +0,0 @@
|
||||
{
|
||||
"Demucs_v3_mdx_url_1": "https://dl.fbaipublicfiles.com/demucs/mdx_final/0d19c1c6-0f06f20e.th",
|
||||
"Demucs_v3_mdx_url_2": "https://dl.fbaipublicfiles.com/demucs/mdx_final/7ecf8ec1-70f50cc9.th",
|
||||
"Demucs_v3_mdx_url_3": "https://dl.fbaipublicfiles.com/demucs/mdx_final/c511e2ab-fe698775.th",
|
||||
"Demucs_v3_mdx_url_4": "https://dl.fbaipublicfiles.com/demucs/mdx_final/7d865c68-3d5dd56b.th",
|
||||
"Demucs_v3_mdx_url_5": "https://raw.githubusercontent.com/facebookresearch/demucs/main/demucs/remote/mdx.yaml",
|
||||
"Demucs_v3_mdx_q_url_1": "https://dl.fbaipublicfiles.com/demucs/mdx_final/6b9c2ca1-3fd82607.th",
|
||||
"Demucs_v3_mdx_q_url_2": "https://dl.fbaipublicfiles.com/demucs/mdx_final/b72baf4e-8778635e.th",
|
||||
"Demucs_v3_mdx_q_url_3": "https://dl.fbaipublicfiles.com/demucs/mdx_final/42e558d4-196e0e1b.th",
|
||||
"Demucs_v3_mdx_q_url_4": "https://dl.fbaipublicfiles.com/demucs/mdx_final/305bc58f-18378783.th",
|
||||
"Demucs_v3_mdx_q_url_5": "https://raw.githubusercontent.com/facebookresearch/demucs/main/demucs/remote/mdx_q.yaml",
|
||||
"Demucs_v3_mdx_extra_url_1": "https://dl.fbaipublicfiles.com/demucs/mdx_final/e51eebcc-c1b80bdd.th",
|
||||
"Demucs_v3_mdx_extra_url_2": "https://dl.fbaipublicfiles.com/demucs/mdx_final/a1d90b5c-ae9d2452.th",
|
||||
"Demucs_v3_mdx_extra_url_3": "https://dl.fbaipublicfiles.com/demucs/mdx_final/5d2d6c55-db83574e.th",
|
||||
"Demucs_v3_mdx_extra_url_4": "https://dl.fbaipublicfiles.com/demucs/mdx_final/cfa93e08-61801ae1.th",
|
||||
"Demucs_v3_mdx_extra_url_5": "https://raw.githubusercontent.com/facebookresearch/demucs/main/demucs/remote/mdx_extra.yaml",
|
||||
"Demucs_v3_mdx_extra_q_url_1": "https://dl.fbaipublicfiles.com/demucs/mdx_final/83fc094f-4a16d450.th",
|
||||
"Demucs_v3_mdx_extra_q_url_2": "https://dl.fbaipublicfiles.com/demucs/mdx_final/464b36d7-e5a9386e.th",
|
||||
"Demucs_v3_mdx_extra_q_url_3": "https://dl.fbaipublicfiles.com/demucs/mdx_final/14fc6a69-a89dd0ee.th",
|
||||
"Demucs_v3_mdx_extra_q_url_4": "https://dl.fbaipublicfiles.com/demucs/mdx_final/7fd6ef75-a905dd85.th",
|
||||
"Demucs_v3_mdx_extra_q_url_5": "https://raw.githubusercontent.com/facebookresearch/demucs/main/demucs/remote/mdx_extra_q.yaml",
|
||||
"Demucs_v3_UVR_url_1": "https://github.com/TRvlvr/model_repo/releases/download/all_public_uvr_models/ebf34a2d.th",
|
||||
"Demucs_v3_UVR_url_2": "https://github.com/TRvlvr/model_repo/releases/download/all_public_uvr_models/ebf34a2db.th",
|
||||
"Demucs_v3_UVR_url_3": "https://github.com/TRvlvr/model_repo/releases/download/all_public_uvr_models/UVR_Demucs_Model_1.yaml",
|
||||
"Demucs_v3_UVR_url_4": "https://github.com/TRvlvr/model_repo/releases/download/all_public_uvr_models/UVR_Demucs_Model_2.yaml",
|
||||
"Demucs_v3_UVR_url_5": "https://github.com/TRvlvr/model_repo/releases/download/all_public_uvr_models/UVR_Demucs_Model_Bag.yaml",
|
||||
"Demucs_v2_demucs_url_1": "https://dl.fbaipublicfiles.com/demucs/v3.0/demucs-e07c671f.th",
|
||||
"Demucs_v2_demucs_extra_url_1": "https://dl.fbaipublicfiles.com/demucs/v3.0/demucs_extra-3646af93.th",
|
||||
"Demucs_v2_demucs48_hq_url_1": "https://dl.fbaipublicfiles.com/demucs/v3.0/demucs48_hq-28a1282c.th",
|
||||
"Demucs_v2_tasnet_url_1": "https://dl.fbaipublicfiles.com/demucs/v3.0/tasnet-beb46fac.th",
|
||||
"Demucs_v2_tasnet_extra_url_1": "https://dl.fbaipublicfiles.com/demucs/v3.0/tasnet_extra-df3777b2.th",
|
||||
"Demucs_v2_demucs_unittest_url_1": "https://dl.fbaipublicfiles.com/demucs/v3.0/demucs_unittest-09ebc15f.th",
|
||||
"Demucs_v1_demucs_url_1": "https://dl.fbaipublicfiles.com/demucs/v2.0/demucs.th",
|
||||
"Demucs_v1_demucs_extra_url_1": "https://dl.fbaipublicfiles.com/demucs/v2.0/demucs_extra.th",
|
||||
"Demucs_v1_light_url_1": "https://dl.fbaipublicfiles.com/demucs/v2.0/light.th",
|
||||
"Demucs_v1_light_extra_url_1": "https://dl.fbaipublicfiles.com/demucs/v2.0/light_extra.th",
|
||||
"Demucs_v1_tasnet_url_1": "https://dl.fbaipublicfiles.com/demucs/v2.0/tasnet.th",
|
||||
"Demucs_v1_tasnet_extra_url_1": "https://dl.fbaipublicfiles.com/demucs/v2.0/tasnet_extra.th",
|
||||
"model_repo_url_1": "https://github.com/TRvlvr/model_repo/releases/download/model_pack_repo/",
|
||||
"single_model_repo_url_1": "https://github.com/TRvlvr/model_repo/releases/download/all_public_uvr_models/",
|
||||
"app_patch": "https://github.com/TRvlvr/model_repo/releases/download/uvr_update_patches/"
|
||||
}
|
||||
@ -1,13 +0,0 @@
|
||||
No Model Selected
|
||||
No Model Selected
|
||||
MDX-Net Model: UVR_MDXNET_Main
|
||||
MDX-Net Model: UVR_MDXNET_1_9703
|
||||
MDX-Net Model: UVR_MDXNET_2_9682
|
||||
MDX-Net Model: UVR_MDXNET_3_9662
|
||||
MDX-Net Model: UVR_MDXNET_9482
|
||||
MDX-Net Model: UVR_MDXNET_KARA
|
||||
Developer Pack: voc_model_epochs_434-439
|
||||
Developer Pack: inst_model_epochs_10-17
|
||||
Developer Pack: inst_model_epochs_26-40
|
||||
Developer Pack: inst_model_epochs_51-57
|
||||
Developer Pack: inst_model_epochs_58-64
|
||||
@ -1 +0,0 @@
|
||||
temp
|
||||
@ -1,25 +0,0 @@
|
||||
No Model Selected
|
||||
No Model Selected
|
||||
VR Arch Model Pack v5: SP Models
|
||||
VR Arch Model Pack v5: HP2 Models
|
||||
VR Arch Model Pack v4: Main Models
|
||||
VR Arch Single Model v5: 1_HP-UVR
|
||||
VR Arch Single Model v5: 2_HP-UVR
|
||||
VR Arch Single Model v5: 3_HP-Vocal-UVR
|
||||
VR Arch Single Model v5: 4_HP-Vocal-UVR
|
||||
VR Arch Single Model v5: 5_HP-Karaoke-UVR
|
||||
VR Arch Single Model v5: 6_HP-Karaoke-UVR
|
||||
VR Arch Single Model v5: 7_HP2-UVR
|
||||
VR Arch Single Model v5: 8_HP2-UVR
|
||||
VR Arch Single Model v5: 9_HP2-UVR
|
||||
VR Arch Single Model v5: 10_SP-UVR-2B-32000-1
|
||||
VR Arch Single Model v5: 11_SP-UVR-2B-32000-2
|
||||
VR Arch Single Model v5: 12_SP-UVR-3B-44100
|
||||
VR Arch Single Model v5: 13_SP-UVR-4B-44100-1
|
||||
VR Arch Single Model v5: 14_SP-UVR-4B-44100-2
|
||||
VR Arch Single Model v5: 15_SP-UVR-MID-44100-1
|
||||
VR Arch Single Model v5: 16_SP-UVR-MID-44100-2
|
||||
VR Arch Single Model v4: MGM_HIGHEND_v4
|
||||
VR Arch Single Model v4: MGM_LOWEND_A_v4
|
||||
VR Arch Single Model v4: MGM_LOWEND_B_v4
|
||||
VR Arch Single Model v4: MGM_MAIN_v4
|
||||
@ -1,16 +0,0 @@
|
||||
No Model
|
||||
No Model
|
||||
MDX-Net: UVR-MDX-NET 1
|
||||
MDX-Net: UVR-MDX-NET 2
|
||||
MDX-Net: UVR-MDX-NET 3
|
||||
MDX-Net: UVR_MDXNET_9482
|
||||
MDX-Net: UVR-MDX-NET Karaoke
|
||||
MDX-Net: bass
|
||||
Demucs: UVR_Demucs_Model_1
|
||||
Demucs: UVR_Demucs_Model_2
|
||||
Demucs: UVR_Demucs_Model_Bag
|
||||
Demucs: Demucs_unittest v2
|
||||
Demucs: mdx_extra
|
||||
Demucs: mdx_q
|
||||
Demucs: Tasnet v1
|
||||
Demucs: Tasnet_extra v1
|
||||
@ -1,13 +0,0 @@
|
||||
No Model
|
||||
No Model
|
||||
1_HP-UVR
|
||||
2_HP-UVR
|
||||
6_HP-Karaoke-UVR
|
||||
11_SP-UVR-2B-32000-2
|
||||
13_SP-UVR-4B-44100-1
|
||||
MGM_HIGHEND_v4 (1)
|
||||
MGM_HIGHEND_v4
|
||||
MGM_LOWEND_A_v4
|
||||
MGM_LOWEND_B_v4
|
||||
MGM_MAIN_v4
|
||||
WIP-Piano-4band-129605kb
|
||||
@ -1,266 +0,0 @@
|
||||
0374836ab954ccd5946b29668c36bfdd
|
||||
e494001f6e38f3ff9f8db07073dbcc38
|
||||
1284cfb6ca5cbe68778a392bf92fe448
|
||||
927f5993571090980f7042a5087ad7d5
|
||||
8a7b42b609383c9bd11c093e357c8d01
|
||||
d7b9b49bf75c78afad0e88786bd729ae
|
||||
8737fb18a216d62ba2fec7948da59721
|
||||
432308a1f31d144ba117dc5f87e97809
|
||||
39eedabe4309ecbcda1df37a5cdd6adf
|
||||
fbee5264485e2c4b4003db897dea3528
|
||||
443ca4b02bf26bc42d9ef114aa25adc8
|
||||
8d700d4589ba30b876114823af40bcea
|
||||
52be0e4c18ddfba376081a7edaca5f56
|
||||
f0fd48dc0422a992e08e18f9cf318678
|
||||
af21b8645202d0b6d93cb383284ad0e6
|
||||
c5f39697c5dd372c4fb9656f01c814cc
|
||||
3f2bd7da922c4b45ac8923cccce41455
|
||||
8c467b5fbce83858b67436e28efa31c3
|
||||
16418df565fbaddd8e91ead6693a7418
|
||||
06f4c4b9e7cc2c3503bdcb9cfad4298b
|
||||
3a28db13d74c27f4ef255e5acb2f66b1
|
||||
2c7e0a31f0aa7947865b66a8ccfdf08f
|
||||
ef878de3d28e6ef376ad35527d86e4dc
|
||||
52b4c815669b7a6c1898453f5c42a617
|
||||
63c28c788f40af6d49ad8013600149bf
|
||||
787c1fa7256250695b14d2c62d57070c
|
||||
3eac6076f597da01788946563cc2756d
|
||||
d3efcccba9417bd0567845958d55bed5
|
||||
5d4b6d50847807833b82009be6519364
|
||||
cd988a4722f06c7fc9798750da9a6216
|
||||
5c0953d6f2d6a6ac3c738c0ba04fb518
|
||||
c488149f2cd37c710da574f3b784e6e3
|
||||
e46ef7b671b13a9a6bada197d8ccc0cf
|
||||
a57f5884b67bc0be16759045d9734b7f
|
||||
9b60b21391809845bb4435bc7096aabf
|
||||
5a4897d7a3afcae86f6948ba9275a49e
|
||||
d521fcb9b6cb6818710bf1609d1a8182
|
||||
3c6420e4e3646d7906d6d1b86d0410bc
|
||||
a85c8174b2c91e780a67b391650ae264
|
||||
b5a906653199f49f0e0215e0b22901b4
|
||||
16de23223764249902734d64f2a0223f
|
||||
2957853f7346c5039fc2f4d1b0f3c65c
|
||||
a8fb404acc3631cfbb8aacf20f39d65d
|
||||
f1e62e9e52477dff0c1b92c218bcd7f0
|
||||
2067c316fe7a097040f6da35451bde94
|
||||
b6653ac3158e60561f05933ebe4d6477
|
||||
274f631b444ce3ec6fbb71778ca58ae5
|
||||
5dc5f89788c7385a99787a15e78cc852
|
||||
740e38a99d835f113ffc396e144fbfa8
|
||||
b21de390242db154b1e6d45b9c44a912
|
||||
8c6751c184707c356534a8d4481e2987
|
||||
057a0878236e70dfdf8e09fc8e3d9126
|
||||
0a15368a0d7b00eb1f49276c316c280f
|
||||
ae35d179e0395e53867f2b7f32de337e
|
||||
87a9c56e36d2cb159ff76929b0edfd3f
|
||||
c488149f2cd37c710da574f3b784e6e3
|
||||
18508ba57b14327516f4f958d7cb5bdd
|
||||
d2b8a00978bcf61eeb3642ffb2653431
|
||||
c1cbaeea910a925be045c4dad52e0f95
|
||||
72d739c4fa1f0c606f357cc6b4b313b0
|
||||
44d6519e1efc6c68d53f82361e382987
|
||||
7b63504b3dd9ed7ef5c6ef9fbbb6d857
|
||||
f0fb53226a97e2d1ed8af46308d0b238
|
||||
27065f326f173ee54c38e349a84caed8
|
||||
71f2b8871d21ce1ae9b3c16042d3b511
|
||||
870a12746e7b7648cc8449c0ea5f8be8
|
||||
93bcd06e20b2ae5388da783bf68c7224
|
||||
09973b74f4dabb01ecdfed706743f45e
|
||||
34c8c8870ed8cac0fbc1dc9d1ae90dec
|
||||
e3078ba5c104e5f4ec7ff17b23f4bc48
|
||||
7a84d9cc7d2a064be2328d5f183e21f8
|
||||
49efdbcfcf32bde3b0cdec517fc14f7d
|
||||
389c7c5b890a73fec8753a675dd12ac0
|
||||
34ec0124d7d734babf76c79ab94c1f5c
|
||||
2ebfd30a6f0dcae400bdadf27c138d72
|
||||
a6520bca33f5d9d47249c9dc72769869
|
||||
fb95e55c41d7b4609da78f7e189d57c7
|
||||
82ea77ce031ff7f388ca61d8022d0783
|
||||
ca4225e905d87de17c4495af7763f853
|
||||
066f3b8bc8e25d949c73ec1995072903
|
||||
1e4cccb2b9279c0f3d120e3df42eade5
|
||||
7148b2d2a566a500df4046b57d64c921
|
||||
4d5b28ed4facc34105c548b4ebe445d3
|
||||
545fcdc3f4214d1f19b5fdd21cde38bf
|
||||
c84cea02dae21a7c94d5cbc0de4cebfc
|
||||
19cb138773f3db150366f8ed3e1b3b44
|
||||
72f8c3d5486fe653c8bb99519097eb93
|
||||
1205327ef4f1b2b12710c4d648b71de2
|
||||
6b2a00d20b89c27e9b4185c2cf9e4aa7
|
||||
eae02d1f9b056ed813f864d30f33834d
|
||||
1c3fecce7632cc859b75c7b88ad2afa2
|
||||
ff62c0bcbcc4f685dd3b7831a0f7c3b7
|
||||
6fbc36b6b8c22477ff32dac19e4fca28
|
||||
7a205daabf8960bf487b9c0b794251bc
|
||||
cc1c53adf5c99f2264f5e01dc1082a5b
|
||||
37822839f05c3befd0c99b9f4882ab86
|
||||
ee0712096e63be0afd56a74f4a9e0ea1
|
||||
c58340ff0f0507240e3a146bdd990cf1
|
||||
1a13fd81b213d7163b49db5da00f53c3
|
||||
b77cf1f489cdde9eb8a9b67b6b59ec68
|
||||
c488149f2cd37c710da574f3b784e6e3
|
||||
ae3bb6e6e27b272999c31ef153a3c87a
|
||||
299a55c8ede63c7897f018b45cb2d5d2
|
||||
679042229a62536cc8edbfc04efd4db6
|
||||
a8bd219e9c47f19d7a6114e423ad408f
|
||||
a55dd195a4545ee38d4cf35b1cc2ec63
|
||||
583d824859c4312bf9f29e33c9e5bfb9
|
||||
4c43c3b27dc8f806a629bf1fed00713f
|
||||
3e50a377fe876b78b0372275e4280deb
|
||||
0aa28867b9527b9d3039d51b64c539bd
|
||||
254ece16803e5c86ab5e848b7ed050f7
|
||||
231532acc25316bec842fb9daf8a2433
|
||||
ff07a7503a0f5242c20d3120e31edaef
|
||||
f3d967caf9a18cca0464c94af5e29b96
|
||||
f482c7a3e906bbd5f7f84344de95884e
|
||||
47541c9bfc1f64fb27ed7fe0e51d5820
|
||||
d11dd98a13e6931844d099162e944175
|
||||
db59cbd3343a3e1c4077e97273947c07
|
||||
c488149f2cd37c710da574f3b784e6e3
|
||||
684895a24ed48e27092604e5ca55147b
|
||||
f56742ceec2bb09e4a8f6bfebd19e3b2
|
||||
0938f3adcf49644ee5952fce8470e1ee
|
||||
18c56f68d4a1ca021ef0daffd586e5a1
|
||||
2fd39f65bbdfb44ae24952fc55ec828f
|
||||
56c8dd32289acfc89ddb1cf1eb2080a1
|
||||
49df79ad1f22a1cc51d3fe6581c04368
|
||||
b39b1edd3defacc1035fedcb9717c13f
|
||||
493598ca0cfd5a4e3a308cf2d7b337cf
|
||||
027e12756f7ac9d72a0243e57208d640
|
||||
2c3920c38cdb3dac251d35f7cad32d76
|
||||
7399b3d9512206b26f885cbbbd5db8a6
|
||||
857af50fb527a7b187e2038d1f5339fe
|
||||
706bd84cee98aa24862c0d84b3ad7d83
|
||||
1310c0587c85752e2b2a958cbb8a865b
|
||||
08e4bf0e8b3d35897f3165db19e2fe45
|
||||
cbbdd039f02090f130a774609f546073
|
||||
1f98187a1cb50b04fcf9b6483aede1fe
|
||||
77c77d58fb898a936f534098c7cedde3
|
||||
68e0861a93e8f69d18e06446dabb4bfd
|
||||
d66fcc3a6869fc9e16e412f7102e7a4b
|
||||
c488149f2cd37c710da574f3b784e6e3
|
||||
cbf08ceba8462bc3b63a4b8219b2127f
|
||||
1c8c65dccc65a874baf6b946323b9fcc
|
||||
29d53240deda5d41501fcf0a766adfce
|
||||
9c555f81fba02566811902d04c623d6c
|
||||
282ba1496fd4242ce2cf2d90e57fe2df
|
||||
5a071be7bdc03e9f10cfabc35256562f
|
||||
24154e229b32e53bc9418ec787e17494
|
||||
e4bb05aef0f088bc085bd70fdf0e23e5
|
||||
75a2e9b4a66c0906aa05841a08796f16
|
||||
76fcb7ec78d402d0fcbab783e6a5bb08
|
||||
afcd755eb6fe6fbd63064d62b97d7415
|
||||
2f749b5bba3fbdf47b6f1786682b6e17
|
||||
d7c92cee09b62970b9dec9290507f1ab
|
||||
4edc1683c236af337baa583a01120d4c
|
||||
ab9bc82a525b84ffddb8e3e0d3cfb9d6
|
||||
12b4b7b4569816cb499a051203f84163
|
||||
7dc1829acd3fc5ed7aee352383464a8f
|
||||
77654e582fa565e1878880c2dc7ccc6a
|
||||
1af9ca98f02117c715a7d08a93a90d11
|
||||
26ee73d5f2cbff2f833e83e3d9d9a970
|
||||
aa04ea15b9607d920ae5ae0e826e2b96
|
||||
602bd5c724ea9d4c2f446c9491ee35f2
|
||||
5a5770df2a95203d87f383ea4bac1c13
|
||||
5c3aac1983493dca30d0512e32d73aec
|
||||
317c5596068ee1d89f1e134f34c41a6f
|
||||
cb7622696ed1a09922c8bf9e08dab19a
|
||||
a10993b3fc9e713175e8c15369208e37
|
||||
54f94ef6ae99039746b29e8c7b095f42
|
||||
fcdf005de9869614d418fda6921736b1
|
||||
738526a1206737e49ed2e3ed525cef75
|
||||
561075aa5e8c15da84a793cc351ff7f1
|
||||
6c31403d96f13e0f99cf70caba5dbcf4
|
||||
d2f6e3956e13abcaa5854061ec66b328
|
||||
eeb018b28ae4d0aeec0a5644f2784319
|
||||
33b6e5e6677d03a5210f04eb9f2b7d25
|
||||
a93ff7868bbe0707e01f23b1bd420dc9
|
||||
3fcc378ca8ff8d2e57e2768393a78f01
|
||||
759ca61e56317dc0c1dfbaec891e0b6d
|
||||
fd5e636fbb4c6eea0383f7803f32649f
|
||||
743f50f5512c8b8d8a045c8987e2f490
|
||||
cc40b9e052b524fa882b812be835e2b1
|
||||
2299e5fd3ff04cdac5473479dab4ca12
|
||||
80ed2351ed5645b10ee59c49de2fa85e
|
||||
72e70ae8fcea406a4175cb8aff365f26
|
||||
ce7eab5316ddbe8492a911fd6e9360ec
|
||||
0a3db22db2ec36d2274be2e6c1a3f281
|
||||
7accc6b329e87fd488a8611be2822c31
|
||||
9d75768514ae1e9d815d44770c2c7528
|
||||
2c099e9ad0ef0ed8fbbe5207d86bf1c3
|
||||
1735d7847e15f75103dab33ae7812593
|
||||
4ccf7af16a1e132915a53b5c1d0489b4
|
||||
9d6dc0eeb89233b2fc2e4cb6b0c03c8d
|
||||
9c174a9728ba905330d53404350e5030
|
||||
b1e90f473873c0de27feadd42f784cde
|
||||
d96712e0a516217a4b15f21da3e09dee
|
||||
9c64543bc5757fd9bd385ad1977d98f6
|
||||
6f3c09ff2b603f88f9cb59643bf60d27
|
||||
195710ebfdac31abc9a8c1cf70af64c5
|
||||
dc69356b009931763e60494ed65b091b
|
||||
8700c86a88cc28873a3fa38dddeb657d
|
||||
d2deadd2e486200fad9d7504a6be272b
|
||||
94c1f1fcc101819c1b0ea619b1d14a4d
|
||||
bd1f3fcfb0080ad24fbaf522f46d93d0
|
||||
7ab6e91e95dc46c43f6fd453a7fa08d2
|
||||
b62bf9ecb1504fafcbb2f1dfd3577adb
|
||||
d79e834a074be79732214c241acbafae
|
||||
248fbfb44e4c7b6252c43f6efd57beeb
|
||||
9d0a3bf0fbfd3ac5a326da42730c59d8
|
||||
0381868fc1424a5a2cbddc0dcd3e03ff
|
||||
99fe35974f3b3e44222d3ea922361441
|
||||
ee649ecf9af62537230d53a1669aded9
|
||||
7c603f02ecae615a6e6db193c833049d
|
||||
70643a56ccda1e41c853b626fc61b2b7
|
||||
7aede03b88b41aed480873dcc1155ff1
|
||||
f82eb261734ae9506f3a2b0982f5a436
|
||||
b7e0ea765ce79b16b3360b6f7c20bf3e
|
||||
e42f1bb671e9007a58a56a0dbcfba96a
|
||||
4944c8ab32140b11ea96055b536195bd
|
||||
2398160da9da6e05ed4cf41c313ab359
|
||||
83640e4eebfa904a335fcce7308f76ee
|
||||
b79bcd180ab4105bc7d477594366ac32
|
||||
9ea57e40b2a085a34432903153b6f553
|
||||
a55bd0247b2b97299a21ca33ea4153d6
|
||||
b917bdac8ecdd58d62c662604fe41156
|
||||
de0e9b8d929f8ea4c6e9fbdbc5702539
|
||||
a1f86a115d6771f2f49c192187433937
|
||||
3c0441d69f6f21890f07b791a2a0f975
|
||||
df50c0c577db9620d65c461514753cc6
|
||||
e2c2fdddc294a14e81ab638877746897
|
||||
6289fc9313430c398a5a154a4184e643
|
||||
45858d3c16515a018a55bdd63a8dcaa0
|
||||
094235d47192c91bd2bfeaa2005c4a95
|
||||
e50ae9b60d3a2527545c2a9c3c840947
|
||||
38e807c4d51ea851182e89b2a26823c9
|
||||
7b0e3b866a0ad9aedac142b017a46d23
|
||||
82f53c138d9ae0442f42357538c2623a
|
||||
e6dd1e0d8fd1dc6ace7026f8749c1fed
|
||||
221c01f90c00e98ceccef860b33ecb76
|
||||
1d39503ae9334f6eee559bdd70f8c24f
|
||||
f4b9da2652eb0dce90a9ea3335cd36c0
|
||||
d403bb7dcbe901eb8f248e87c03fb7cd
|
||||
41e7fab08927a26a7afd6a8830bc2d8c
|
||||
539e6c09143a295cea430c7459bbff33
|
||||
fd83d05f82a4edf4a01c56ce3a8d6e63
|
||||
8fead59e6616e753f9f1c27156353947
|
||||
61be36cf0a2880d4b38c575b65699b4a
|
||||
83232153c59247187570377af0a1d223
|
||||
dc9777a4cef96dc2f46a5381ba058256
|
||||
322d9cd44c707f6b54e55775700ca514
|
||||
7587ebafebed7d9bdaa6165db2c43608
|
||||
83ddc19beea20010318bce4722d890d5
|
||||
bb64c8f176f3c4d87ff260be720d66c3
|
||||
202c811375f61cd6d5840a40cf2cb4ba
|
||||
7b3460f9e8dd28202111ec582e965d55
|
||||
dd087f5adcd94116225963397ac25889
|
||||
5e5865e8d2f48d93066c866aa31c5983
|
||||
ff2d856e56b5768688a1d1de00c143dc
|
||||
90245b98df873843287f68caf9da596b
|
||||
fb793952993781dbee22336bc6296987
|
||||
5107088c9476c07bdd32af1b9d67108e
|
||||
969bfa927f27ca1c98c00d3ae589f623
|
||||
37969f2b2add6199fb00780899ce8c0d
|
||||
122742f2e218d003cc22d108df9c668a
|
||||
c72c77ff2373bc0e0d9beebf35c0dc07
|
||||
43eefd064d333877c59ee7362d044e0c
|
||||
@ -1,56 +0,0 @@
|
||||
c488149f2cd37c710da574f3b784e6e3
|
||||
2223f985aa47e99a0348f987ea2a3db6
|
||||
3fc1fa83cebb144dca7d379a3767ac77
|
||||
dfe5e1f42c42dcc262f2a2e2bc4ca7aa
|
||||
69108243963e4ed3fe22fdb37b812fe8
|
||||
ec3d6dc24c75e06a72fe1a2bd970b7b0
|
||||
aee23fcb90ef135bbd036f6e1c1ad9f9
|
||||
4561b9240c9acaf13eee2bd0186b4df4
|
||||
da0f43e81eb60b0bdacedd2de4758ee9
|
||||
5ee77a84e40bbd66a9d2b08b6a4c8496
|
||||
8b0526d02937c08adc95dfd57652c915
|
||||
08ff1d3a743eb2377c96b3de793990a0
|
||||
af292ce84aa6125ab75020f31e183f5b
|
||||
7f56ee6b9ee402802c403f348fea58cb
|
||||
038bb0d0a9f3c89b5671189384cfdf91
|
||||
db79054b3ae6c30f655e8ff12d64caa7
|
||||
b2e16d43ef559782a32874ffdace8b82
|
||||
c488149f2cd37c710da574f3b784e6e3
|
||||
b71bb782cad9b8c6ea0a72c6ae69e8b4
|
||||
9d478026519f140e14e9c1bb32fcc522
|
||||
e81c4b46ec685b3ce227f426f884cfe0
|
||||
e9edcfacbbbbc513734082e1c1f7f6ad
|
||||
8d0bb54171a0996f2610be0fdc5743a8
|
||||
fbe5c4eecc1f3ad4b38364d77a774013
|
||||
66209e099302542c627c29bbb6976c59
|
||||
ee7b7dcda4e4353730cf89afc881c8cb
|
||||
0d09ab08de2cf0efe123144589345a33
|
||||
be2cff51fc6fcae172d30d3f1c142ee6
|
||||
b53e4cc77f9c47cbc6184cc7158f65ca
|
||||
c38f6a06977b9bc287c83f1e80ee0d9f
|
||||
0e05e776801b7016714015043e97373b
|
||||
fd39ea4282a92d3cb5f8d502e4c44ffa
|
||||
c937b7662b6f5cd92ff82fe595f160e7
|
||||
c0e0e9c44f815c1d0c0737343d702923
|
||||
c488149f2cd37c710da574f3b784e6e3
|
||||
823ef02eadf24f652609959e3a35206e
|
||||
674e8851d69bef3e4e1724d21247a8dc
|
||||
d3543c7b5d214515a7bd55d57b2413a0
|
||||
eb1179fb56fbcab660d9b69f9ac9cbe7
|
||||
0f7f0b21feb6b1b6e843f6047a3af13a
|
||||
b6f8de9a8316d5bd330a84eb22078ef4
|
||||
6dd6104d3ba4b3e9f47fba7d6c84dd9c
|
||||
1579df63864bf6ec11f2484cb2cfca7e
|
||||
3d2af588b96dc0e84be66804684e7c56
|
||||
6e566b37b3cec881ec93cfe415a710ec
|
||||
455c272b691c1001aa9b9cad5dfedd20
|
||||
ec892e0ea6f973d8c15645e621ee8fe1
|
||||
4a0b13b03e4db47191f387c2ced82f73
|
||||
e5d6e895bcbe4ca62ca9cc8808d8da6e
|
||||
4a0b13b03e4db47191f387c2ced82f73
|
||||
e5d6e895bcbe4ca62ca9cc8808d8da6e
|
||||
9df9f3bf4c7151fc36fb822d7728f433
|
||||
4475c1d3f79482cb2082f1bd13899e1b
|
||||
3d3523c8e0ab0a355748f38449331918
|
||||
24bb2808feae6efb2aaae9db3778886c
|
||||
947e6dd9e4aea2811eb3fb26d4bde615
|
||||
@ -1,5 +0,0 @@
|
||||
1bbcb39d8a4be721d9322e62f13de1c1
|
||||
94422d1d6eb7019eff97dbef2daba979
|
||||
d3b87173f484864674ee2a21cd7b35f2
|
||||
053f663b23c70c6c1f52938fb480f5b8
|
||||
76929c1b5b9b804f89f4ebb78712c668
|
||||
@ -1 +0,0 @@
|
||||
cache_goes_here
|
||||
@ -1 +0,0 @@
|
||||
cache_goes_here
|
||||
Binary file not shown.
Binary file not shown.
Binary file not shown.
Binary file not shown.
Binary file not shown.
Binary file not shown.
Binary file not shown.
Binary file not shown.
116
lib_v5/layers.py
116
lib_v5/layers.py
@ -1,116 +0,0 @@
|
||||
import torch
|
||||
from torch import nn
|
||||
import torch.nn.functional as F
|
||||
|
||||
from lib_v5 import spec_utils
|
||||
|
||||
|
||||
class Conv2DBNActiv(nn.Module):
|
||||
|
||||
def __init__(self, nin, nout, ksize=3, stride=1, pad=1, dilation=1, activ=nn.ReLU):
|
||||
super(Conv2DBNActiv, self).__init__()
|
||||
self.conv = nn.Sequential(
|
||||
nn.Conv2d(
|
||||
nin, nout,
|
||||
kernel_size=ksize,
|
||||
stride=stride,
|
||||
padding=pad,
|
||||
dilation=dilation,
|
||||
bias=False),
|
||||
nn.BatchNorm2d(nout),
|
||||
activ()
|
||||
)
|
||||
|
||||
def __call__(self, x):
|
||||
return self.conv(x)
|
||||
|
||||
|
||||
class SeperableConv2DBNActiv(nn.Module):
|
||||
|
||||
def __init__(self, nin, nout, ksize=3, stride=1, pad=1, dilation=1, activ=nn.ReLU):
|
||||
super(SeperableConv2DBNActiv, self).__init__()
|
||||
self.conv = nn.Sequential(
|
||||
nn.Conv2d(
|
||||
nin, nin,
|
||||
kernel_size=ksize,
|
||||
stride=stride,
|
||||
padding=pad,
|
||||
dilation=dilation,
|
||||
groups=nin,
|
||||
bias=False),
|
||||
nn.Conv2d(
|
||||
nin, nout,
|
||||
kernel_size=1,
|
||||
bias=False),
|
||||
nn.BatchNorm2d(nout),
|
||||
activ()
|
||||
)
|
||||
|
||||
def __call__(self, x):
|
||||
return self.conv(x)
|
||||
|
||||
|
||||
class Encoder(nn.Module):
|
||||
|
||||
def __init__(self, nin, nout, ksize=3, stride=1, pad=1, activ=nn.LeakyReLU):
|
||||
super(Encoder, self).__init__()
|
||||
self.conv1 = Conv2DBNActiv(nin, nout, ksize, 1, pad, activ=activ)
|
||||
self.conv2 = Conv2DBNActiv(nout, nout, ksize, stride, pad, activ=activ)
|
||||
|
||||
def __call__(self, x):
|
||||
skip = self.conv1(x)
|
||||
h = self.conv2(skip)
|
||||
|
||||
return h, skip
|
||||
|
||||
|
||||
class Decoder(nn.Module):
|
||||
|
||||
def __init__(self, nin, nout, ksize=3, stride=1, pad=1, activ=nn.ReLU, dropout=False):
|
||||
super(Decoder, self).__init__()
|
||||
self.conv = Conv2DBNActiv(nin, nout, ksize, 1, pad, activ=activ)
|
||||
self.dropout = nn.Dropout2d(0.1) if dropout else None
|
||||
|
||||
def __call__(self, x, skip=None):
|
||||
x = F.interpolate(x, scale_factor=2, mode='bilinear', align_corners=True)
|
||||
if skip is not None:
|
||||
skip = spec_utils.crop_center(skip, x)
|
||||
x = torch.cat([x, skip], dim=1)
|
||||
h = self.conv(x)
|
||||
|
||||
if self.dropout is not None:
|
||||
h = self.dropout(h)
|
||||
|
||||
return h
|
||||
|
||||
|
||||
class ASPPModule(nn.Module):
|
||||
|
||||
def __init__(self, nin, nout, dilations=(4, 8, 16), activ=nn.ReLU):
|
||||
super(ASPPModule, self).__init__()
|
||||
self.conv1 = nn.Sequential(
|
||||
nn.AdaptiveAvgPool2d((1, None)),
|
||||
Conv2DBNActiv(nin, nin, 1, 1, 0, activ=activ)
|
||||
)
|
||||
self.conv2 = Conv2DBNActiv(nin, nin, 1, 1, 0, activ=activ)
|
||||
self.conv3 = SeperableConv2DBNActiv(
|
||||
nin, nin, 3, 1, dilations[0], dilations[0], activ=activ)
|
||||
self.conv4 = SeperableConv2DBNActiv(
|
||||
nin, nin, 3, 1, dilations[1], dilations[1], activ=activ)
|
||||
self.conv5 = SeperableConv2DBNActiv(
|
||||
nin, nin, 3, 1, dilations[2], dilations[2], activ=activ)
|
||||
self.bottleneck = nn.Sequential(
|
||||
Conv2DBNActiv(nin * 5, nout, 1, 1, 0, activ=activ),
|
||||
nn.Dropout2d(0.1)
|
||||
)
|
||||
|
||||
def forward(self, x):
|
||||
_, _, h, w = x.size()
|
||||
feat1 = F.interpolate(self.conv1(x), size=(h, w), mode='bilinear', align_corners=True)
|
||||
feat2 = self.conv2(x)
|
||||
feat3 = self.conv3(x)
|
||||
feat4 = self.conv4(x)
|
||||
feat5 = self.conv5(x)
|
||||
out = torch.cat((feat1, feat2, feat3, feat4, feat5), dim=1)
|
||||
bottle = self.bottleneck(out)
|
||||
return bottle
|
||||
@ -1,116 +0,0 @@
|
||||
import torch
|
||||
from torch import nn
|
||||
import torch.nn.functional as F
|
||||
|
||||
from lib_v5 import spec_utils
|
||||
|
||||
|
||||
class Conv2DBNActiv(nn.Module):
|
||||
|
||||
def __init__(self, nin, nout, ksize=3, stride=1, pad=1, dilation=1, activ=nn.ReLU):
|
||||
super(Conv2DBNActiv, self).__init__()
|
||||
self.conv = nn.Sequential(
|
||||
nn.Conv2d(
|
||||
nin, nout,
|
||||
kernel_size=ksize,
|
||||
stride=stride,
|
||||
padding=pad,
|
||||
dilation=dilation,
|
||||
bias=False),
|
||||
nn.BatchNorm2d(nout),
|
||||
activ()
|
||||
)
|
||||
|
||||
def __call__(self, x):
|
||||
return self.conv(x)
|
||||
|
||||
|
||||
class SeperableConv2DBNActiv(nn.Module):
|
||||
|
||||
def __init__(self, nin, nout, ksize=3, stride=1, pad=1, dilation=1, activ=nn.ReLU):
|
||||
super(SeperableConv2DBNActiv, self).__init__()
|
||||
self.conv = nn.Sequential(
|
||||
nn.Conv2d(
|
||||
nin, nin,
|
||||
kernel_size=ksize,
|
||||
stride=stride,
|
||||
padding=pad,
|
||||
dilation=dilation,
|
||||
groups=nin,
|
||||
bias=False),
|
||||
nn.Conv2d(
|
||||
nin, nout,
|
||||
kernel_size=1,
|
||||
bias=False),
|
||||
nn.BatchNorm2d(nout),
|
||||
activ()
|
||||
)
|
||||
|
||||
def __call__(self, x):
|
||||
return self.conv(x)
|
||||
|
||||
|
||||
class Encoder(nn.Module):
|
||||
|
||||
def __init__(self, nin, nout, ksize=3, stride=1, pad=1, activ=nn.LeakyReLU):
|
||||
super(Encoder, self).__init__()
|
||||
self.conv1 = Conv2DBNActiv(nin, nout, ksize, 1, pad, activ=activ)
|
||||
self.conv2 = Conv2DBNActiv(nout, nout, ksize, stride, pad, activ=activ)
|
||||
|
||||
def __call__(self, x):
|
||||
skip = self.conv1(x)
|
||||
h = self.conv2(skip)
|
||||
|
||||
return h, skip
|
||||
|
||||
|
||||
class Decoder(nn.Module):
|
||||
|
||||
def __init__(self, nin, nout, ksize=3, stride=1, pad=1, activ=nn.ReLU, dropout=False):
|
||||
super(Decoder, self).__init__()
|
||||
self.conv = Conv2DBNActiv(nin, nout, ksize, 1, pad, activ=activ)
|
||||
self.dropout = nn.Dropout2d(0.1) if dropout else None
|
||||
|
||||
def __call__(self, x, skip=None):
|
||||
x = F.interpolate(x, scale_factor=2, mode='bilinear', align_corners=True)
|
||||
if skip is not None:
|
||||
skip = spec_utils.crop_center(skip, x)
|
||||
x = torch.cat([x, skip], dim=1)
|
||||
h = self.conv(x)
|
||||
|
||||
if self.dropout is not None:
|
||||
h = self.dropout(h)
|
||||
|
||||
return h
|
||||
|
||||
|
||||
class ASPPModule(nn.Module):
|
||||
|
||||
def __init__(self, nin, nout, dilations=(4, 8, 16), activ=nn.ReLU):
|
||||
super(ASPPModule, self).__init__()
|
||||
self.conv1 = nn.Sequential(
|
||||
nn.AdaptiveAvgPool2d((1, None)),
|
||||
Conv2DBNActiv(nin, nin, 1, 1, 0, activ=activ)
|
||||
)
|
||||
self.conv2 = Conv2DBNActiv(nin, nin, 1, 1, 0, activ=activ)
|
||||
self.conv3 = SeperableConv2DBNActiv(
|
||||
nin, nin, 3, 1, dilations[0], dilations[0], activ=activ)
|
||||
self.conv4 = SeperableConv2DBNActiv(
|
||||
nin, nin, 3, 1, dilations[1], dilations[1], activ=activ)
|
||||
self.conv5 = SeperableConv2DBNActiv(
|
||||
nin, nin, 3, 1, dilations[2], dilations[2], activ=activ)
|
||||
self.bottleneck = nn.Sequential(
|
||||
Conv2DBNActiv(nin * 5, nout, 1, 1, 0, activ=activ),
|
||||
nn.Dropout2d(0.1)
|
||||
)
|
||||
|
||||
def forward(self, x):
|
||||
_, _, h, w = x.size()
|
||||
feat1 = F.interpolate(self.conv1(x), size=(h, w), mode='bilinear', align_corners=True)
|
||||
feat2 = self.conv2(x)
|
||||
feat3 = self.conv3(x)
|
||||
feat4 = self.conv4(x)
|
||||
feat5 = self.conv5(x)
|
||||
out = torch.cat((feat1, feat2, feat3, feat4, feat5), dim=1)
|
||||
bottle = self.bottleneck(out)
|
||||
return bottle
|
||||
@ -1,116 +0,0 @@
|
||||
import torch
|
||||
from torch import nn
|
||||
import torch.nn.functional as F
|
||||
|
||||
from lib_v5 import spec_utils
|
||||
|
||||
|
||||
class Conv2DBNActiv(nn.Module):
|
||||
|
||||
def __init__(self, nin, nout, ksize=3, stride=1, pad=1, dilation=1, activ=nn.ReLU):
|
||||
super(Conv2DBNActiv, self).__init__()
|
||||
self.conv = nn.Sequential(
|
||||
nn.Conv2d(
|
||||
nin, nout,
|
||||
kernel_size=ksize,
|
||||
stride=stride,
|
||||
padding=pad,
|
||||
dilation=dilation,
|
||||
bias=False),
|
||||
nn.BatchNorm2d(nout),
|
||||
activ()
|
||||
)
|
||||
|
||||
def __call__(self, x):
|
||||
return self.conv(x)
|
||||
|
||||
|
||||
class SeperableConv2DBNActiv(nn.Module):
|
||||
|
||||
def __init__(self, nin, nout, ksize=3, stride=1, pad=1, dilation=1, activ=nn.ReLU):
|
||||
super(SeperableConv2DBNActiv, self).__init__()
|
||||
self.conv = nn.Sequential(
|
||||
nn.Conv2d(
|
||||
nin, nin,
|
||||
kernel_size=ksize,
|
||||
stride=stride,
|
||||
padding=pad,
|
||||
dilation=dilation,
|
||||
groups=nin,
|
||||
bias=False),
|
||||
nn.Conv2d(
|
||||
nin, nout,
|
||||
kernel_size=1,
|
||||
bias=False),
|
||||
nn.BatchNorm2d(nout),
|
||||
activ()
|
||||
)
|
||||
|
||||
def __call__(self, x):
|
||||
return self.conv(x)
|
||||
|
||||
|
||||
class Encoder(nn.Module):
|
||||
|
||||
def __init__(self, nin, nout, ksize=3, stride=1, pad=1, activ=nn.LeakyReLU):
|
||||
super(Encoder, self).__init__()
|
||||
self.conv1 = Conv2DBNActiv(nin, nout, ksize, 1, pad, activ=activ)
|
||||
self.conv2 = Conv2DBNActiv(nout, nout, ksize, stride, pad, activ=activ)
|
||||
|
||||
def __call__(self, x):
|
||||
skip = self.conv1(x)
|
||||
h = self.conv2(skip)
|
||||
|
||||
return h, skip
|
||||
|
||||
|
||||
class Decoder(nn.Module):
|
||||
|
||||
def __init__(self, nin, nout, ksize=3, stride=1, pad=1, activ=nn.ReLU, dropout=False):
|
||||
super(Decoder, self).__init__()
|
||||
self.conv = Conv2DBNActiv(nin, nout, ksize, 1, pad, activ=activ)
|
||||
self.dropout = nn.Dropout2d(0.1) if dropout else None
|
||||
|
||||
def __call__(self, x, skip=None):
|
||||
x = F.interpolate(x, scale_factor=2, mode='bilinear', align_corners=True)
|
||||
if skip is not None:
|
||||
skip = spec_utils.crop_center(skip, x)
|
||||
x = torch.cat([x, skip], dim=1)
|
||||
h = self.conv(x)
|
||||
|
||||
if self.dropout is not None:
|
||||
h = self.dropout(h)
|
||||
|
||||
return h
|
||||
|
||||
|
||||
class ASPPModule(nn.Module):
|
||||
|
||||
def __init__(self, nin, nout, dilations=(4, 8, 16), activ=nn.ReLU):
|
||||
super(ASPPModule, self).__init__()
|
||||
self.conv1 = nn.Sequential(
|
||||
nn.AdaptiveAvgPool2d((1, None)),
|
||||
Conv2DBNActiv(nin, nin, 1, 1, 0, activ=activ)
|
||||
)
|
||||
self.conv2 = Conv2DBNActiv(nin, nin, 1, 1, 0, activ=activ)
|
||||
self.conv3 = SeperableConv2DBNActiv(
|
||||
nin, nin, 3, 1, dilations[0], dilations[0], activ=activ)
|
||||
self.conv4 = SeperableConv2DBNActiv(
|
||||
nin, nin, 3, 1, dilations[1], dilations[1], activ=activ)
|
||||
self.conv5 = SeperableConv2DBNActiv(
|
||||
nin, nin, 3, 1, dilations[2], dilations[2], activ=activ)
|
||||
self.bottleneck = nn.Sequential(
|
||||
Conv2DBNActiv(nin * 5, nout, 1, 1, 0, activ=activ),
|
||||
nn.Dropout2d(0.1)
|
||||
)
|
||||
|
||||
def forward(self, x):
|
||||
_, _, h, w = x.size()
|
||||
feat1 = F.interpolate(self.conv1(x), size=(h, w), mode='bilinear', align_corners=True)
|
||||
feat2 = self.conv2(x)
|
||||
feat3 = self.conv3(x)
|
||||
feat4 = self.conv4(x)
|
||||
feat5 = self.conv5(x)
|
||||
out = torch.cat((feat1, feat2, feat3, feat4, feat5), dim=1)
|
||||
bottle = self.bottleneck(out)
|
||||
return bottle
|
||||
@ -1,119 +0,0 @@
|
||||
import torch
|
||||
from torch import nn
|
||||
import torch.nn.functional as F
|
||||
|
||||
from lib_v5 import spec_utils
|
||||
|
||||
|
||||
class Conv2DBNActiv(nn.Module):
|
||||
|
||||
def __init__(self, nin, nout, ksize=3, stride=1, pad=1, dilation=1, activ=nn.ReLU):
|
||||
super(Conv2DBNActiv, self).__init__()
|
||||
self.conv = nn.Sequential(
|
||||
nn.Conv2d(
|
||||
nin, nout,
|
||||
kernel_size=ksize,
|
||||
stride=stride,
|
||||
padding=pad,
|
||||
dilation=dilation,
|
||||
bias=False),
|
||||
nn.BatchNorm2d(nout),
|
||||
activ()
|
||||
)
|
||||
|
||||
def __call__(self, x):
|
||||
return self.conv(x)
|
||||
|
||||
|
||||
class SeperableConv2DBNActiv(nn.Module):
|
||||
|
||||
def __init__(self, nin, nout, ksize=3, stride=1, pad=1, dilation=1, activ=nn.ReLU):
|
||||
super(SeperableConv2DBNActiv, self).__init__()
|
||||
self.conv = nn.Sequential(
|
||||
nn.Conv2d(
|
||||
nin, nin,
|
||||
kernel_size=ksize,
|
||||
stride=stride,
|
||||
padding=pad,
|
||||
dilation=dilation,
|
||||
groups=nin,
|
||||
bias=False),
|
||||
nn.Conv2d(
|
||||
nin, nout,
|
||||
kernel_size=1,
|
||||
bias=False),
|
||||
nn.BatchNorm2d(nout),
|
||||
activ()
|
||||
)
|
||||
|
||||
def __call__(self, x):
|
||||
return self.conv(x)
|
||||
|
||||
|
||||
class Encoder(nn.Module):
|
||||
|
||||
def __init__(self, nin, nout, ksize=3, stride=1, pad=1, activ=nn.LeakyReLU):
|
||||
super(Encoder, self).__init__()
|
||||
self.conv1 = Conv2DBNActiv(nin, nout, ksize, 1, pad, activ=activ)
|
||||
self.conv2 = Conv2DBNActiv(nout, nout, ksize, stride, pad, activ=activ)
|
||||
|
||||
def __call__(self, x):
|
||||
skip = self.conv1(x)
|
||||
h = self.conv2(skip)
|
||||
|
||||
return h, skip
|
||||
|
||||
|
||||
class Decoder(nn.Module):
|
||||
|
||||
def __init__(self, nin, nout, ksize=3, stride=1, pad=1, activ=nn.ReLU, dropout=False):
|
||||
super(Decoder, self).__init__()
|
||||
self.conv = Conv2DBNActiv(nin, nout, ksize, 1, pad, activ=activ)
|
||||
self.dropout = nn.Dropout2d(0.1) if dropout else None
|
||||
|
||||
def __call__(self, x, skip=None):
|
||||
x = F.interpolate(x, scale_factor=2, mode='bilinear', align_corners=True)
|
||||
if skip is not None:
|
||||
skip = spec_utils.crop_center(skip, x)
|
||||
x = torch.cat([x, skip], dim=1)
|
||||
h = self.conv(x)
|
||||
|
||||
if self.dropout is not None:
|
||||
h = self.dropout(h)
|
||||
|
||||
return h
|
||||
|
||||
|
||||
class ASPPModule(nn.Module):
|
||||
|
||||
def __init__(self, nin, nout, dilations=(4, 8, 16, 32), activ=nn.ReLU):
|
||||
super(ASPPModule, self).__init__()
|
||||
self.conv1 = nn.Sequential(
|
||||
nn.AdaptiveAvgPool2d((1, None)),
|
||||
Conv2DBNActiv(nin, nin, 1, 1, 0, activ=activ)
|
||||
)
|
||||
self.conv2 = Conv2DBNActiv(nin, nin, 1, 1, 0, activ=activ)
|
||||
self.conv3 = SeperableConv2DBNActiv(
|
||||
nin, nin, 3, 1, dilations[0], dilations[0], activ=activ)
|
||||
self.conv4 = SeperableConv2DBNActiv(
|
||||
nin, nin, 3, 1, dilations[1], dilations[1], activ=activ)
|
||||
self.conv5 = SeperableConv2DBNActiv(
|
||||
nin, nin, 3, 1, dilations[2], dilations[2], activ=activ)
|
||||
self.conv6 = SeperableConv2DBNActiv(
|
||||
nin, nin, 3, 1, dilations[2], dilations[2], activ=activ)
|
||||
self.bottleneck = nn.Sequential(
|
||||
Conv2DBNActiv(nin * 6, nout, 1, 1, 0, activ=activ),
|
||||
nn.Dropout2d(0.1)
|
||||
)
|
||||
|
||||
def forward(self, x):
|
||||
_, _, h, w = x.size()
|
||||
feat1 = F.interpolate(self.conv1(x), size=(h, w), mode='bilinear', align_corners=True)
|
||||
feat2 = self.conv2(x)
|
||||
feat3 = self.conv3(x)
|
||||
feat4 = self.conv4(x)
|
||||
feat5 = self.conv5(x)
|
||||
feat6 = self.conv6(x)
|
||||
out = torch.cat((feat1, feat2, feat3, feat4, feat5, feat6), dim=1)
|
||||
bottle = self.bottleneck(out)
|
||||
return bottle
|
||||
@ -1,122 +0,0 @@
|
||||
import torch
|
||||
from torch import nn
|
||||
import torch.nn.functional as F
|
||||
|
||||
from lib_v5 import spec_utils
|
||||
|
||||
|
||||
class Conv2DBNActiv(nn.Module):
|
||||
|
||||
def __init__(self, nin, nout, ksize=3, stride=1, pad=1, dilation=1, activ=nn.ReLU):
|
||||
super(Conv2DBNActiv, self).__init__()
|
||||
self.conv = nn.Sequential(
|
||||
nn.Conv2d(
|
||||
nin, nout,
|
||||
kernel_size=ksize,
|
||||
stride=stride,
|
||||
padding=pad,
|
||||
dilation=dilation,
|
||||
bias=False),
|
||||
nn.BatchNorm2d(nout),
|
||||
activ()
|
||||
)
|
||||
|
||||
def __call__(self, x):
|
||||
return self.conv(x)
|
||||
|
||||
|
||||
class SeperableConv2DBNActiv(nn.Module):
|
||||
|
||||
def __init__(self, nin, nout, ksize=3, stride=1, pad=1, dilation=1, activ=nn.ReLU):
|
||||
super(SeperableConv2DBNActiv, self).__init__()
|
||||
self.conv = nn.Sequential(
|
||||
nn.Conv2d(
|
||||
nin, nin,
|
||||
kernel_size=ksize,
|
||||
stride=stride,
|
||||
padding=pad,
|
||||
dilation=dilation,
|
||||
groups=nin,
|
||||
bias=False),
|
||||
nn.Conv2d(
|
||||
nin, nout,
|
||||
kernel_size=1,
|
||||
bias=False),
|
||||
nn.BatchNorm2d(nout),
|
||||
activ()
|
||||
)
|
||||
|
||||
def __call__(self, x):
|
||||
return self.conv(x)
|
||||
|
||||
|
||||
class Encoder(nn.Module):
|
||||
|
||||
def __init__(self, nin, nout, ksize=3, stride=1, pad=1, activ=nn.LeakyReLU):
|
||||
super(Encoder, self).__init__()
|
||||
self.conv1 = Conv2DBNActiv(nin, nout, ksize, 1, pad, activ=activ)
|
||||
self.conv2 = Conv2DBNActiv(nout, nout, ksize, stride, pad, activ=activ)
|
||||
|
||||
def __call__(self, x):
|
||||
skip = self.conv1(x)
|
||||
h = self.conv2(skip)
|
||||
|
||||
return h, skip
|
||||
|
||||
|
||||
class Decoder(nn.Module):
|
||||
|
||||
def __init__(self, nin, nout, ksize=3, stride=1, pad=1, activ=nn.ReLU, dropout=False):
|
||||
super(Decoder, self).__init__()
|
||||
self.conv = Conv2DBNActiv(nin, nout, ksize, 1, pad, activ=activ)
|
||||
self.dropout = nn.Dropout2d(0.1) if dropout else None
|
||||
|
||||
def __call__(self, x, skip=None):
|
||||
x = F.interpolate(x, scale_factor=2, mode='bilinear', align_corners=True)
|
||||
if skip is not None:
|
||||
skip = spec_utils.crop_center(skip, x)
|
||||
x = torch.cat([x, skip], dim=1)
|
||||
h = self.conv(x)
|
||||
|
||||
if self.dropout is not None:
|
||||
h = self.dropout(h)
|
||||
|
||||
return h
|
||||
|
||||
|
||||
class ASPPModule(nn.Module):
|
||||
|
||||
def __init__(self, nin, nout, dilations=(4, 8, 16, 32, 64), activ=nn.ReLU):
|
||||
super(ASPPModule, self).__init__()
|
||||
self.conv1 = nn.Sequential(
|
||||
nn.AdaptiveAvgPool2d((1, None)),
|
||||
Conv2DBNActiv(nin, nin, 1, 1, 0, activ=activ)
|
||||
)
|
||||
self.conv2 = Conv2DBNActiv(nin, nin, 1, 1, 0, activ=activ)
|
||||
self.conv3 = SeperableConv2DBNActiv(
|
||||
nin, nin, 3, 1, dilations[0], dilations[0], activ=activ)
|
||||
self.conv4 = SeperableConv2DBNActiv(
|
||||
nin, nin, 3, 1, dilations[1], dilations[1], activ=activ)
|
||||
self.conv5 = SeperableConv2DBNActiv(
|
||||
nin, nin, 3, 1, dilations[2], dilations[2], activ=activ)
|
||||
self.conv6 = SeperableConv2DBNActiv(
|
||||
nin, nin, 3, 1, dilations[2], dilations[2], activ=activ)
|
||||
self.conv7 = SeperableConv2DBNActiv(
|
||||
nin, nin, 3, 1, dilations[2], dilations[2], activ=activ)
|
||||
self.bottleneck = nn.Sequential(
|
||||
Conv2DBNActiv(nin * 7, nout, 1, 1, 0, activ=activ),
|
||||
nn.Dropout2d(0.1)
|
||||
)
|
||||
|
||||
def forward(self, x):
|
||||
_, _, h, w = x.size()
|
||||
feat1 = F.interpolate(self.conv1(x), size=(h, w), mode='bilinear', align_corners=True)
|
||||
feat2 = self.conv2(x)
|
||||
feat3 = self.conv3(x)
|
||||
feat4 = self.conv4(x)
|
||||
feat5 = self.conv5(x)
|
||||
feat6 = self.conv6(x)
|
||||
feat7 = self.conv7(x)
|
||||
out = torch.cat((feat1, feat2, feat3, feat4, feat5, feat6, feat7), dim=1)
|
||||
bottle = self.bottleneck(out)
|
||||
return bottle
|
||||
@ -1,122 +0,0 @@
|
||||
import torch
|
||||
from torch import nn
|
||||
import torch.nn.functional as F
|
||||
|
||||
from lib_v5 import spec_utils
|
||||
|
||||
|
||||
class Conv2DBNActiv(nn.Module):
|
||||
|
||||
def __init__(self, nin, nout, ksize=3, stride=1, pad=1, dilation=1, activ=nn.ReLU):
|
||||
super(Conv2DBNActiv, self).__init__()
|
||||
self.conv = nn.Sequential(
|
||||
nn.Conv2d(
|
||||
nin, nout,
|
||||
kernel_size=ksize,
|
||||
stride=stride,
|
||||
padding=pad,
|
||||
dilation=dilation,
|
||||
bias=False),
|
||||
nn.BatchNorm2d(nout),
|
||||
activ()
|
||||
)
|
||||
|
||||
def __call__(self, x):
|
||||
return self.conv(x)
|
||||
|
||||
|
||||
class SeperableConv2DBNActiv(nn.Module):
|
||||
|
||||
def __init__(self, nin, nout, ksize=3, stride=1, pad=1, dilation=1, activ=nn.ReLU):
|
||||
super(SeperableConv2DBNActiv, self).__init__()
|
||||
self.conv = nn.Sequential(
|
||||
nn.Conv2d(
|
||||
nin, nin,
|
||||
kernel_size=ksize,
|
||||
stride=stride,
|
||||
padding=pad,
|
||||
dilation=dilation,
|
||||
groups=nin,
|
||||
bias=False),
|
||||
nn.Conv2d(
|
||||
nin, nout,
|
||||
kernel_size=1,
|
||||
bias=False),
|
||||
nn.BatchNorm2d(nout),
|
||||
activ()
|
||||
)
|
||||
|
||||
def __call__(self, x):
|
||||
return self.conv(x)
|
||||
|
||||
|
||||
class Encoder(nn.Module):
|
||||
|
||||
def __init__(self, nin, nout, ksize=3, stride=1, pad=1, activ=nn.LeakyReLU):
|
||||
super(Encoder, self).__init__()
|
||||
self.conv1 = Conv2DBNActiv(nin, nout, ksize, 1, pad, activ=activ)
|
||||
self.conv2 = Conv2DBNActiv(nout, nout, ksize, stride, pad, activ=activ)
|
||||
|
||||
def __call__(self, x):
|
||||
skip = self.conv1(x)
|
||||
h = self.conv2(skip)
|
||||
|
||||
return h, skip
|
||||
|
||||
|
||||
class Decoder(nn.Module):
|
||||
|
||||
def __init__(self, nin, nout, ksize=3, stride=1, pad=1, activ=nn.ReLU, dropout=False):
|
||||
super(Decoder, self).__init__()
|
||||
self.conv = Conv2DBNActiv(nin, nout, ksize, 1, pad, activ=activ)
|
||||
self.dropout = nn.Dropout2d(0.1) if dropout else None
|
||||
|
||||
def __call__(self, x, skip=None):
|
||||
x = F.interpolate(x, scale_factor=2, mode='bilinear', align_corners=True)
|
||||
if skip is not None:
|
||||
skip = spec_utils.crop_center(skip, x)
|
||||
x = torch.cat([x, skip], dim=1)
|
||||
h = self.conv(x)
|
||||
|
||||
if self.dropout is not None:
|
||||
h = self.dropout(h)
|
||||
|
||||
return h
|
||||
|
||||
|
||||
class ASPPModule(nn.Module):
|
||||
|
||||
def __init__(self, nin, nout, dilations=(4, 8, 16, 32, 64), activ=nn.ReLU):
|
||||
super(ASPPModule, self).__init__()
|
||||
self.conv1 = nn.Sequential(
|
||||
nn.AdaptiveAvgPool2d((1, None)),
|
||||
Conv2DBNActiv(nin, nin, 1, 1, 0, activ=activ)
|
||||
)
|
||||
self.conv2 = Conv2DBNActiv(nin, nin, 1, 1, 0, activ=activ)
|
||||
self.conv3 = SeperableConv2DBNActiv(
|
||||
nin, nin, 3, 1, dilations[0], dilations[0], activ=activ)
|
||||
self.conv4 = SeperableConv2DBNActiv(
|
||||
nin, nin, 3, 1, dilations[1], dilations[1], activ=activ)
|
||||
self.conv5 = SeperableConv2DBNActiv(
|
||||
nin, nin, 3, 1, dilations[2], dilations[2], activ=activ)
|
||||
self.conv6 = SeperableConv2DBNActiv(
|
||||
nin, nin, 3, 1, dilations[2], dilations[2], activ=activ)
|
||||
self.conv7 = SeperableConv2DBNActiv(
|
||||
nin, nin, 3, 1, dilations[2], dilations[2], activ=activ)
|
||||
self.bottleneck = nn.Sequential(
|
||||
Conv2DBNActiv(nin * 7, nout, 1, 1, 0, activ=activ),
|
||||
nn.Dropout2d(0.1)
|
||||
)
|
||||
|
||||
def forward(self, x):
|
||||
_, _, h, w = x.size()
|
||||
feat1 = F.interpolate(self.conv1(x), size=(h, w), mode='bilinear', align_corners=True)
|
||||
feat2 = self.conv2(x)
|
||||
feat3 = self.conv3(x)
|
||||
feat4 = self.conv4(x)
|
||||
feat5 = self.conv5(x)
|
||||
feat6 = self.conv6(x)
|
||||
feat7 = self.conv7(x)
|
||||
out = torch.cat((feat1, feat2, feat3, feat4, feat5, feat6, feat7), dim=1)
|
||||
bottle = self.bottleneck(out)
|
||||
return bottle
|
||||
@ -1,122 +0,0 @@
|
||||
import torch
|
||||
from torch import nn
|
||||
import torch.nn.functional as F
|
||||
|
||||
from lib_v5 import spec_utils
|
||||
|
||||
|
||||
class Conv2DBNActiv(nn.Module):
|
||||
|
||||
def __init__(self, nin, nout, ksize=3, stride=1, pad=1, dilation=1, activ=nn.ReLU):
|
||||
super(Conv2DBNActiv, self).__init__()
|
||||
self.conv = nn.Sequential(
|
||||
nn.Conv2d(
|
||||
nin, nout,
|
||||
kernel_size=ksize,
|
||||
stride=stride,
|
||||
padding=pad,
|
||||
dilation=dilation,
|
||||
bias=False),
|
||||
nn.BatchNorm2d(nout),
|
||||
activ()
|
||||
)
|
||||
|
||||
def __call__(self, x):
|
||||
return self.conv(x)
|
||||
|
||||
|
||||
class SeperableConv2DBNActiv(nn.Module):
|
||||
|
||||
def __init__(self, nin, nout, ksize=3, stride=1, pad=1, dilation=1, activ=nn.ReLU):
|
||||
super(SeperableConv2DBNActiv, self).__init__()
|
||||
self.conv = nn.Sequential(
|
||||
nn.Conv2d(
|
||||
nin, nin,
|
||||
kernel_size=ksize,
|
||||
stride=stride,
|
||||
padding=pad,
|
||||
dilation=dilation,
|
||||
groups=nin,
|
||||
bias=False),
|
||||
nn.Conv2d(
|
||||
nin, nout,
|
||||
kernel_size=1,
|
||||
bias=False),
|
||||
nn.BatchNorm2d(nout),
|
||||
activ()
|
||||
)
|
||||
|
||||
def __call__(self, x):
|
||||
return self.conv(x)
|
||||
|
||||
|
||||
class Encoder(nn.Module):
|
||||
|
||||
def __init__(self, nin, nout, ksize=3, stride=1, pad=1, activ=nn.LeakyReLU):
|
||||
super(Encoder, self).__init__()
|
||||
self.conv1 = Conv2DBNActiv(nin, nout, ksize, 1, pad, activ=activ)
|
||||
self.conv2 = Conv2DBNActiv(nout, nout, ksize, stride, pad, activ=activ)
|
||||
|
||||
def __call__(self, x):
|
||||
skip = self.conv1(x)
|
||||
h = self.conv2(skip)
|
||||
|
||||
return h, skip
|
||||
|
||||
|
||||
class Decoder(nn.Module):
|
||||
|
||||
def __init__(self, nin, nout, ksize=3, stride=1, pad=1, activ=nn.ReLU, dropout=False):
|
||||
super(Decoder, self).__init__()
|
||||
self.conv = Conv2DBNActiv(nin, nout, ksize, 1, pad, activ=activ)
|
||||
self.dropout = nn.Dropout2d(0.1) if dropout else None
|
||||
|
||||
def __call__(self, x, skip=None):
|
||||
x = F.interpolate(x, scale_factor=2, mode='bilinear', align_corners=True)
|
||||
if skip is not None:
|
||||
skip = spec_utils.crop_center(skip, x)
|
||||
x = torch.cat([x, skip], dim=1)
|
||||
h = self.conv(x)
|
||||
|
||||
if self.dropout is not None:
|
||||
h = self.dropout(h)
|
||||
|
||||
return h
|
||||
|
||||
|
||||
class ASPPModule(nn.Module):
|
||||
|
||||
def __init__(self, nin, nout, dilations=(4, 8, 16, 32, 64), activ=nn.ReLU):
|
||||
super(ASPPModule, self).__init__()
|
||||
self.conv1 = nn.Sequential(
|
||||
nn.AdaptiveAvgPool2d((1, None)),
|
||||
Conv2DBNActiv(nin, nin, 1, 1, 0, activ=activ)
|
||||
)
|
||||
self.conv2 = Conv2DBNActiv(nin, nin, 1, 1, 0, activ=activ)
|
||||
self.conv3 = SeperableConv2DBNActiv(
|
||||
nin, nin, 3, 1, dilations[0], dilations[0], activ=activ)
|
||||
self.conv4 = SeperableConv2DBNActiv(
|
||||
nin, nin, 3, 1, dilations[1], dilations[1], activ=activ)
|
||||
self.conv5 = SeperableConv2DBNActiv(
|
||||
nin, nin, 3, 1, dilations[2], dilations[2], activ=activ)
|
||||
self.conv6 = SeperableConv2DBNActiv(
|
||||
nin, nin, 3, 1, dilations[2], dilations[2], activ=activ)
|
||||
self.conv7 = SeperableConv2DBNActiv(
|
||||
nin, nin, 3, 1, dilations[2], dilations[2], activ=activ)
|
||||
self.bottleneck = nn.Sequential(
|
||||
Conv2DBNActiv(nin * 7, nout, 1, 1, 0, activ=activ),
|
||||
nn.Dropout2d(0.1)
|
||||
)
|
||||
|
||||
def forward(self, x):
|
||||
_, _, h, w = x.size()
|
||||
feat1 = F.interpolate(self.conv1(x), size=(h, w), mode='bilinear', align_corners=True)
|
||||
feat2 = self.conv2(x)
|
||||
feat3 = self.conv3(x)
|
||||
feat4 = self.conv4(x)
|
||||
feat5 = self.conv5(x)
|
||||
feat6 = self.conv6(x)
|
||||
feat7 = self.conv7(x)
|
||||
out = torch.cat((feat1, feat2, feat3, feat4, feat5, feat6, feat7), dim=1)
|
||||
bottle = self.bottleneck(out)
|
||||
return bottle
|
||||
@ -1,60 +0,0 @@
|
||||
import json
|
||||
import os
|
||||
import pathlib
|
||||
|
||||
default_param = {}
|
||||
default_param['bins'] = 768
|
||||
default_param['unstable_bins'] = 9 # training only
|
||||
default_param['reduction_bins'] = 762 # training only
|
||||
default_param['sr'] = 44100
|
||||
default_param['pre_filter_start'] = 757
|
||||
default_param['pre_filter_stop'] = 768
|
||||
default_param['band'] = {}
|
||||
|
||||
|
||||
default_param['band'][1] = {
|
||||
'sr': 11025,
|
||||
'hl': 128,
|
||||
'n_fft': 960,
|
||||
'crop_start': 0,
|
||||
'crop_stop': 245,
|
||||
'lpf_start': 61, # inference only
|
||||
'res_type': 'polyphase'
|
||||
}
|
||||
|
||||
default_param['band'][2] = {
|
||||
'sr': 44100,
|
||||
'hl': 512,
|
||||
'n_fft': 1536,
|
||||
'crop_start': 24,
|
||||
'crop_stop': 547,
|
||||
'hpf_start': 81, # inference only
|
||||
'res_type': 'sinc_best'
|
||||
}
|
||||
|
||||
|
||||
def int_keys(d):
|
||||
r = {}
|
||||
for k, v in d:
|
||||
if k.isdigit():
|
||||
k = int(k)
|
||||
r[k] = v
|
||||
return r
|
||||
|
||||
|
||||
class ModelParameters(object):
|
||||
def __init__(self, config_path=''):
|
||||
if '.pth' == pathlib.Path(config_path).suffix:
|
||||
import zipfile
|
||||
|
||||
with zipfile.ZipFile(config_path, 'r') as zip:
|
||||
self.param = json.loads(zip.read('param.json'), object_pairs_hook=int_keys)
|
||||
elif '.json' == pathlib.Path(config_path).suffix:
|
||||
with open(config_path, 'r') as f:
|
||||
self.param = json.loads(f.read(), object_pairs_hook=int_keys)
|
||||
else:
|
||||
self.param = default_param
|
||||
|
||||
for k in ['mid_side', 'mid_side_b', 'mid_side_b2', 'stereo_w', 'stereo_n', 'reverse']:
|
||||
if not k in self.param:
|
||||
self.param[k] = False
|
||||
@ -1,19 +0,0 @@
|
||||
{
|
||||
"bins": 1024,
|
||||
"unstable_bins": 0,
|
||||
"reduction_bins": 0,
|
||||
"band": {
|
||||
"1": {
|
||||
"sr": 16000,
|
||||
"hl": 512,
|
||||
"n_fft": 2048,
|
||||
"crop_start": 0,
|
||||
"crop_stop": 1024,
|
||||
"hpf_start": -1,
|
||||
"res_type": "sinc_best"
|
||||
}
|
||||
},
|
||||
"sr": 16000,
|
||||
"pre_filter_start": 1023,
|
||||
"pre_filter_stop": 1024
|
||||
}
|
||||
@ -1,19 +0,0 @@
|
||||
{
|
||||
"bins": 1024,
|
||||
"unstable_bins": 0,
|
||||
"reduction_bins": 0,
|
||||
"band": {
|
||||
"1": {
|
||||
"sr": 32000,
|
||||
"hl": 512,
|
||||
"n_fft": 2048,
|
||||
"crop_start": 0,
|
||||
"crop_stop": 1024,
|
||||
"hpf_start": -1,
|
||||
"res_type": "kaiser_fast"
|
||||
}
|
||||
},
|
||||
"sr": 32000,
|
||||
"pre_filter_start": 1000,
|
||||
"pre_filter_stop": 1021
|
||||
}
|
||||
@ -1,19 +0,0 @@
|
||||
{
|
||||
"bins": 1024,
|
||||
"unstable_bins": 0,
|
||||
"reduction_bins": 0,
|
||||
"band": {
|
||||
"1": {
|
||||
"sr": 33075,
|
||||
"hl": 384,
|
||||
"n_fft": 2048,
|
||||
"crop_start": 0,
|
||||
"crop_stop": 1024,
|
||||
"hpf_start": -1,
|
||||
"res_type": "sinc_best"
|
||||
}
|
||||
},
|
||||
"sr": 33075,
|
||||
"pre_filter_start": 1000,
|
||||
"pre_filter_stop": 1021
|
||||
}
|
||||
@ -1,19 +0,0 @@
|
||||
{
|
||||
"bins": 1024,
|
||||
"unstable_bins": 0,
|
||||
"reduction_bins": 0,
|
||||
"band": {
|
||||
"1": {
|
||||
"sr": 44100,
|
||||
"hl": 1024,
|
||||
"n_fft": 2048,
|
||||
"crop_start": 0,
|
||||
"crop_stop": 1024,
|
||||
"hpf_start": -1,
|
||||
"res_type": "sinc_best"
|
||||
}
|
||||
},
|
||||
"sr": 44100,
|
||||
"pre_filter_start": 1023,
|
||||
"pre_filter_stop": 1024
|
||||
}
|
||||
@ -1,19 +0,0 @@
|
||||
{
|
||||
"bins": 256,
|
||||
"unstable_bins": 0,
|
||||
"reduction_bins": 0,
|
||||
"band": {
|
||||
"1": {
|
||||
"sr": 44100,
|
||||
"hl": 256,
|
||||
"n_fft": 512,
|
||||
"crop_start": 0,
|
||||
"crop_stop": 256,
|
||||
"hpf_start": -1,
|
||||
"res_type": "sinc_best"
|
||||
}
|
||||
},
|
||||
"sr": 44100,
|
||||
"pre_filter_start": 256,
|
||||
"pre_filter_stop": 256
|
||||
}
|
||||
@ -1,19 +0,0 @@
|
||||
{
|
||||
"bins": 1024,
|
||||
"unstable_bins": 0,
|
||||
"reduction_bins": 0,
|
||||
"band": {
|
||||
"1": {
|
||||
"sr": 44100,
|
||||
"hl": 512,
|
||||
"n_fft": 2048,
|
||||
"crop_start": 0,
|
||||
"crop_stop": 1024,
|
||||
"hpf_start": -1,
|
||||
"res_type": "sinc_best"
|
||||
}
|
||||
},
|
||||
"sr": 44100,
|
||||
"pre_filter_start": 1023,
|
||||
"pre_filter_stop": 1024
|
||||
}
|
||||
@ -1,19 +0,0 @@
|
||||
{
|
||||
"bins": 1024,
|
||||
"unstable_bins": 0,
|
||||
"reduction_bins": 0,
|
||||
"band": {
|
||||
"1": {
|
||||
"sr": 44100,
|
||||
"hl": 512,
|
||||
"n_fft": 2048,
|
||||
"crop_start": 0,
|
||||
"crop_stop": 700,
|
||||
"hpf_start": -1,
|
||||
"res_type": "sinc_best"
|
||||
}
|
||||
},
|
||||
"sr": 44100,
|
||||
"pre_filter_start": 1023,
|
||||
"pre_filter_stop": 700
|
||||
}
|
||||
@ -1,30 +0,0 @@
|
||||
{
|
||||
"bins": 768,
|
||||
"unstable_bins": 7,
|
||||
"reduction_bins": 705,
|
||||
"band": {
|
||||
"1": {
|
||||
"sr": 6000,
|
||||
"hl": 66,
|
||||
"n_fft": 512,
|
||||
"crop_start": 0,
|
||||
"crop_stop": 240,
|
||||
"lpf_start": 60,
|
||||
"lpf_stop": 118,
|
||||
"res_type": "sinc_fastest"
|
||||
},
|
||||
"2": {
|
||||
"sr": 32000,
|
||||
"hl": 352,
|
||||
"n_fft": 1024,
|
||||
"crop_start": 22,
|
||||
"crop_stop": 505,
|
||||
"hpf_start": 44,
|
||||
"hpf_stop": 23,
|
||||
"res_type": "sinc_medium"
|
||||
}
|
||||
},
|
||||
"sr": 32000,
|
||||
"pre_filter_start": 710,
|
||||
"pre_filter_stop": 731
|
||||
}
|
||||
@ -1,30 +0,0 @@
|
||||
{
|
||||
"bins": 512,
|
||||
"unstable_bins": 7,
|
||||
"reduction_bins": 510,
|
||||
"band": {
|
||||
"1": {
|
||||
"sr": 11025,
|
||||
"hl": 160,
|
||||
"n_fft": 768,
|
||||
"crop_start": 0,
|
||||
"crop_stop": 192,
|
||||
"lpf_start": 41,
|
||||
"lpf_stop": 139,
|
||||
"res_type": "sinc_fastest"
|
||||
},
|
||||
"2": {
|
||||
"sr": 44100,
|
||||
"hl": 640,
|
||||
"n_fft": 1024,
|
||||
"crop_start": 10,
|
||||
"crop_stop": 320,
|
||||
"hpf_start": 47,
|
||||
"hpf_stop": 15,
|
||||
"res_type": "sinc_medium"
|
||||
}
|
||||
},
|
||||
"sr": 44100,
|
||||
"pre_filter_start": 510,
|
||||
"pre_filter_stop": 512
|
||||
}
|
||||
@ -1,30 +0,0 @@
|
||||
{
|
||||
"bins": 768,
|
||||
"unstable_bins": 7,
|
||||
"reduction_bins": 705,
|
||||
"band": {
|
||||
"1": {
|
||||
"sr": 6000,
|
||||
"hl": 66,
|
||||
"n_fft": 512,
|
||||
"crop_start": 0,
|
||||
"crop_stop": 240,
|
||||
"lpf_start": 60,
|
||||
"lpf_stop": 240,
|
||||
"res_type": "sinc_fastest"
|
||||
},
|
||||
"2": {
|
||||
"sr": 48000,
|
||||
"hl": 528,
|
||||
"n_fft": 1536,
|
||||
"crop_start": 22,
|
||||
"crop_stop": 505,
|
||||
"hpf_start": 82,
|
||||
"hpf_stop": 22,
|
||||
"res_type": "sinc_medium"
|
||||
}
|
||||
},
|
||||
"sr": 48000,
|
||||
"pre_filter_start": 710,
|
||||
"pre_filter_stop": 731
|
||||
}
|
||||
@ -1,42 +0,0 @@
|
||||
{
|
||||
"bins": 768,
|
||||
"unstable_bins": 5,
|
||||
"reduction_bins": 733,
|
||||
"band": {
|
||||
"1": {
|
||||
"sr": 11025,
|
||||
"hl": 128,
|
||||
"n_fft": 768,
|
||||
"crop_start": 0,
|
||||
"crop_stop": 278,
|
||||
"lpf_start": 28,
|
||||
"lpf_stop": 140,
|
||||
"res_type": "polyphase"
|
||||
},
|
||||
"2": {
|
||||
"sr": 22050,
|
||||
"hl": 256,
|
||||
"n_fft": 768,
|
||||
"crop_start": 14,
|
||||
"crop_stop": 322,
|
||||
"hpf_start": 70,
|
||||
"hpf_stop": 14,
|
||||
"lpf_start": 283,
|
||||
"lpf_stop": 314,
|
||||
"res_type": "polyphase"
|
||||
},
|
||||
"3": {
|
||||
"sr": 44100,
|
||||
"hl": 512,
|
||||
"n_fft": 768,
|
||||
"crop_start": 131,
|
||||
"crop_stop": 313,
|
||||
"hpf_start": 154,
|
||||
"hpf_stop": 141,
|
||||
"res_type": "sinc_medium"
|
||||
}
|
||||
},
|
||||
"sr": 44100,
|
||||
"pre_filter_start": 757,
|
||||
"pre_filter_stop": 768
|
||||
}
|
||||
@ -1,43 +0,0 @@
|
||||
{
|
||||
"mid_side": true,
|
||||
"bins": 768,
|
||||
"unstable_bins": 5,
|
||||
"reduction_bins": 733,
|
||||
"band": {
|
||||
"1": {
|
||||
"sr": 11025,
|
||||
"hl": 128,
|
||||
"n_fft": 768,
|
||||
"crop_start": 0,
|
||||
"crop_stop": 278,
|
||||
"lpf_start": 28,
|
||||
"lpf_stop": 140,
|
||||
"res_type": "polyphase"
|
||||
},
|
||||
"2": {
|
||||
"sr": 22050,
|
||||
"hl": 256,
|
||||
"n_fft": 768,
|
||||
"crop_start": 14,
|
||||
"crop_stop": 322,
|
||||
"hpf_start": 70,
|
||||
"hpf_stop": 14,
|
||||
"lpf_start": 283,
|
||||
"lpf_stop": 314,
|
||||
"res_type": "polyphase"
|
||||
},
|
||||
"3": {
|
||||
"sr": 44100,
|
||||
"hl": 512,
|
||||
"n_fft": 768,
|
||||
"crop_start": 131,
|
||||
"crop_stop": 313,
|
||||
"hpf_start": 154,
|
||||
"hpf_stop": 141,
|
||||
"res_type": "sinc_medium"
|
||||
}
|
||||
},
|
||||
"sr": 44100,
|
||||
"pre_filter_start": 757,
|
||||
"pre_filter_stop": 768
|
||||
}
|
||||
@ -1,43 +0,0 @@
|
||||
{
|
||||
"mid_side_b2": true,
|
||||
"bins": 640,
|
||||
"unstable_bins": 7,
|
||||
"reduction_bins": 565,
|
||||
"band": {
|
||||
"1": {
|
||||
"sr": 11025,
|
||||
"hl": 108,
|
||||
"n_fft": 1024,
|
||||
"crop_start": 0,
|
||||
"crop_stop": 187,
|
||||
"lpf_start": 92,
|
||||
"lpf_stop": 186,
|
||||
"res_type": "polyphase"
|
||||
},
|
||||
"2": {
|
||||
"sr": 22050,
|
||||
"hl": 216,
|
||||
"n_fft": 768,
|
||||
"crop_start": 0,
|
||||
"crop_stop": 212,
|
||||
"hpf_start": 68,
|
||||
"hpf_stop": 34,
|
||||
"lpf_start": 174,
|
||||
"lpf_stop": 209,
|
||||
"res_type": "polyphase"
|
||||
},
|
||||
"3": {
|
||||
"sr": 44100,
|
||||
"hl": 432,
|
||||
"n_fft": 640,
|
||||
"crop_start": 66,
|
||||
"crop_stop": 307,
|
||||
"hpf_start": 86,
|
||||
"hpf_stop": 72,
|
||||
"res_type": "kaiser_fast"
|
||||
}
|
||||
},
|
||||
"sr": 44100,
|
||||
"pre_filter_start": 639,
|
||||
"pre_filter_stop": 640
|
||||
}
|
||||
@ -1,54 +0,0 @@
|
||||
{
|
||||
"bins": 768,
|
||||
"unstable_bins": 7,
|
||||
"reduction_bins": 668,
|
||||
"band": {
|
||||
"1": {
|
||||
"sr": 11025,
|
||||
"hl": 128,
|
||||
"n_fft": 1024,
|
||||
"crop_start": 0,
|
||||
"crop_stop": 186,
|
||||
"lpf_start": 37,
|
||||
"lpf_stop": 73,
|
||||
"res_type": "polyphase"
|
||||
},
|
||||
"2": {
|
||||
"sr": 11025,
|
||||
"hl": 128,
|
||||
"n_fft": 512,
|
||||
"crop_start": 4,
|
||||
"crop_stop": 185,
|
||||
"hpf_start": 36,
|
||||
"hpf_stop": 18,
|
||||
"lpf_start": 93,
|
||||
"lpf_stop": 185,
|
||||
"res_type": "polyphase"
|
||||
},
|
||||
"3": {
|
||||
"sr": 22050,
|
||||
"hl": 256,
|
||||
"n_fft": 512,
|
||||
"crop_start": 46,
|
||||
"crop_stop": 186,
|
||||
"hpf_start": 93,
|
||||
"hpf_stop": 46,
|
||||
"lpf_start": 164,
|
||||
"lpf_stop": 186,
|
||||
"res_type": "polyphase"
|
||||
},
|
||||
"4": {
|
||||
"sr": 44100,
|
||||
"hl": 512,
|
||||
"n_fft": 768,
|
||||
"crop_start": 121,
|
||||
"crop_stop": 382,
|
||||
"hpf_start": 138,
|
||||
"hpf_stop": 123,
|
||||
"res_type": "sinc_medium"
|
||||
}
|
||||
},
|
||||
"sr": 44100,
|
||||
"pre_filter_start": 740,
|
||||
"pre_filter_stop": 768
|
||||
}
|
||||
@ -1,55 +0,0 @@
|
||||
{
|
||||
"bins": 768,
|
||||
"unstable_bins": 7,
|
||||
"mid_side": true,
|
||||
"reduction_bins": 668,
|
||||
"band": {
|
||||
"1": {
|
||||
"sr": 11025,
|
||||
"hl": 128,
|
||||
"n_fft": 1024,
|
||||
"crop_start": 0,
|
||||
"crop_stop": 186,
|
||||
"lpf_start": 37,
|
||||
"lpf_stop": 73,
|
||||
"res_type": "polyphase"
|
||||
},
|
||||
"2": {
|
||||
"sr": 11025,
|
||||
"hl": 128,
|
||||
"n_fft": 512,
|
||||
"crop_start": 4,
|
||||
"crop_stop": 185,
|
||||
"hpf_start": 36,
|
||||
"hpf_stop": 18,
|
||||
"lpf_start": 93,
|
||||
"lpf_stop": 185,
|
||||
"res_type": "polyphase"
|
||||
},
|
||||
"3": {
|
||||
"sr": 22050,
|
||||
"hl": 256,
|
||||
"n_fft": 512,
|
||||
"crop_start": 46,
|
||||
"crop_stop": 186,
|
||||
"hpf_start": 93,
|
||||
"hpf_stop": 46,
|
||||
"lpf_start": 164,
|
||||
"lpf_stop": 186,
|
||||
"res_type": "polyphase"
|
||||
},
|
||||
"4": {
|
||||
"sr": 44100,
|
||||
"hl": 512,
|
||||
"n_fft": 768,
|
||||
"crop_start": 121,
|
||||
"crop_stop": 382,
|
||||
"hpf_start": 138,
|
||||
"hpf_stop": 123,
|
||||
"res_type": "sinc_medium"
|
||||
}
|
||||
},
|
||||
"sr": 44100,
|
||||
"pre_filter_start": 740,
|
||||
"pre_filter_stop": 768
|
||||
}
|
||||
@ -1,55 +0,0 @@
|
||||
{
|
||||
"mid_side_b": true,
|
||||
"bins": 768,
|
||||
"unstable_bins": 7,
|
||||
"reduction_bins": 668,
|
||||
"band": {
|
||||
"1": {
|
||||
"sr": 11025,
|
||||
"hl": 128,
|
||||
"n_fft": 1024,
|
||||
"crop_start": 0,
|
||||
"crop_stop": 186,
|
||||
"lpf_start": 37,
|
||||
"lpf_stop": 73,
|
||||
"res_type": "polyphase"
|
||||
},
|
||||
"2": {
|
||||
"sr": 11025,
|
||||
"hl": 128,
|
||||
"n_fft": 512,
|
||||
"crop_start": 4,
|
||||
"crop_stop": 185,
|
||||
"hpf_start": 36,
|
||||
"hpf_stop": 18,
|
||||
"lpf_start": 93,
|
||||
"lpf_stop": 185,
|
||||
"res_type": "polyphase"
|
||||
},
|
||||
"3": {
|
||||
"sr": 22050,
|
||||
"hl": 256,
|
||||
"n_fft": 512,
|
||||
"crop_start": 46,
|
||||
"crop_stop": 186,
|
||||
"hpf_start": 93,
|
||||
"hpf_stop": 46,
|
||||
"lpf_start": 164,
|
||||
"lpf_stop": 186,
|
||||
"res_type": "polyphase"
|
||||
},
|
||||
"4": {
|
||||
"sr": 44100,
|
||||
"hl": 512,
|
||||
"n_fft": 768,
|
||||
"crop_start": 121,
|
||||
"crop_stop": 382,
|
||||
"hpf_start": 138,
|
||||
"hpf_stop": 123,
|
||||
"res_type": "sinc_medium"
|
||||
}
|
||||
},
|
||||
"sr": 44100,
|
||||
"pre_filter_start": 740,
|
||||
"pre_filter_stop": 768
|
||||
}
|
||||
@ -1,55 +0,0 @@
|
||||
{
|
||||
"mid_side_b": true,
|
||||
"bins": 768,
|
||||
"unstable_bins": 7,
|
||||
"reduction_bins": 668,
|
||||
"band": {
|
||||
"1": {
|
||||
"sr": 11025,
|
||||
"hl": 128,
|
||||
"n_fft": 1024,
|
||||
"crop_start": 0,
|
||||
"crop_stop": 186,
|
||||
"lpf_start": 37,
|
||||
"lpf_stop": 73,
|
||||
"res_type": "polyphase"
|
||||
},
|
||||
"2": {
|
||||
"sr": 11025,
|
||||
"hl": 128,
|
||||
"n_fft": 512,
|
||||
"crop_start": 4,
|
||||
"crop_stop": 185,
|
||||
"hpf_start": 36,
|
||||
"hpf_stop": 18,
|
||||
"lpf_start": 93,
|
||||
"lpf_stop": 185,
|
||||
"res_type": "polyphase"
|
||||
},
|
||||
"3": {
|
||||
"sr": 22050,
|
||||
"hl": 256,
|
||||
"n_fft": 512,
|
||||
"crop_start": 46,
|
||||
"crop_stop": 186,
|
||||
"hpf_start": 93,
|
||||
"hpf_stop": 46,
|
||||
"lpf_start": 164,
|
||||
"lpf_stop": 186,
|
||||
"res_type": "polyphase"
|
||||
},
|
||||
"4": {
|
||||
"sr": 44100,
|
||||
"hl": 512,
|
||||
"n_fft": 768,
|
||||
"crop_start": 121,
|
||||
"crop_stop": 382,
|
||||
"hpf_start": 138,
|
||||
"hpf_stop": 123,
|
||||
"res_type": "sinc_medium"
|
||||
}
|
||||
},
|
||||
"sr": 44100,
|
||||
"pre_filter_start": 740,
|
||||
"pre_filter_stop": 768
|
||||
}
|
||||
@ -1,55 +0,0 @@
|
||||
{
|
||||
"reverse": true,
|
||||
"bins": 768,
|
||||
"unstable_bins": 7,
|
||||
"reduction_bins": 668,
|
||||
"band": {
|
||||
"1": {
|
||||
"sr": 11025,
|
||||
"hl": 128,
|
||||
"n_fft": 1024,
|
||||
"crop_start": 0,
|
||||
"crop_stop": 186,
|
||||
"lpf_start": 37,
|
||||
"lpf_stop": 73,
|
||||
"res_type": "polyphase"
|
||||
},
|
||||
"2": {
|
||||
"sr": 11025,
|
||||
"hl": 128,
|
||||
"n_fft": 512,
|
||||
"crop_start": 4,
|
||||
"crop_stop": 185,
|
||||
"hpf_start": 36,
|
||||
"hpf_stop": 18,
|
||||
"lpf_start": 93,
|
||||
"lpf_stop": 185,
|
||||
"res_type": "polyphase"
|
||||
},
|
||||
"3": {
|
||||
"sr": 22050,
|
||||
"hl": 256,
|
||||
"n_fft": 512,
|
||||
"crop_start": 46,
|
||||
"crop_stop": 186,
|
||||
"hpf_start": 93,
|
||||
"hpf_stop": 46,
|
||||
"lpf_start": 164,
|
||||
"lpf_stop": 186,
|
||||
"res_type": "polyphase"
|
||||
},
|
||||
"4": {
|
||||
"sr": 44100,
|
||||
"hl": 512,
|
||||
"n_fft": 768,
|
||||
"crop_start": 121,
|
||||
"crop_stop": 382,
|
||||
"hpf_start": 138,
|
||||
"hpf_stop": 123,
|
||||
"res_type": "sinc_medium"
|
||||
}
|
||||
},
|
||||
"sr": 44100,
|
||||
"pre_filter_start": 740,
|
||||
"pre_filter_stop": 768
|
||||
}
|
||||
@ -1,55 +0,0 @@
|
||||
{
|
||||
"stereo_w": true,
|
||||
"bins": 768,
|
||||
"unstable_bins": 7,
|
||||
"reduction_bins": 668,
|
||||
"band": {
|
||||
"1": {
|
||||
"sr": 11025,
|
||||
"hl": 128,
|
||||
"n_fft": 1024,
|
||||
"crop_start": 0,
|
||||
"crop_stop": 186,
|
||||
"lpf_start": 37,
|
||||
"lpf_stop": 73,
|
||||
"res_type": "polyphase"
|
||||
},
|
||||
"2": {
|
||||
"sr": 11025,
|
||||
"hl": 128,
|
||||
"n_fft": 512,
|
||||
"crop_start": 4,
|
||||
"crop_stop": 185,
|
||||
"hpf_start": 36,
|
||||
"hpf_stop": 18,
|
||||
"lpf_start": 93,
|
||||
"lpf_stop": 185,
|
||||
"res_type": "polyphase"
|
||||
},
|
||||
"3": {
|
||||
"sr": 22050,
|
||||
"hl": 256,
|
||||
"n_fft": 512,
|
||||
"crop_start": 46,
|
||||
"crop_stop": 186,
|
||||
"hpf_start": 93,
|
||||
"hpf_stop": 46,
|
||||
"lpf_start": 164,
|
||||
"lpf_stop": 186,
|
||||
"res_type": "polyphase"
|
||||
},
|
||||
"4": {
|
||||
"sr": 44100,
|
||||
"hl": 512,
|
||||
"n_fft": 768,
|
||||
"crop_start": 121,
|
||||
"crop_stop": 382,
|
||||
"hpf_start": 138,
|
||||
"hpf_stop": 123,
|
||||
"res_type": "sinc_medium"
|
||||
}
|
||||
},
|
||||
"sr": 44100,
|
||||
"pre_filter_start": 740,
|
||||
"pre_filter_stop": 768
|
||||
}
|
||||
@ -1,54 +0,0 @@
|
||||
{
|
||||
"bins": 672,
|
||||
"unstable_bins": 8,
|
||||
"reduction_bins": 637,
|
||||
"band": {
|
||||
"1": {
|
||||
"sr": 7350,
|
||||
"hl": 80,
|
||||
"n_fft": 640,
|
||||
"crop_start": 0,
|
||||
"crop_stop": 85,
|
||||
"lpf_start": 25,
|
||||
"lpf_stop": 53,
|
||||
"res_type": "polyphase"
|
||||
},
|
||||
"2": {
|
||||
"sr": 7350,
|
||||
"hl": 80,
|
||||
"n_fft": 320,
|
||||
"crop_start": 4,
|
||||
"crop_stop": 87,
|
||||
"hpf_start": 25,
|
||||
"hpf_stop": 12,
|
||||
"lpf_start": 31,
|
||||
"lpf_stop": 62,
|
||||
"res_type": "polyphase"
|
||||
},
|
||||
"3": {
|
||||
"sr": 14700,
|
||||
"hl": 160,
|
||||
"n_fft": 512,
|
||||
"crop_start": 17,
|
||||
"crop_stop": 216,
|
||||
"hpf_start": 48,
|
||||
"hpf_stop": 24,
|
||||
"lpf_start": 139,
|
||||
"lpf_stop": 210,
|
||||
"res_type": "polyphase"
|
||||
},
|
||||
"4": {
|
||||
"sr": 44100,
|
||||
"hl": 480,
|
||||
"n_fft": 960,
|
||||
"crop_start": 78,
|
||||
"crop_stop": 383,
|
||||
"hpf_start": 130,
|
||||
"hpf_stop": 86,
|
||||
"res_type": "kaiser_fast"
|
||||
}
|
||||
},
|
||||
"sr": 44100,
|
||||
"pre_filter_start": 668,
|
||||
"pre_filter_stop": 672
|
||||
}
|
||||
@ -1,55 +0,0 @@
|
||||
{
|
||||
"bins": 672,
|
||||
"unstable_bins": 8,
|
||||
"reduction_bins": 637,
|
||||
"band": {
|
||||
"1": {
|
||||
"sr": 7350,
|
||||
"hl": 80,
|
||||
"n_fft": 640,
|
||||
"crop_start": 0,
|
||||
"crop_stop": 85,
|
||||
"lpf_start": 25,
|
||||
"lpf_stop": 53,
|
||||
"res_type": "polyphase"
|
||||
},
|
||||
"2": {
|
||||
"sr": 7350,
|
||||
"hl": 80,
|
||||
"n_fft": 320,
|
||||
"crop_start": 4,
|
||||
"crop_stop": 87,
|
||||
"hpf_start": 25,
|
||||
"hpf_stop": 12,
|
||||
"lpf_start": 31,
|
||||
"lpf_stop": 62,
|
||||
"res_type": "polyphase"
|
||||
},
|
||||
"3": {
|
||||
"sr": 14700,
|
||||
"hl": 160,
|
||||
"n_fft": 512,
|
||||
"crop_start": 17,
|
||||
"crop_stop": 216,
|
||||
"hpf_start": 48,
|
||||
"hpf_stop": 24,
|
||||
"lpf_start": 139,
|
||||
"lpf_stop": 210,
|
||||
"res_type": "polyphase"
|
||||
},
|
||||
"4": {
|
||||
"sr": 44100,
|
||||
"hl": 480,
|
||||
"n_fft": 960,
|
||||
"crop_start": 78,
|
||||
"crop_stop": 383,
|
||||
"hpf_start": 130,
|
||||
"hpf_stop": 86,
|
||||
"convert_channels": "stereo_n",
|
||||
"res_type": "kaiser_fast"
|
||||
}
|
||||
},
|
||||
"sr": 44100,
|
||||
"pre_filter_start": 668,
|
||||
"pre_filter_stop": 672
|
||||
}
|
||||
@ -1 +0,0 @@
|
||||
Auto
|
||||
@ -1,43 +0,0 @@
|
||||
{
|
||||
"mid_side_b2": true,
|
||||
"bins": 1280,
|
||||
"unstable_bins": 7,
|
||||
"reduction_bins": 565,
|
||||
"band": {
|
||||
"1": {
|
||||
"sr": 11025,
|
||||
"hl": 108,
|
||||
"n_fft": 2048,
|
||||
"crop_start": 0,
|
||||
"crop_stop": 374,
|
||||
"lpf_start": 92,
|
||||
"lpf_stop": 186,
|
||||
"res_type": "polyphase"
|
||||
},
|
||||
"2": {
|
||||
"sr": 22050,
|
||||
"hl": 216,
|
||||
"n_fft": 1536,
|
||||
"crop_start": 0,
|
||||
"crop_stop": 424,
|
||||
"hpf_start": 68,
|
||||
"hpf_stop": 34,
|
||||
"lpf_start": 348,
|
||||
"lpf_stop": 418,
|
||||
"res_type": "polyphase"
|
||||
},
|
||||
"3": {
|
||||
"sr": 44100,
|
||||
"hl": 432,
|
||||
"n_fft": 1280,
|
||||
"crop_start": 132,
|
||||
"crop_stop": 614,
|
||||
"hpf_start": 172,
|
||||
"hpf_stop": 144,
|
||||
"res_type": "polyphase"
|
||||
}
|
||||
},
|
||||
"sr": 44100,
|
||||
"pre_filter_start": 1280,
|
||||
"pre_filter_stop": 1280
|
||||
}
|
||||
@ -1,166 +0,0 @@
|
||||
def provide_model_param_hash(model_hash):
|
||||
#v5 Models
|
||||
if model_hash == '47939caf0cfe52a0e81442b85b971dfd':
|
||||
model_params_set=str('lib_v5/modelparams/4band_44100.json')
|
||||
param_name=str('4band_44100')
|
||||
elif model_hash == '4e4ecb9764c50a8c414fee6e10395bbe':
|
||||
model_params_set=str('lib_v5/modelparams/4band_v2.json')
|
||||
param_name=str('4band_v2')
|
||||
elif model_hash == 'e60a1e84803ce4efc0a6551206cc4b71':
|
||||
model_params_set=str('lib_v5/modelparams/4band_44100.json')
|
||||
param_name=str('4band_44100')
|
||||
elif model_hash == 'a82f14e75892e55e994376edbf0c8435':
|
||||
model_params_set=str('lib_v5/modelparams/4band_44100.json')
|
||||
param_name=str('4band_44100')
|
||||
elif model_hash == '6dd9eaa6f0420af9f1d403aaafa4cc06':
|
||||
model_params_set=str('lib_v5/modelparams/4band_v2_sn.json')
|
||||
param_name=str('4band_v2_sn')
|
||||
elif model_hash == '5c7bbca45a187e81abbbd351606164e5':
|
||||
model_params_set=str('lib_v5/modelparams/3band_44100_msb2.json')
|
||||
param_name=str('3band_44100_msb2')
|
||||
elif model_hash == 'd6b2cb685a058a091e5e7098192d3233':
|
||||
model_params_set=str('lib_v5/modelparams/3band_44100_msb2.json')
|
||||
param_name=str('3band_44100_msb2')
|
||||
elif model_hash == 'c1b9f38170a7c90e96f027992eb7c62b':
|
||||
model_params_set=str('lib_v5/modelparams/4band_44100.json')
|
||||
param_name=str('4band_44100')
|
||||
elif model_hash == 'c3448ec923fa0edf3d03a19e633faa53':
|
||||
model_params_set=str('lib_v5/modelparams/4band_44100.json')
|
||||
param_name=str('4band_44100')
|
||||
elif model_hash == '68aa2c8093d0080704b200d140f59e54':
|
||||
model_params_set=str('lib_v5/modelparams/3band_44100.json')
|
||||
param_name=str('3band_44100.json')
|
||||
elif model_hash == 'fdc83be5b798e4bd29fe00fe6600e147':
|
||||
model_params_set=str('lib_v5/modelparams/3band_44100_mid.json')
|
||||
param_name=str('3band_44100_mid.json')
|
||||
elif model_hash == '2ce34bc92fd57f55db16b7a4def3d745':
|
||||
model_params_set=str('lib_v5/modelparams/3band_44100_mid.json')
|
||||
param_name=str('3band_44100_mid.json')
|
||||
elif model_hash == '52fdca89576f06cf4340b74a4730ee5f':
|
||||
model_params_set=str('lib_v5/modelparams/4band_44100.json')
|
||||
param_name=str('4band_44100.json')
|
||||
elif model_hash == '41191165b05d38fc77f072fa9e8e8a30':
|
||||
model_params_set=str('lib_v5/modelparams/4band_44100.json')
|
||||
param_name=str('4band_44100.json')
|
||||
elif model_hash == '89e83b511ad474592689e562d5b1f80e':
|
||||
model_params_set=str('lib_v5/modelparams/2band_32000.json')
|
||||
param_name=str('2band_32000.json')
|
||||
elif model_hash == '0b954da81d453b716b114d6d7c95177f':
|
||||
model_params_set=str('lib_v5/modelparams/2band_32000.json')
|
||||
param_name=str('2band_32000.json')
|
||||
|
||||
#v4 Models
|
||||
|
||||
elif model_hash == '6a00461c51c2920fd68937d4609ed6c8':
|
||||
model_params_set=str('lib_v5/modelparams/1band_sr16000_hl512.json')
|
||||
param_name=str('1band_sr16000_hl512')
|
||||
elif model_hash == '0ab504864d20f1bd378fe9c81ef37140':
|
||||
model_params_set=str('lib_v5/modelparams/1band_sr32000_hl512.json')
|
||||
param_name=str('1band_sr32000_hl512')
|
||||
elif model_hash == '7dd21065bf91c10f7fccb57d7d83b07f':
|
||||
model_params_set=str('lib_v5/modelparams/1band_sr32000_hl512.json')
|
||||
param_name=str('1band_sr32000_hl512')
|
||||
elif model_hash == '80ab74d65e515caa3622728d2de07d23':
|
||||
model_params_set=str('lib_v5/modelparams/1band_sr32000_hl512.json')
|
||||
param_name=str('1band_sr32000_hl512')
|
||||
elif model_hash == 'edc115e7fc523245062200c00caa847f':
|
||||
model_params_set=str('lib_v5/modelparams/1band_sr33075_hl384.json')
|
||||
param_name=str('1band_sr33075_hl384')
|
||||
elif model_hash == '28063e9f6ab5b341c5f6d3c67f2045b7':
|
||||
model_params_set=str('lib_v5/modelparams/1band_sr33075_hl384.json')
|
||||
param_name=str('1band_sr33075_hl384')
|
||||
elif model_hash == 'b58090534c52cbc3e9b5104bad666ef2':
|
||||
model_params_set=str('lib_v5/modelparams/1band_sr44100_hl512.json')
|
||||
param_name=str('1band_sr44100_hl512')
|
||||
elif model_hash == '0cdab9947f1b0928705f518f3c78ea8f':
|
||||
model_params_set=str('lib_v5/modelparams/1band_sr44100_hl512.json')
|
||||
param_name=str('1band_sr44100_hl512')
|
||||
elif model_hash == 'ae702fed0238afb5346db8356fe25f13':
|
||||
model_params_set=str('lib_v5/modelparams/1band_sr44100_hl1024.json')
|
||||
param_name=str('1band_sr44100_hl1024')
|
||||
else:
|
||||
model_params_set=str('Not Found Using Hash')
|
||||
param_name=str('Not Found Using Hash')
|
||||
|
||||
model_params = model_params_set, param_name
|
||||
|
||||
return model_params
|
||||
|
||||
def provide_model_param_name(ModelName):
|
||||
#1 Band
|
||||
if '1band_sr16000_hl512' in ModelName:
|
||||
model_params_set=str('lib_v5/modelparams/1band_sr16000_hl512.json')
|
||||
param_name=str('1band_sr16000_hl512')
|
||||
elif '1band_sr32000_hl512' in ModelName:
|
||||
model_params_set=str('lib_v5/modelparams/1band_sr32000_hl512.json')
|
||||
param_name=str('1band_sr32000_hl512')
|
||||
elif '1band_sr33075_hl384' in ModelName:
|
||||
model_params_set=str('lib_v5/modelparams/1band_sr33075_hl384.json')
|
||||
param_name=str('1band_sr33075_hl384')
|
||||
elif '1band_sr44100_hl256' in ModelName:
|
||||
model_params_set=str('lib_v5/modelparams/1band_sr44100_hl256.json')
|
||||
param_name=str('1band_sr44100_hl256')
|
||||
elif '1band_sr44100_hl512' in ModelName:
|
||||
model_params_set=str('lib_v5/modelparams/1band_sr44100_hl512.json')
|
||||
param_name=str('1band_sr44100_hl512')
|
||||
elif '1band_sr44100_hl1024' in ModelName:
|
||||
model_params_set=str('lib_v5/modelparams/1band_sr44100_hl1024.json')
|
||||
param_name=str('1band_sr44100_hl1024')
|
||||
|
||||
#2 Band
|
||||
elif '2band_44100_lofi' in ModelName:
|
||||
model_params_set=str('lib_v5/modelparams/2band_44100_lofi.json')
|
||||
param_name=str('2band_44100_lofi')
|
||||
elif '2band_32000' in ModelName:
|
||||
model_params_set=str('lib_v5/modelparams/2band_32000.json')
|
||||
param_name=str('2band_32000')
|
||||
elif '2band_48000' in ModelName:
|
||||
model_params_set=str('lib_v5/modelparams/2band_48000.json')
|
||||
param_name=str('2band_48000')
|
||||
|
||||
#3 Band
|
||||
elif '3band_44100' in ModelName:
|
||||
model_params_set=str('lib_v5/modelparams/3band_44100.json')
|
||||
param_name=str('3band_44100')
|
||||
elif '3band_44100_mid' in ModelName:
|
||||
model_params_set=str('lib_v5/modelparams/3band_44100_mid.json')
|
||||
param_name=str('3band_44100_mid')
|
||||
elif '3band_44100_msb2' in ModelName:
|
||||
model_params_set=str('lib_v5/modelparams/3band_44100_msb2.json')
|
||||
param_name=str('3band_44100_msb2')
|
||||
|
||||
#4 Band
|
||||
elif '4band_44100' in ModelName:
|
||||
model_params_set=str('lib_v5/modelparams/4band_44100.json')
|
||||
param_name=str('4band_44100')
|
||||
elif '4band_44100_mid' in ModelName:
|
||||
model_params_set=str('lib_v5/modelparams/4band_44100_mid.json')
|
||||
param_name=str('4band_44100_mid')
|
||||
elif '4band_44100_msb' in ModelName:
|
||||
model_params_set=str('lib_v5/modelparams/4band_44100_msb.json')
|
||||
param_name=str('4band_44100_msb')
|
||||
elif '4band_44100_msb2' in ModelName:
|
||||
model_params_set=str('lib_v5/modelparams/4band_44100_msb2.json')
|
||||
param_name=str('4band_44100_msb2')
|
||||
elif '4band_44100_reverse' in ModelName:
|
||||
model_params_set=str('lib_v5/modelparams/4band_44100_reverse.json')
|
||||
param_name=str('4band_44100_reverse')
|
||||
elif '4band_44100_sw' in ModelName:
|
||||
model_params_set=str('lib_v5/modelparams/4band_44100_sw.json')
|
||||
param_name=str('4band_44100_sw')
|
||||
elif '4band_v2' in ModelName:
|
||||
model_params_set=str('lib_v5/modelparams/4band_v2.json')
|
||||
param_name=str('4band_v2')
|
||||
elif '4band_v2_sn' in ModelName:
|
||||
model_params_set=str('lib_v5/modelparams/4band_v2_sn.json')
|
||||
param_name=str('4band_v2_sn')
|
||||
elif 'tmodelparam' in ModelName:
|
||||
model_params_set=str('lib_v5/modelparams/tmodelparam.json')
|
||||
param_name=str('User Model Param Set')
|
||||
else:
|
||||
model_params_set=str('Not Found Using Name')
|
||||
param_name=str('Not Found Using Name')
|
||||
|
||||
model_params = model_params_set, param_name
|
||||
|
||||
return model_params
|
||||
113
lib_v5/nets.py
113
lib_v5/nets.py
@ -1,113 +0,0 @@
|
||||
import torch
|
||||
from torch import nn
|
||||
import torch.nn.functional as F
|
||||
|
||||
from lib_v5 import layers
|
||||
from lib_v5 import spec_utils
|
||||
|
||||
|
||||
class BaseASPPNet(nn.Module):
|
||||
|
||||
def __init__(self, nin, ch, dilations=(4, 8, 16)):
|
||||
super(BaseASPPNet, self).__init__()
|
||||
self.enc1 = layers.Encoder(nin, ch, 3, 2, 1)
|
||||
self.enc2 = layers.Encoder(ch, ch * 2, 3, 2, 1)
|
||||
self.enc3 = layers.Encoder(ch * 2, ch * 4, 3, 2, 1)
|
||||
self.enc4 = layers.Encoder(ch * 4, ch * 8, 3, 2, 1)
|
||||
|
||||
self.aspp = layers.ASPPModule(ch * 8, ch * 16, dilations)
|
||||
|
||||
self.dec4 = layers.Decoder(ch * (8 + 16), ch * 8, 3, 1, 1)
|
||||
self.dec3 = layers.Decoder(ch * (4 + 8), ch * 4, 3, 1, 1)
|
||||
self.dec2 = layers.Decoder(ch * (2 + 4), ch * 2, 3, 1, 1)
|
||||
self.dec1 = layers.Decoder(ch * (1 + 2), ch, 3, 1, 1)
|
||||
|
||||
def __call__(self, x):
|
||||
h, e1 = self.enc1(x)
|
||||
h, e2 = self.enc2(h)
|
||||
h, e3 = self.enc3(h)
|
||||
h, e4 = self.enc4(h)
|
||||
|
||||
h = self.aspp(h)
|
||||
|
||||
h = self.dec4(h, e4)
|
||||
h = self.dec3(h, e3)
|
||||
h = self.dec2(h, e2)
|
||||
h = self.dec1(h, e1)
|
||||
|
||||
return h
|
||||
|
||||
|
||||
class CascadedASPPNet(nn.Module):
|
||||
|
||||
def __init__(self, n_fft):
|
||||
super(CascadedASPPNet, self).__init__()
|
||||
self.stg1_low_band_net = BaseASPPNet(2, 16)
|
||||
self.stg1_high_band_net = BaseASPPNet(2, 16)
|
||||
|
||||
self.stg2_bridge = layers.Conv2DBNActiv(18, 8, 1, 1, 0)
|
||||
self.stg2_full_band_net = BaseASPPNet(8, 16)
|
||||
|
||||
self.stg3_bridge = layers.Conv2DBNActiv(34, 16, 1, 1, 0)
|
||||
self.stg3_full_band_net = BaseASPPNet(16, 32)
|
||||
|
||||
self.out = nn.Conv2d(32, 2, 1, bias=False)
|
||||
self.aux1_out = nn.Conv2d(16, 2, 1, bias=False)
|
||||
self.aux2_out = nn.Conv2d(16, 2, 1, bias=False)
|
||||
|
||||
self.max_bin = n_fft // 2
|
||||
self.output_bin = n_fft // 2 + 1
|
||||
|
||||
self.offset = 128
|
||||
|
||||
def forward(self, x, aggressiveness=None):
|
||||
mix = x.detach()
|
||||
x = x.clone()
|
||||
|
||||
x = x[:, :, :self.max_bin]
|
||||
|
||||
bandw = x.size()[2] // 2
|
||||
aux1 = torch.cat([
|
||||
self.stg1_low_band_net(x[:, :, :bandw]),
|
||||
self.stg1_high_band_net(x[:, :, bandw:])
|
||||
], dim=2)
|
||||
|
||||
h = torch.cat([x, aux1], dim=1)
|
||||
aux2 = self.stg2_full_band_net(self.stg2_bridge(h))
|
||||
|
||||
h = torch.cat([x, aux1, aux2], dim=1)
|
||||
h = self.stg3_full_band_net(self.stg3_bridge(h))
|
||||
|
||||
mask = torch.sigmoid(self.out(h))
|
||||
mask = F.pad(
|
||||
input=mask,
|
||||
pad=(0, 0, 0, self.output_bin - mask.size()[2]),
|
||||
mode='replicate')
|
||||
|
||||
if self.training:
|
||||
aux1 = torch.sigmoid(self.aux1_out(aux1))
|
||||
aux1 = F.pad(
|
||||
input=aux1,
|
||||
pad=(0, 0, 0, self.output_bin - aux1.size()[2]),
|
||||
mode='replicate')
|
||||
aux2 = torch.sigmoid(self.aux2_out(aux2))
|
||||
aux2 = F.pad(
|
||||
input=aux2,
|
||||
pad=(0, 0, 0, self.output_bin - aux2.size()[2]),
|
||||
mode='replicate')
|
||||
return mask * mix, aux1 * mix, aux2 * mix
|
||||
else:
|
||||
if aggressiveness:
|
||||
mask[:, :, :aggressiveness['split_bin']] = torch.pow(mask[:, :, :aggressiveness['split_bin']], 1 + aggressiveness['value'] / 3)
|
||||
mask[:, :, aggressiveness['split_bin']:] = torch.pow(mask[:, :, aggressiveness['split_bin']:], 1 + aggressiveness['value'])
|
||||
|
||||
return mask * mix
|
||||
|
||||
def predict(self, x_mag, aggressiveness=None):
|
||||
h = self.forward(x_mag, aggressiveness)
|
||||
|
||||
if self.offset > 0:
|
||||
h = h[:, :, :, self.offset:-self.offset]
|
||||
assert h.size()[3] > 0
|
||||
|
||||
return h
|
||||
@ -1,112 +0,0 @@
|
||||
import torch
|
||||
from torch import nn
|
||||
import torch.nn.functional as F
|
||||
|
||||
from lib_v5 import layers_123821KB as layers
|
||||
|
||||
|
||||
class BaseASPPNet(nn.Module):
|
||||
|
||||
def __init__(self, nin, ch, dilations=(4, 8, 16)):
|
||||
super(BaseASPPNet, self).__init__()
|
||||
self.enc1 = layers.Encoder(nin, ch, 3, 2, 1)
|
||||
self.enc2 = layers.Encoder(ch, ch * 2, 3, 2, 1)
|
||||
self.enc3 = layers.Encoder(ch * 2, ch * 4, 3, 2, 1)
|
||||
self.enc4 = layers.Encoder(ch * 4, ch * 8, 3, 2, 1)
|
||||
|
||||
self.aspp = layers.ASPPModule(ch * 8, ch * 16, dilations)
|
||||
|
||||
self.dec4 = layers.Decoder(ch * (8 + 16), ch * 8, 3, 1, 1)
|
||||
self.dec3 = layers.Decoder(ch * (4 + 8), ch * 4, 3, 1, 1)
|
||||
self.dec2 = layers.Decoder(ch * (2 + 4), ch * 2, 3, 1, 1)
|
||||
self.dec1 = layers.Decoder(ch * (1 + 2), ch, 3, 1, 1)
|
||||
|
||||
def __call__(self, x):
|
||||
h, e1 = self.enc1(x)
|
||||
h, e2 = self.enc2(h)
|
||||
h, e3 = self.enc3(h)
|
||||
h, e4 = self.enc4(h)
|
||||
|
||||
h = self.aspp(h)
|
||||
|
||||
h = self.dec4(h, e4)
|
||||
h = self.dec3(h, e3)
|
||||
h = self.dec2(h, e2)
|
||||
h = self.dec1(h, e1)
|
||||
|
||||
return h
|
||||
|
||||
|
||||
class CascadedASPPNet(nn.Module):
|
||||
|
||||
def __init__(self, n_fft):
|
||||
super(CascadedASPPNet, self).__init__()
|
||||
self.stg1_low_band_net = BaseASPPNet(2, 32)
|
||||
self.stg1_high_band_net = BaseASPPNet(2, 32)
|
||||
|
||||
self.stg2_bridge = layers.Conv2DBNActiv(34, 16, 1, 1, 0)
|
||||
self.stg2_full_band_net = BaseASPPNet(16, 32)
|
||||
|
||||
self.stg3_bridge = layers.Conv2DBNActiv(66, 32, 1, 1, 0)
|
||||
self.stg3_full_band_net = BaseASPPNet(32, 64)
|
||||
|
||||
self.out = nn.Conv2d(64, 2, 1, bias=False)
|
||||
self.aux1_out = nn.Conv2d(32, 2, 1, bias=False)
|
||||
self.aux2_out = nn.Conv2d(32, 2, 1, bias=False)
|
||||
|
||||
self.max_bin = n_fft // 2
|
||||
self.output_bin = n_fft // 2 + 1
|
||||
|
||||
self.offset = 128
|
||||
|
||||
def forward(self, x, aggressiveness=None):
|
||||
mix = x.detach()
|
||||
x = x.clone()
|
||||
|
||||
x = x[:, :, :self.max_bin]
|
||||
|
||||
bandw = x.size()[2] // 2
|
||||
aux1 = torch.cat([
|
||||
self.stg1_low_band_net(x[:, :, :bandw]),
|
||||
self.stg1_high_band_net(x[:, :, bandw:])
|
||||
], dim=2)
|
||||
|
||||
h = torch.cat([x, aux1], dim=1)
|
||||
aux2 = self.stg2_full_band_net(self.stg2_bridge(h))
|
||||
|
||||
h = torch.cat([x, aux1, aux2], dim=1)
|
||||
h = self.stg3_full_band_net(self.stg3_bridge(h))
|
||||
|
||||
mask = torch.sigmoid(self.out(h))
|
||||
mask = F.pad(
|
||||
input=mask,
|
||||
pad=(0, 0, 0, self.output_bin - mask.size()[2]),
|
||||
mode='replicate')
|
||||
|
||||
if self.training:
|
||||
aux1 = torch.sigmoid(self.aux1_out(aux1))
|
||||
aux1 = F.pad(
|
||||
input=aux1,
|
||||
pad=(0, 0, 0, self.output_bin - aux1.size()[2]),
|
||||
mode='replicate')
|
||||
aux2 = torch.sigmoid(self.aux2_out(aux2))
|
||||
aux2 = F.pad(
|
||||
input=aux2,
|
||||
pad=(0, 0, 0, self.output_bin - aux2.size()[2]),
|
||||
mode='replicate')
|
||||
return mask * mix, aux1 * mix, aux2 * mix
|
||||
else:
|
||||
if aggressiveness:
|
||||
mask[:, :, :aggressiveness['split_bin']] = torch.pow(mask[:, :, :aggressiveness['split_bin']], 1 + aggressiveness['value'] / 3)
|
||||
mask[:, :, aggressiveness['split_bin']:] = torch.pow(mask[:, :, aggressiveness['split_bin']:], 1 + aggressiveness['value'])
|
||||
|
||||
return mask * mix
|
||||
|
||||
def predict(self, x_mag, aggressiveness=None):
|
||||
h = self.forward(x_mag, aggressiveness)
|
||||
|
||||
if self.offset > 0:
|
||||
h = h[:, :, :, self.offset:-self.offset]
|
||||
assert h.size()[3] > 0
|
||||
|
||||
return h
|
||||
@ -1,112 +0,0 @@
|
||||
import torch
|
||||
from torch import nn
|
||||
import torch.nn.functional as F
|
||||
|
||||
from lib_v5 import layers_123821KB as layers
|
||||
|
||||
|
||||
class BaseASPPNet(nn.Module):
|
||||
|
||||
def __init__(self, nin, ch, dilations=(4, 8, 16)):
|
||||
super(BaseASPPNet, self).__init__()
|
||||
self.enc1 = layers.Encoder(nin, ch, 3, 2, 1)
|
||||
self.enc2 = layers.Encoder(ch, ch * 2, 3, 2, 1)
|
||||
self.enc3 = layers.Encoder(ch * 2, ch * 4, 3, 2, 1)
|
||||
self.enc4 = layers.Encoder(ch * 4, ch * 8, 3, 2, 1)
|
||||
|
||||
self.aspp = layers.ASPPModule(ch * 8, ch * 16, dilations)
|
||||
|
||||
self.dec4 = layers.Decoder(ch * (8 + 16), ch * 8, 3, 1, 1)
|
||||
self.dec3 = layers.Decoder(ch * (4 + 8), ch * 4, 3, 1, 1)
|
||||
self.dec2 = layers.Decoder(ch * (2 + 4), ch * 2, 3, 1, 1)
|
||||
self.dec1 = layers.Decoder(ch * (1 + 2), ch, 3, 1, 1)
|
||||
|
||||
def __call__(self, x):
|
||||
h, e1 = self.enc1(x)
|
||||
h, e2 = self.enc2(h)
|
||||
h, e3 = self.enc3(h)
|
||||
h, e4 = self.enc4(h)
|
||||
|
||||
h = self.aspp(h)
|
||||
|
||||
h = self.dec4(h, e4)
|
||||
h = self.dec3(h, e3)
|
||||
h = self.dec2(h, e2)
|
||||
h = self.dec1(h, e1)
|
||||
|
||||
return h
|
||||
|
||||
|
||||
class CascadedASPPNet(nn.Module):
|
||||
|
||||
def __init__(self, n_fft):
|
||||
super(CascadedASPPNet, self).__init__()
|
||||
self.stg1_low_band_net = BaseASPPNet(2, 32)
|
||||
self.stg1_high_band_net = BaseASPPNet(2, 32)
|
||||
|
||||
self.stg2_bridge = layers.Conv2DBNActiv(34, 16, 1, 1, 0)
|
||||
self.stg2_full_band_net = BaseASPPNet(16, 32)
|
||||
|
||||
self.stg3_bridge = layers.Conv2DBNActiv(66, 32, 1, 1, 0)
|
||||
self.stg3_full_band_net = BaseASPPNet(32, 64)
|
||||
|
||||
self.out = nn.Conv2d(64, 2, 1, bias=False)
|
||||
self.aux1_out = nn.Conv2d(32, 2, 1, bias=False)
|
||||
self.aux2_out = nn.Conv2d(32, 2, 1, bias=False)
|
||||
|
||||
self.max_bin = n_fft // 2
|
||||
self.output_bin = n_fft // 2 + 1
|
||||
|
||||
self.offset = 128
|
||||
|
||||
def forward(self, x, aggressiveness=None):
|
||||
mix = x.detach()
|
||||
x = x.clone()
|
||||
|
||||
x = x[:, :, :self.max_bin]
|
||||
|
||||
bandw = x.size()[2] // 2
|
||||
aux1 = torch.cat([
|
||||
self.stg1_low_band_net(x[:, :, :bandw]),
|
||||
self.stg1_high_band_net(x[:, :, bandw:])
|
||||
], dim=2)
|
||||
|
||||
h = torch.cat([x, aux1], dim=1)
|
||||
aux2 = self.stg2_full_band_net(self.stg2_bridge(h))
|
||||
|
||||
h = torch.cat([x, aux1, aux2], dim=1)
|
||||
h = self.stg3_full_band_net(self.stg3_bridge(h))
|
||||
|
||||
mask = torch.sigmoid(self.out(h))
|
||||
mask = F.pad(
|
||||
input=mask,
|
||||
pad=(0, 0, 0, self.output_bin - mask.size()[2]),
|
||||
mode='replicate')
|
||||
|
||||
if self.training:
|
||||
aux1 = torch.sigmoid(self.aux1_out(aux1))
|
||||
aux1 = F.pad(
|
||||
input=aux1,
|
||||
pad=(0, 0, 0, self.output_bin - aux1.size()[2]),
|
||||
mode='replicate')
|
||||
aux2 = torch.sigmoid(self.aux2_out(aux2))
|
||||
aux2 = F.pad(
|
||||
input=aux2,
|
||||
pad=(0, 0, 0, self.output_bin - aux2.size()[2]),
|
||||
mode='replicate')
|
||||
return mask * mix, aux1 * mix, aux2 * mix
|
||||
else:
|
||||
if aggressiveness:
|
||||
mask[:, :, :aggressiveness['split_bin']] = torch.pow(mask[:, :, :aggressiveness['split_bin']], 1 + aggressiveness['value'] / 3)
|
||||
mask[:, :, aggressiveness['split_bin']:] = torch.pow(mask[:, :, aggressiveness['split_bin']:], 1 + aggressiveness['value'])
|
||||
|
||||
return mask * mix
|
||||
|
||||
def predict(self, x_mag, aggressiveness=None):
|
||||
h = self.forward(x_mag, aggressiveness)
|
||||
|
||||
if self.offset > 0:
|
||||
h = h[:, :, :, self.offset:-self.offset]
|
||||
assert h.size()[3] > 0
|
||||
|
||||
return h
|
||||
@ -1,116 +0,0 @@
|
||||
import torch
|
||||
from torch import nn
|
||||
import torch.nn.functional as F
|
||||
|
||||
from lib_v5 import layers_129605KB as layers
|
||||
|
||||
|
||||
class BaseASPPNet(nn.Module):
|
||||
|
||||
def __init__(self, nin, ch, dilations=(4, 8, 16, 32)):
|
||||
super(BaseASPPNet, self).__init__()
|
||||
self.enc1 = layers.Encoder(nin, ch, 3, 2, 1)
|
||||
self.enc2 = layers.Encoder(ch, ch * 2, 3, 2, 1)
|
||||
self.enc3 = layers.Encoder(ch * 2, ch * 4, 3, 2, 1)
|
||||
self.enc4 = layers.Encoder(ch * 4, ch * 8, 3, 2, 1)
|
||||
self.enc5 = layers.Encoder(ch * 8, ch * 16, 3, 2, 1)
|
||||
|
||||
self.aspp = layers.ASPPModule(ch * 16, ch * 32, dilations)
|
||||
|
||||
self.dec5 = layers.Decoder(ch * (16 + 32), ch * 16, 3, 1, 1)
|
||||
self.dec4 = layers.Decoder(ch * (8 + 16), ch * 8, 3, 1, 1)
|
||||
self.dec3 = layers.Decoder(ch * (4 + 8), ch * 4, 3, 1, 1)
|
||||
self.dec2 = layers.Decoder(ch * (2 + 4), ch * 2, 3, 1, 1)
|
||||
self.dec1 = layers.Decoder(ch * (1 + 2), ch, 3, 1, 1)
|
||||
|
||||
def __call__(self, x):
|
||||
h, e1 = self.enc1(x)
|
||||
h, e2 = self.enc2(h)
|
||||
h, e3 = self.enc3(h)
|
||||
h, e4 = self.enc4(h)
|
||||
h, e5 = self.enc5(h)
|
||||
|
||||
h = self.aspp(h)
|
||||
|
||||
h = self.dec5(h, e5)
|
||||
h = self.dec4(h, e4)
|
||||
h = self.dec3(h, e3)
|
||||
h = self.dec2(h, e2)
|
||||
h = self.dec1(h, e1)
|
||||
|
||||
return h
|
||||
|
||||
|
||||
class CascadedASPPNet(nn.Module):
|
||||
|
||||
def __init__(self, n_fft):
|
||||
super(CascadedASPPNet, self).__init__()
|
||||
self.stg1_low_band_net = BaseASPPNet(2, 16)
|
||||
self.stg1_high_band_net = BaseASPPNet(2, 16)
|
||||
|
||||
self.stg2_bridge = layers.Conv2DBNActiv(18, 8, 1, 1, 0)
|
||||
self.stg2_full_band_net = BaseASPPNet(8, 16)
|
||||
|
||||
self.stg3_bridge = layers.Conv2DBNActiv(34, 16, 1, 1, 0)
|
||||
self.stg3_full_band_net = BaseASPPNet(16, 32)
|
||||
|
||||
self.out = nn.Conv2d(32, 2, 1, bias=False)
|
||||
self.aux1_out = nn.Conv2d(16, 2, 1, bias=False)
|
||||
self.aux2_out = nn.Conv2d(16, 2, 1, bias=False)
|
||||
|
||||
self.max_bin = n_fft // 2
|
||||
self.output_bin = n_fft // 2 + 1
|
||||
|
||||
self.offset = 128
|
||||
|
||||
def forward(self, x, aggressiveness=None):
|
||||
mix = x.detach()
|
||||
x = x.clone()
|
||||
|
||||
x = x[:, :, :self.max_bin]
|
||||
|
||||
bandw = x.size()[2] // 2
|
||||
aux1 = torch.cat([
|
||||
self.stg1_low_band_net(x[:, :, :bandw]),
|
||||
self.stg1_high_band_net(x[:, :, bandw:])
|
||||
], dim=2)
|
||||
|
||||
h = torch.cat([x, aux1], dim=1)
|
||||
aux2 = self.stg2_full_band_net(self.stg2_bridge(h))
|
||||
|
||||
h = torch.cat([x, aux1, aux2], dim=1)
|
||||
h = self.stg3_full_band_net(self.stg3_bridge(h))
|
||||
|
||||
mask = torch.sigmoid(self.out(h))
|
||||
mask = F.pad(
|
||||
input=mask,
|
||||
pad=(0, 0, 0, self.output_bin - mask.size()[2]),
|
||||
mode='replicate')
|
||||
|
||||
if self.training:
|
||||
aux1 = torch.sigmoid(self.aux1_out(aux1))
|
||||
aux1 = F.pad(
|
||||
input=aux1,
|
||||
pad=(0, 0, 0, self.output_bin - aux1.size()[2]),
|
||||
mode='replicate')
|
||||
aux2 = torch.sigmoid(self.aux2_out(aux2))
|
||||
aux2 = F.pad(
|
||||
input=aux2,
|
||||
pad=(0, 0, 0, self.output_bin - aux2.size()[2]),
|
||||
mode='replicate')
|
||||
return mask * mix, aux1 * mix, aux2 * mix
|
||||
else:
|
||||
if aggressiveness:
|
||||
mask[:, :, :aggressiveness['split_bin']] = torch.pow(mask[:, :, :aggressiveness['split_bin']], 1 + aggressiveness['value'] / 3)
|
||||
mask[:, :, aggressiveness['split_bin']:] = torch.pow(mask[:, :, aggressiveness['split_bin']:], 1 + aggressiveness['value'])
|
||||
|
||||
return mask * mix
|
||||
|
||||
def predict(self, x_mag, aggressiveness=None):
|
||||
h = self.forward(x_mag, aggressiveness)
|
||||
|
||||
if self.offset > 0:
|
||||
h = h[:, :, :, self.offset:-self.offset]
|
||||
assert h.size()[3] > 0
|
||||
|
||||
return h
|
||||
@ -1,112 +0,0 @@
|
||||
import torch
|
||||
from torch import nn
|
||||
import torch.nn.functional as F
|
||||
|
||||
from lib_v5 import layers_33966KB as layers
|
||||
|
||||
|
||||
class BaseASPPNet(nn.Module):
|
||||
|
||||
def __init__(self, nin, ch, dilations=(4, 8, 16, 32)):
|
||||
super(BaseASPPNet, self).__init__()
|
||||
self.enc1 = layers.Encoder(nin, ch, 3, 2, 1)
|
||||
self.enc2 = layers.Encoder(ch, ch * 2, 3, 2, 1)
|
||||
self.enc3 = layers.Encoder(ch * 2, ch * 4, 3, 2, 1)
|
||||
self.enc4 = layers.Encoder(ch * 4, ch * 8, 3, 2, 1)
|
||||
|
||||
self.aspp = layers.ASPPModule(ch * 8, ch * 16, dilations)
|
||||
|
||||
self.dec4 = layers.Decoder(ch * (8 + 16), ch * 8, 3, 1, 1)
|
||||
self.dec3 = layers.Decoder(ch * (4 + 8), ch * 4, 3, 1, 1)
|
||||
self.dec2 = layers.Decoder(ch * (2 + 4), ch * 2, 3, 1, 1)
|
||||
self.dec1 = layers.Decoder(ch * (1 + 2), ch, 3, 1, 1)
|
||||
|
||||
def __call__(self, x):
|
||||
h, e1 = self.enc1(x)
|
||||
h, e2 = self.enc2(h)
|
||||
h, e3 = self.enc3(h)
|
||||
h, e4 = self.enc4(h)
|
||||
|
||||
h = self.aspp(h)
|
||||
|
||||
h = self.dec4(h, e4)
|
||||
h = self.dec3(h, e3)
|
||||
h = self.dec2(h, e2)
|
||||
h = self.dec1(h, e1)
|
||||
|
||||
return h
|
||||
|
||||
|
||||
class CascadedASPPNet(nn.Module):
|
||||
|
||||
def __init__(self, n_fft):
|
||||
super(CascadedASPPNet, self).__init__()
|
||||
self.stg1_low_band_net = BaseASPPNet(2, 16)
|
||||
self.stg1_high_band_net = BaseASPPNet(2, 16)
|
||||
|
||||
self.stg2_bridge = layers.Conv2DBNActiv(18, 8, 1, 1, 0)
|
||||
self.stg2_full_band_net = BaseASPPNet(8, 16)
|
||||
|
||||
self.stg3_bridge = layers.Conv2DBNActiv(34, 16, 1, 1, 0)
|
||||
self.stg3_full_band_net = BaseASPPNet(16, 32)
|
||||
|
||||
self.out = nn.Conv2d(32, 2, 1, bias=False)
|
||||
self.aux1_out = nn.Conv2d(16, 2, 1, bias=False)
|
||||
self.aux2_out = nn.Conv2d(16, 2, 1, bias=False)
|
||||
|
||||
self.max_bin = n_fft // 2
|
||||
self.output_bin = n_fft // 2 + 1
|
||||
|
||||
self.offset = 128
|
||||
|
||||
def forward(self, x, aggressiveness=None):
|
||||
mix = x.detach()
|
||||
x = x.clone()
|
||||
|
||||
x = x[:, :, :self.max_bin]
|
||||
|
||||
bandw = x.size()[2] // 2
|
||||
aux1 = torch.cat([
|
||||
self.stg1_low_band_net(x[:, :, :bandw]),
|
||||
self.stg1_high_band_net(x[:, :, bandw:])
|
||||
], dim=2)
|
||||
|
||||
h = torch.cat([x, aux1], dim=1)
|
||||
aux2 = self.stg2_full_band_net(self.stg2_bridge(h))
|
||||
|
||||
h = torch.cat([x, aux1, aux2], dim=1)
|
||||
h = self.stg3_full_band_net(self.stg3_bridge(h))
|
||||
|
||||
mask = torch.sigmoid(self.out(h))
|
||||
mask = F.pad(
|
||||
input=mask,
|
||||
pad=(0, 0, 0, self.output_bin - mask.size()[2]),
|
||||
mode='replicate')
|
||||
|
||||
if self.training:
|
||||
aux1 = torch.sigmoid(self.aux1_out(aux1))
|
||||
aux1 = F.pad(
|
||||
input=aux1,
|
||||
pad=(0, 0, 0, self.output_bin - aux1.size()[2]),
|
||||
mode='replicate')
|
||||
aux2 = torch.sigmoid(self.aux2_out(aux2))
|
||||
aux2 = F.pad(
|
||||
input=aux2,
|
||||
pad=(0, 0, 0, self.output_bin - aux2.size()[2]),
|
||||
mode='replicate')
|
||||
return mask * mix, aux1 * mix, aux2 * mix
|
||||
else:
|
||||
if aggressiveness:
|
||||
mask[:, :, :aggressiveness['split_bin']] = torch.pow(mask[:, :, :aggressiveness['split_bin']], 1 + aggressiveness['value'] / 3)
|
||||
mask[:, :, aggressiveness['split_bin']:] = torch.pow(mask[:, :, aggressiveness['split_bin']:], 1 + aggressiveness['value'])
|
||||
|
||||
return mask * mix
|
||||
|
||||
def predict(self, x_mag, aggressiveness=None):
|
||||
h = self.forward(x_mag, aggressiveness)
|
||||
|
||||
if self.offset > 0:
|
||||
h = h[:, :, :, self.offset:-self.offset]
|
||||
assert h.size()[3] > 0
|
||||
|
||||
return h
|
||||
@ -1,113 +0,0 @@
|
||||
import torch
|
||||
import numpy as np
|
||||
from torch import nn
|
||||
import torch.nn.functional as F
|
||||
|
||||
from lib_v5 import layers_537238KB as layers
|
||||
|
||||
|
||||
class BaseASPPNet(nn.Module):
|
||||
|
||||
def __init__(self, nin, ch, dilations=(4, 8, 16)):
|
||||
super(BaseASPPNet, self).__init__()
|
||||
self.enc1 = layers.Encoder(nin, ch, 3, 2, 1)
|
||||
self.enc2 = layers.Encoder(ch, ch * 2, 3, 2, 1)
|
||||
self.enc3 = layers.Encoder(ch * 2, ch * 4, 3, 2, 1)
|
||||
self.enc4 = layers.Encoder(ch * 4, ch * 8, 3, 2, 1)
|
||||
|
||||
self.aspp = layers.ASPPModule(ch * 8, ch * 16, dilations)
|
||||
|
||||
self.dec4 = layers.Decoder(ch * (8 + 16), ch * 8, 3, 1, 1)
|
||||
self.dec3 = layers.Decoder(ch * (4 + 8), ch * 4, 3, 1, 1)
|
||||
self.dec2 = layers.Decoder(ch * (2 + 4), ch * 2, 3, 1, 1)
|
||||
self.dec1 = layers.Decoder(ch * (1 + 2), ch, 3, 1, 1)
|
||||
|
||||
def __call__(self, x):
|
||||
h, e1 = self.enc1(x)
|
||||
h, e2 = self.enc2(h)
|
||||
h, e3 = self.enc3(h)
|
||||
h, e4 = self.enc4(h)
|
||||
|
||||
h = self.aspp(h)
|
||||
|
||||
h = self.dec4(h, e4)
|
||||
h = self.dec3(h, e3)
|
||||
h = self.dec2(h, e2)
|
||||
h = self.dec1(h, e1)
|
||||
|
||||
return h
|
||||
|
||||
|
||||
class CascadedASPPNet(nn.Module):
|
||||
|
||||
def __init__(self, n_fft):
|
||||
super(CascadedASPPNet, self).__init__()
|
||||
self.stg1_low_band_net = BaseASPPNet(2, 64)
|
||||
self.stg1_high_band_net = BaseASPPNet(2, 64)
|
||||
|
||||
self.stg2_bridge = layers.Conv2DBNActiv(66, 32, 1, 1, 0)
|
||||
self.stg2_full_band_net = BaseASPPNet(32, 64)
|
||||
|
||||
self.stg3_bridge = layers.Conv2DBNActiv(130, 64, 1, 1, 0)
|
||||
self.stg3_full_band_net = BaseASPPNet(64, 128)
|
||||
|
||||
self.out = nn.Conv2d(128, 2, 1, bias=False)
|
||||
self.aux1_out = nn.Conv2d(64, 2, 1, bias=False)
|
||||
self.aux2_out = nn.Conv2d(64, 2, 1, bias=False)
|
||||
|
||||
self.max_bin = n_fft // 2
|
||||
self.output_bin = n_fft // 2 + 1
|
||||
|
||||
self.offset = 128
|
||||
|
||||
def forward(self, x, aggressiveness=None):
|
||||
mix = x.detach()
|
||||
x = x.clone()
|
||||
|
||||
x = x[:, :, :self.max_bin]
|
||||
|
||||
bandw = x.size()[2] // 2
|
||||
aux1 = torch.cat([
|
||||
self.stg1_low_band_net(x[:, :, :bandw]),
|
||||
self.stg1_high_band_net(x[:, :, bandw:])
|
||||
], dim=2)
|
||||
|
||||
h = torch.cat([x, aux1], dim=1)
|
||||
aux2 = self.stg2_full_band_net(self.stg2_bridge(h))
|
||||
|
||||
h = torch.cat([x, aux1, aux2], dim=1)
|
||||
h = self.stg3_full_band_net(self.stg3_bridge(h))
|
||||
|
||||
mask = torch.sigmoid(self.out(h))
|
||||
mask = F.pad(
|
||||
input=mask,
|
||||
pad=(0, 0, 0, self.output_bin - mask.size()[2]),
|
||||
mode='replicate')
|
||||
|
||||
if self.training:
|
||||
aux1 = torch.sigmoid(self.aux1_out(aux1))
|
||||
aux1 = F.pad(
|
||||
input=aux1,
|
||||
pad=(0, 0, 0, self.output_bin - aux1.size()[2]),
|
||||
mode='replicate')
|
||||
aux2 = torch.sigmoid(self.aux2_out(aux2))
|
||||
aux2 = F.pad(
|
||||
input=aux2,
|
||||
pad=(0, 0, 0, self.output_bin - aux2.size()[2]),
|
||||
mode='replicate')
|
||||
return mask * mix, aux1 * mix, aux2 * mix
|
||||
else:
|
||||
if aggressiveness:
|
||||
mask[:, :, :aggressiveness['split_bin']] = torch.pow(mask[:, :, :aggressiveness['split_bin']], 1 + aggressiveness['value'] / 3)
|
||||
mask[:, :, aggressiveness['split_bin']:] = torch.pow(mask[:, :, aggressiveness['split_bin']:], 1 + aggressiveness['value'])
|
||||
|
||||
return mask * mix
|
||||
|
||||
def predict(self, x_mag, aggressiveness=None):
|
||||
h = self.forward(x_mag, aggressiveness)
|
||||
|
||||
if self.offset > 0:
|
||||
h = h[:, :, :, self.offset:-self.offset]
|
||||
assert h.size()[3] > 0
|
||||
|
||||
return h
|
||||
@ -1,113 +0,0 @@
|
||||
import torch
|
||||
import numpy as np
|
||||
from torch import nn
|
||||
import torch.nn.functional as F
|
||||
|
||||
from lib_v5 import layers_537238KB as layers
|
||||
|
||||
|
||||
class BaseASPPNet(nn.Module):
|
||||
|
||||
def __init__(self, nin, ch, dilations=(4, 8, 16)):
|
||||
super(BaseASPPNet, self).__init__()
|
||||
self.enc1 = layers.Encoder(nin, ch, 3, 2, 1)
|
||||
self.enc2 = layers.Encoder(ch, ch * 2, 3, 2, 1)
|
||||
self.enc3 = layers.Encoder(ch * 2, ch * 4, 3, 2, 1)
|
||||
self.enc4 = layers.Encoder(ch * 4, ch * 8, 3, 2, 1)
|
||||
|
||||
self.aspp = layers.ASPPModule(ch * 8, ch * 16, dilations)
|
||||
|
||||
self.dec4 = layers.Decoder(ch * (8 + 16), ch * 8, 3, 1, 1)
|
||||
self.dec3 = layers.Decoder(ch * (4 + 8), ch * 4, 3, 1, 1)
|
||||
self.dec2 = layers.Decoder(ch * (2 + 4), ch * 2, 3, 1, 1)
|
||||
self.dec1 = layers.Decoder(ch * (1 + 2), ch, 3, 1, 1)
|
||||
|
||||
def __call__(self, x):
|
||||
h, e1 = self.enc1(x)
|
||||
h, e2 = self.enc2(h)
|
||||
h, e3 = self.enc3(h)
|
||||
h, e4 = self.enc4(h)
|
||||
|
||||
h = self.aspp(h)
|
||||
|
||||
h = self.dec4(h, e4)
|
||||
h = self.dec3(h, e3)
|
||||
h = self.dec2(h, e2)
|
||||
h = self.dec1(h, e1)
|
||||
|
||||
return h
|
||||
|
||||
|
||||
class CascadedASPPNet(nn.Module):
|
||||
|
||||
def __init__(self, n_fft):
|
||||
super(CascadedASPPNet, self).__init__()
|
||||
self.stg1_low_band_net = BaseASPPNet(2, 64)
|
||||
self.stg1_high_band_net = BaseASPPNet(2, 64)
|
||||
|
||||
self.stg2_bridge = layers.Conv2DBNActiv(66, 32, 1, 1, 0)
|
||||
self.stg2_full_band_net = BaseASPPNet(32, 64)
|
||||
|
||||
self.stg3_bridge = layers.Conv2DBNActiv(130, 64, 1, 1, 0)
|
||||
self.stg3_full_band_net = BaseASPPNet(64, 128)
|
||||
|
||||
self.out = nn.Conv2d(128, 2, 1, bias=False)
|
||||
self.aux1_out = nn.Conv2d(64, 2, 1, bias=False)
|
||||
self.aux2_out = nn.Conv2d(64, 2, 1, bias=False)
|
||||
|
||||
self.max_bin = n_fft // 2
|
||||
self.output_bin = n_fft // 2 + 1
|
||||
|
||||
self.offset = 128
|
||||
|
||||
def forward(self, x, aggressiveness=None):
|
||||
mix = x.detach()
|
||||
x = x.clone()
|
||||
|
||||
x = x[:, :, :self.max_bin]
|
||||
|
||||
bandw = x.size()[2] // 2
|
||||
aux1 = torch.cat([
|
||||
self.stg1_low_band_net(x[:, :, :bandw]),
|
||||
self.stg1_high_band_net(x[:, :, bandw:])
|
||||
], dim=2)
|
||||
|
||||
h = torch.cat([x, aux1], dim=1)
|
||||
aux2 = self.stg2_full_band_net(self.stg2_bridge(h))
|
||||
|
||||
h = torch.cat([x, aux1, aux2], dim=1)
|
||||
h = self.stg3_full_band_net(self.stg3_bridge(h))
|
||||
|
||||
mask = torch.sigmoid(self.out(h))
|
||||
mask = F.pad(
|
||||
input=mask,
|
||||
pad=(0, 0, 0, self.output_bin - mask.size()[2]),
|
||||
mode='replicate')
|
||||
|
||||
if self.training:
|
||||
aux1 = torch.sigmoid(self.aux1_out(aux1))
|
||||
aux1 = F.pad(
|
||||
input=aux1,
|
||||
pad=(0, 0, 0, self.output_bin - aux1.size()[2]),
|
||||
mode='replicate')
|
||||
aux2 = torch.sigmoid(self.aux2_out(aux2))
|
||||
aux2 = F.pad(
|
||||
input=aux2,
|
||||
pad=(0, 0, 0, self.output_bin - aux2.size()[2]),
|
||||
mode='replicate')
|
||||
return mask * mix, aux1 * mix, aux2 * mix
|
||||
else:
|
||||
if aggressiveness:
|
||||
mask[:, :, :aggressiveness['split_bin']] = torch.pow(mask[:, :, :aggressiveness['split_bin']], 1 + aggressiveness['value'] / 3)
|
||||
mask[:, :, aggressiveness['split_bin']:] = torch.pow(mask[:, :, aggressiveness['split_bin']:], 1 + aggressiveness['value'])
|
||||
|
||||
return mask * mix
|
||||
|
||||
def predict(self, x_mag, aggressiveness=None):
|
||||
h = self.forward(x_mag, aggressiveness)
|
||||
|
||||
if self.offset > 0:
|
||||
h = h[:, :, :, self.offset:-self.offset]
|
||||
assert h.size()[3] > 0
|
||||
|
||||
return h
|
||||
File diff suppressed because one or more lines are too long
File diff suppressed because one or more lines are too long
File diff suppressed because one or more lines are too long
@ -1 +0,0 @@
|
||||
Sox goes here
|
||||
File diff suppressed because one or more lines are too long
@ -1,549 +0,0 @@
|
||||
import os
|
||||
|
||||
import librosa
|
||||
import numpy as np
|
||||
import soundfile as sf
|
||||
import math
|
||||
import json
|
||||
import hashlib
|
||||
from tqdm import tqdm
|
||||
|
||||
|
||||
def crop_center(h1, h2):
|
||||
h1_shape = h1.size()
|
||||
h2_shape = h2.size()
|
||||
|
||||
if h1_shape[3] == h2_shape[3]:
|
||||
return h1
|
||||
elif h1_shape[3] < h2_shape[3]:
|
||||
raise ValueError('h1_shape[3] must be greater than h2_shape[3]')
|
||||
|
||||
# s_freq = (h2_shape[2] - h1_shape[2]) // 2
|
||||
# e_freq = s_freq + h1_shape[2]
|
||||
s_time = (h1_shape[3] - h2_shape[3]) // 2
|
||||
e_time = s_time + h2_shape[3]
|
||||
h1 = h1[:, :, :, s_time:e_time]
|
||||
|
||||
return h1
|
||||
|
||||
|
||||
def wave_to_spectrogram(wave, hop_length, n_fft, mid_side=False, mid_side_b2=False, reverse=False):
|
||||
if reverse:
|
||||
wave_left = np.flip(np.asfortranarray(wave[0]))
|
||||
wave_right = np.flip(np.asfortranarray(wave[1]))
|
||||
elif mid_side:
|
||||
wave_left = np.asfortranarray(np.add(wave[0], wave[1]) / 2)
|
||||
wave_right = np.asfortranarray(np.subtract(wave[0], wave[1]))
|
||||
elif mid_side_b2:
|
||||
wave_left = np.asfortranarray(np.add(wave[1], wave[0] * .5))
|
||||
wave_right = np.asfortranarray(np.subtract(wave[0], wave[1] * .5))
|
||||
else:
|
||||
wave_left = np.asfortranarray(wave[0])
|
||||
wave_right = np.asfortranarray(wave[1])
|
||||
|
||||
spec_left = librosa.stft(wave_left, n_fft, hop_length=hop_length)
|
||||
spec_right = librosa.stft(wave_right, n_fft, hop_length=hop_length)
|
||||
|
||||
spec = np.asfortranarray([spec_left, spec_right])
|
||||
|
||||
return spec
|
||||
|
||||
|
||||
def wave_to_spectrogram_mt(wave, hop_length, n_fft, mid_side=False, mid_side_b2=False, reverse=False):
|
||||
import threading
|
||||
|
||||
if reverse:
|
||||
wave_left = np.flip(np.asfortranarray(wave[0]))
|
||||
wave_right = np.flip(np.asfortranarray(wave[1]))
|
||||
elif mid_side:
|
||||
wave_left = np.asfortranarray(np.add(wave[0], wave[1]) / 2)
|
||||
wave_right = np.asfortranarray(np.subtract(wave[0], wave[1]))
|
||||
elif mid_side_b2:
|
||||
wave_left = np.asfortranarray(np.add(wave[1], wave[0] * .5))
|
||||
wave_right = np.asfortranarray(np.subtract(wave[0], wave[1] * .5))
|
||||
else:
|
||||
wave_left = np.asfortranarray(wave[0])
|
||||
wave_right = np.asfortranarray(wave[1])
|
||||
|
||||
def run_thread(**kwargs):
|
||||
global spec_left
|
||||
spec_left = librosa.stft(**kwargs)
|
||||
|
||||
thread = threading.Thread(target=run_thread, kwargs={'y': wave_left, 'n_fft': n_fft, 'hop_length': hop_length})
|
||||
thread.start()
|
||||
spec_right = librosa.stft(wave_right, n_fft, hop_length=hop_length)
|
||||
thread.join()
|
||||
|
||||
spec = np.asfortranarray([spec_left, spec_right])
|
||||
|
||||
return spec
|
||||
|
||||
def normalize(wave_res):
|
||||
"""Save output music files"""
|
||||
maxv = np.abs(wave_res).max()
|
||||
if maxv > 1.0:
|
||||
print(f"\nNormalization Set On: Input above threshold for clipping. The result was normalized. Max:{maxv}\n")
|
||||
wave_res /= maxv
|
||||
else:
|
||||
print(f"\nNormalization Set On: Input not above threshold for clipping. Max:{maxv}\n")
|
||||
|
||||
return wave_res
|
||||
|
||||
def nonormalize(wave_res):
|
||||
"""Save output music files"""
|
||||
maxv = np.abs(wave_res).max()
|
||||
if maxv > 1.0:
|
||||
print(f"\nNormalization Set Off: Input above threshold for clipping. The result was not normalized. Max:{maxv}\n")
|
||||
else:
|
||||
print(f"\nNormalization Set Off: Input not above threshold for clipping. Max:{maxv}\n")
|
||||
|
||||
return wave_res
|
||||
|
||||
def combine_spectrograms(specs, mp):
|
||||
l = min([specs[i].shape[2] for i in specs])
|
||||
spec_c = np.zeros(shape=(2, mp.param['bins'] + 1, l), dtype=np.complex64)
|
||||
offset = 0
|
||||
bands_n = len(mp.param['band'])
|
||||
|
||||
for d in range(1, bands_n + 1):
|
||||
h = mp.param['band'][d]['crop_stop'] - mp.param['band'][d]['crop_start']
|
||||
spec_c[:, offset:offset+h, :l] = specs[d][:, mp.param['band'][d]['crop_start']:mp.param['band'][d]['crop_stop'], :l]
|
||||
offset += h
|
||||
|
||||
if offset > mp.param['bins']:
|
||||
raise ValueError('Too much bins')
|
||||
|
||||
# lowpass fiter
|
||||
if mp.param['pre_filter_start'] > 0: # and mp.param['band'][bands_n]['res_type'] in ['scipy', 'polyphase']:
|
||||
if bands_n == 1:
|
||||
spec_c = fft_lp_filter(spec_c, mp.param['pre_filter_start'], mp.param['pre_filter_stop'])
|
||||
else:
|
||||
gp = 1
|
||||
for b in range(mp.param['pre_filter_start'] + 1, mp.param['pre_filter_stop']):
|
||||
g = math.pow(10, -(b - mp.param['pre_filter_start']) * (3.5 - gp) / 20.0)
|
||||
gp = g
|
||||
spec_c[:, b, :] *= g
|
||||
|
||||
return np.asfortranarray(spec_c)
|
||||
|
||||
|
||||
def spectrogram_to_image(spec, mode='magnitude'):
|
||||
if mode == 'magnitude':
|
||||
if np.iscomplexobj(spec):
|
||||
y = np.abs(spec)
|
||||
else:
|
||||
y = spec
|
||||
y = np.log10(y ** 2 + 1e-8)
|
||||
elif mode == 'phase':
|
||||
if np.iscomplexobj(spec):
|
||||
y = np.angle(spec)
|
||||
else:
|
||||
y = spec
|
||||
|
||||
y -= y.min()
|
||||
y *= 255 / y.max()
|
||||
img = np.uint8(y)
|
||||
|
||||
if y.ndim == 3:
|
||||
img = img.transpose(1, 2, 0)
|
||||
img = np.concatenate([
|
||||
np.max(img, axis=2, keepdims=True), img
|
||||
], axis=2)
|
||||
|
||||
return img
|
||||
|
||||
|
||||
def reduce_vocal_aggressively(X, y, softmask):
|
||||
v = X - y
|
||||
y_mag_tmp = np.abs(y)
|
||||
v_mag_tmp = np.abs(v)
|
||||
|
||||
v_mask = v_mag_tmp > y_mag_tmp
|
||||
y_mag = np.clip(y_mag_tmp - v_mag_tmp * v_mask * softmask, 0, np.inf)
|
||||
|
||||
return y_mag * np.exp(1.j * np.angle(y))
|
||||
|
||||
|
||||
def mask_silence(mag, ref, thres=0.2, min_range=64, fade_size=32):
|
||||
if min_range < fade_size * 2:
|
||||
raise ValueError('min_range must be >= fade_area * 2')
|
||||
|
||||
mag = mag.copy()
|
||||
|
||||
idx = np.where(ref.mean(axis=(0, 1)) < thres)[0]
|
||||
starts = np.insert(idx[np.where(np.diff(idx) != 1)[0] + 1], 0, idx[0])
|
||||
ends = np.append(idx[np.where(np.diff(idx) != 1)[0]], idx[-1])
|
||||
uninformative = np.where(ends - starts > min_range)[0]
|
||||
if len(uninformative) > 0:
|
||||
starts = starts[uninformative]
|
||||
ends = ends[uninformative]
|
||||
old_e = None
|
||||
for s, e in zip(starts, ends):
|
||||
if old_e is not None and s - old_e < fade_size:
|
||||
s = old_e - fade_size * 2
|
||||
|
||||
if s != 0:
|
||||
weight = np.linspace(0, 1, fade_size)
|
||||
mag[:, :, s:s + fade_size] += weight * ref[:, :, s:s + fade_size]
|
||||
else:
|
||||
s -= fade_size
|
||||
|
||||
if e != mag.shape[2]:
|
||||
weight = np.linspace(1, 0, fade_size)
|
||||
mag[:, :, e - fade_size:e] += weight * ref[:, :, e - fade_size:e]
|
||||
else:
|
||||
e += fade_size
|
||||
|
||||
mag[:, :, s + fade_size:e - fade_size] += ref[:, :, s + fade_size:e - fade_size]
|
||||
old_e = e
|
||||
|
||||
return mag
|
||||
|
||||
|
||||
def align_wave_head_and_tail(a, b):
|
||||
l = min([a[0].size, b[0].size])
|
||||
|
||||
return a[:l,:l], b[:l,:l]
|
||||
|
||||
|
||||
def cache_or_load(mix_path, inst_path, mp):
|
||||
mix_basename = os.path.splitext(os.path.basename(mix_path))[0]
|
||||
inst_basename = os.path.splitext(os.path.basename(inst_path))[0]
|
||||
|
||||
cache_dir = 'mph{}'.format(hashlib.sha1(json.dumps(mp.param, sort_keys=True).encode('utf-8')).hexdigest())
|
||||
mix_cache_dir = os.path.join('cache', cache_dir)
|
||||
inst_cache_dir = os.path.join('cache', cache_dir)
|
||||
|
||||
os.makedirs(mix_cache_dir, exist_ok=True)
|
||||
os.makedirs(inst_cache_dir, exist_ok=True)
|
||||
|
||||
mix_cache_path = os.path.join(mix_cache_dir, mix_basename + '.npy')
|
||||
inst_cache_path = os.path.join(inst_cache_dir, inst_basename + '.npy')
|
||||
|
||||
if os.path.exists(mix_cache_path) and os.path.exists(inst_cache_path):
|
||||
X_spec_m = np.load(mix_cache_path)
|
||||
y_spec_m = np.load(inst_cache_path)
|
||||
else:
|
||||
X_wave, y_wave, X_spec_s, y_spec_s = {}, {}, {}, {}
|
||||
|
||||
for d in range(len(mp.param['band']), 0, -1):
|
||||
bp = mp.param['band'][d]
|
||||
|
||||
if d == len(mp.param['band']): # high-end band
|
||||
X_wave[d], _ = librosa.load(
|
||||
mix_path, bp['sr'], False, dtype=np.float32, res_type=bp['res_type'])
|
||||
y_wave[d], _ = librosa.load(
|
||||
inst_path, bp['sr'], False, dtype=np.float32, res_type=bp['res_type'])
|
||||
else: # lower bands
|
||||
X_wave[d] = librosa.resample(X_wave[d+1], mp.param['band'][d+1]['sr'], bp['sr'], res_type=bp['res_type'])
|
||||
y_wave[d] = librosa.resample(y_wave[d+1], mp.param['band'][d+1]['sr'], bp['sr'], res_type=bp['res_type'])
|
||||
|
||||
X_wave[d], y_wave[d] = align_wave_head_and_tail(X_wave[d], y_wave[d])
|
||||
|
||||
X_spec_s[d] = wave_to_spectrogram(X_wave[d], bp['hl'], bp['n_fft'], mp.param['mid_side'], mp.param['mid_side_b2'], mp.param['reverse'])
|
||||
y_spec_s[d] = wave_to_spectrogram(y_wave[d], bp['hl'], bp['n_fft'], mp.param['mid_side'], mp.param['mid_side_b2'], mp.param['reverse'])
|
||||
|
||||
del X_wave, y_wave
|
||||
|
||||
X_spec_m = combine_spectrograms(X_spec_s, mp)
|
||||
y_spec_m = combine_spectrograms(y_spec_s, mp)
|
||||
|
||||
if X_spec_m.shape != y_spec_m.shape:
|
||||
raise ValueError('The combined spectrograms are different: ' + mix_path)
|
||||
|
||||
_, ext = os.path.splitext(mix_path)
|
||||
|
||||
np.save(mix_cache_path, X_spec_m)
|
||||
np.save(inst_cache_path, y_spec_m)
|
||||
|
||||
return X_spec_m, y_spec_m
|
||||
|
||||
|
||||
def spectrogram_to_wave(spec, hop_length, mid_side, mid_side_b2, reverse, clamp=False):
|
||||
spec_left = np.asfortranarray(spec[0])
|
||||
spec_right = np.asfortranarray(spec[1])
|
||||
|
||||
wave_left = librosa.istft(spec_left, hop_length=hop_length)
|
||||
wave_right = librosa.istft(spec_right, hop_length=hop_length)
|
||||
|
||||
if reverse:
|
||||
return np.asfortranarray([np.flip(wave_left), np.flip(wave_right)])
|
||||
elif mid_side:
|
||||
return np.asfortranarray([np.add(wave_left, wave_right / 2), np.subtract(wave_left, wave_right / 2)])
|
||||
elif mid_side_b2:
|
||||
return np.asfortranarray([np.add(wave_right / 1.25, .4 * wave_left), np.subtract(wave_left / 1.25, .4 * wave_right)])
|
||||
else:
|
||||
return np.asfortranarray([wave_left, wave_right])
|
||||
|
||||
|
||||
def spectrogram_to_wave_mt(spec, hop_length, mid_side, reverse, mid_side_b2):
|
||||
import threading
|
||||
|
||||
spec_left = np.asfortranarray(spec[0])
|
||||
spec_right = np.asfortranarray(spec[1])
|
||||
|
||||
def run_thread(**kwargs):
|
||||
global wave_left
|
||||
wave_left = librosa.istft(**kwargs)
|
||||
|
||||
thread = threading.Thread(target=run_thread, kwargs={'stft_matrix': spec_left, 'hop_length': hop_length})
|
||||
thread.start()
|
||||
wave_right = librosa.istft(spec_right, hop_length=hop_length)
|
||||
thread.join()
|
||||
|
||||
if reverse:
|
||||
return np.asfortranarray([np.flip(wave_left), np.flip(wave_right)])
|
||||
elif mid_side:
|
||||
return np.asfortranarray([np.add(wave_left, wave_right / 2), np.subtract(wave_left, wave_right / 2)])
|
||||
elif mid_side_b2:
|
||||
return np.asfortranarray([np.add(wave_right / 1.25, .4 * wave_left), np.subtract(wave_left / 1.25, .4 * wave_right)])
|
||||
else:
|
||||
return np.asfortranarray([wave_left, wave_right])
|
||||
|
||||
|
||||
def cmb_spectrogram_to_wave(spec_m, mp, extra_bins_h=None, extra_bins=None):
|
||||
wave_band = {}
|
||||
bands_n = len(mp.param['band'])
|
||||
offset = 0
|
||||
|
||||
for d in range(1, bands_n + 1):
|
||||
bp = mp.param['band'][d]
|
||||
spec_s = np.ndarray(shape=(2, bp['n_fft'] // 2 + 1, spec_m.shape[2]), dtype=complex)
|
||||
h = bp['crop_stop'] - bp['crop_start']
|
||||
spec_s[:, bp['crop_start']:bp['crop_stop'], :] = spec_m[:, offset:offset+h, :]
|
||||
|
||||
offset += h
|
||||
if d == bands_n: # higher
|
||||
if extra_bins_h: # if --high_end_process bypass
|
||||
max_bin = bp['n_fft'] // 2
|
||||
spec_s[:, max_bin-extra_bins_h:max_bin, :] = extra_bins[:, :extra_bins_h, :]
|
||||
if bp['hpf_start'] > 0:
|
||||
spec_s = fft_hp_filter(spec_s, bp['hpf_start'], bp['hpf_stop'] - 1)
|
||||
if bands_n == 1:
|
||||
wave = spectrogram_to_wave(spec_s, bp['hl'], mp.param['mid_side'], mp.param['mid_side_b2'], mp.param['reverse'])
|
||||
else:
|
||||
wave = np.add(wave, spectrogram_to_wave(spec_s, bp['hl'], mp.param['mid_side'], mp.param['mid_side_b2'], mp.param['reverse']))
|
||||
else:
|
||||
sr = mp.param['band'][d+1]['sr']
|
||||
if d == 1: # lower
|
||||
spec_s = fft_lp_filter(spec_s, bp['lpf_start'], bp['lpf_stop'])
|
||||
wave = librosa.resample(spectrogram_to_wave(spec_s, bp['hl'], mp.param['mid_side'], mp.param['mid_side_b2'], mp.param['reverse']), bp['sr'], sr, res_type="sinc_fastest")
|
||||
else: # mid
|
||||
spec_s = fft_hp_filter(spec_s, bp['hpf_start'], bp['hpf_stop'] - 1)
|
||||
spec_s = fft_lp_filter(spec_s, bp['lpf_start'], bp['lpf_stop'])
|
||||
wave2 = np.add(wave, spectrogram_to_wave(spec_s, bp['hl'], mp.param['mid_side'], mp.param['mid_side_b2'], mp.param['reverse']))
|
||||
wave = librosa.resample(wave2, bp['sr'], sr, res_type="sinc_fastest")
|
||||
|
||||
return wave.T
|
||||
|
||||
def cmb_spectrogram_to_wave_d(spec_m, mp, extra_bins_h=None, extra_bins=None, demucs=True):
|
||||
wave_band = {}
|
||||
bands_n = len(mp.param['band'])
|
||||
offset = 0
|
||||
|
||||
for d in range(1, bands_n + 1):
|
||||
bp = mp.param['band'][d]
|
||||
spec_s = np.ndarray(shape=(2, bp['n_fft'] // 2 + 1, spec_m.shape[2]), dtype=complex)
|
||||
h = bp['crop_stop'] - bp['crop_start']
|
||||
spec_s[:, bp['crop_start']:bp['crop_stop'], :] = spec_m[:, offset:offset+h, :]
|
||||
|
||||
offset += h
|
||||
if d == bands_n: # higher
|
||||
if extra_bins_h: # if --high_end_process bypass
|
||||
max_bin = bp['n_fft'] // 2
|
||||
spec_s[:, max_bin-extra_bins_h:max_bin, :] = extra_bins[:, :extra_bins_h, :]
|
||||
if bp['hpf_start'] > 0:
|
||||
spec_s = fft_hp_filter(spec_s, bp['hpf_start'], bp['hpf_stop'] - 1)
|
||||
if bands_n == 1:
|
||||
wave = spectrogram_to_wave(spec_s, bp['hl'], mp.param['mid_side'], mp.param['mid_side_b2'], mp.param['reverse'])
|
||||
else:
|
||||
wave = np.add(wave, spectrogram_to_wave(spec_s, bp['hl'], mp.param['mid_side'], mp.param['mid_side_b2'], mp.param['reverse']))
|
||||
else:
|
||||
sr = mp.param['band'][d+1]['sr']
|
||||
if d == 1: # lower
|
||||
spec_s = fft_lp_filter(spec_s, bp['lpf_start'], bp['lpf_stop'])
|
||||
wave = librosa.resample(spectrogram_to_wave(spec_s, bp['hl'], mp.param['mid_side'], mp.param['mid_side_b2'], mp.param['reverse']), bp['sr'], sr, res_type="sinc_fastest")
|
||||
else: # mid
|
||||
spec_s = fft_hp_filter(spec_s, bp['hpf_start'], bp['hpf_stop'] - 1)
|
||||
spec_s = fft_lp_filter(spec_s, bp['lpf_start'], bp['lpf_stop'])
|
||||
wave2 = np.add(wave, spectrogram_to_wave(spec_s, bp['hl'], mp.param['mid_side'], mp.param['mid_side_b2'], mp.param['reverse']))
|
||||
wave = librosa.resample(wave2, bp['sr'], sr, res_type="sinc_fastest")
|
||||
|
||||
#print(demucs)
|
||||
|
||||
if demucs == True:
|
||||
wave = librosa.resample(wave, bp['sr'], 44100, res_type="sinc_fastest")
|
||||
return wave
|
||||
else:
|
||||
return wave
|
||||
|
||||
def fft_lp_filter(spec, bin_start, bin_stop):
|
||||
g = 1.0
|
||||
for b in range(bin_start, bin_stop):
|
||||
g -= 1 / (bin_stop - bin_start)
|
||||
spec[:, b, :] = g * spec[:, b, :]
|
||||
|
||||
spec[:, bin_stop:, :] *= 0
|
||||
|
||||
return spec
|
||||
|
||||
|
||||
def fft_hp_filter(spec, bin_start, bin_stop):
|
||||
g = 1.0
|
||||
for b in range(bin_start, bin_stop, -1):
|
||||
g -= 1 / (bin_start - bin_stop)
|
||||
spec[:, b, :] = g * spec[:, b, :]
|
||||
|
||||
spec[:, 0:bin_stop+1, :] *= 0
|
||||
|
||||
return spec
|
||||
|
||||
|
||||
def mirroring(a, spec_m, input_high_end, mp):
|
||||
if 'mirroring' == a:
|
||||
mirror = np.flip(np.abs(spec_m[:, mp.param['pre_filter_start']-10-input_high_end.shape[1]:mp.param['pre_filter_start']-10, :]), 1)
|
||||
mirror = mirror * np.exp(1.j * np.angle(input_high_end))
|
||||
|
||||
return np.where(np.abs(input_high_end) <= np.abs(mirror), input_high_end, mirror)
|
||||
|
||||
if 'mirroring2' == a:
|
||||
mirror = np.flip(np.abs(spec_m[:, mp.param['pre_filter_start']-10-input_high_end.shape[1]:mp.param['pre_filter_start']-10, :]), 1)
|
||||
mi = np.multiply(mirror, input_high_end * 1.7)
|
||||
|
||||
return np.where(np.abs(input_high_end) <= np.abs(mi), input_high_end, mi)
|
||||
|
||||
|
||||
def ensembling(a, specs):
|
||||
for i in range(1, len(specs)):
|
||||
if i == 1:
|
||||
spec = specs[0]
|
||||
|
||||
ln = min([spec.shape[2], specs[i].shape[2]])
|
||||
spec = spec[:,:,:ln]
|
||||
specs[i] = specs[i][:,:,:ln]
|
||||
|
||||
if 'min_mag' == a:
|
||||
spec = np.where(np.abs(specs[i]) <= np.abs(spec), specs[i], spec)
|
||||
if 'max_mag' == a:
|
||||
spec = np.where(np.abs(specs[i]) >= np.abs(spec), specs[i], spec)
|
||||
|
||||
return spec
|
||||
|
||||
def stft(wave, nfft, hl):
|
||||
wave_left = np.asfortranarray(wave[0])
|
||||
wave_right = np.asfortranarray(wave[1])
|
||||
spec_left = librosa.stft(wave_left, nfft, hop_length=hl)
|
||||
spec_right = librosa.stft(wave_right, nfft, hop_length=hl)
|
||||
spec = np.asfortranarray([spec_left, spec_right])
|
||||
|
||||
return spec
|
||||
|
||||
def istft(spec, hl):
|
||||
spec_left = np.asfortranarray(spec[0])
|
||||
spec_right = np.asfortranarray(spec[1])
|
||||
|
||||
wave_left = librosa.istft(spec_left, hop_length=hl)
|
||||
wave_right = librosa.istft(spec_right, hop_length=hl)
|
||||
wave = np.asfortranarray([wave_left, wave_right])
|
||||
|
||||
|
||||
if __name__ == "__main__":
|
||||
import cv2
|
||||
import sys
|
||||
import time
|
||||
import argparse
|
||||
from model_param_init import ModelParameters
|
||||
|
||||
p = argparse.ArgumentParser()
|
||||
p.add_argument('--algorithm', '-a', type=str, choices=['invert', 'invert_p', 'min_mag', 'max_mag', 'deep', 'align'], default='min_mag')
|
||||
p.add_argument('--model_params', '-m', type=str, default=os.path.join('modelparams', '1band_sr44100_hl512.json'))
|
||||
p.add_argument('--output_name', '-o', type=str, default='output')
|
||||
p.add_argument('--vocals_only', '-v', action='store_true')
|
||||
p.add_argument('input', nargs='+')
|
||||
args = p.parse_args()
|
||||
|
||||
start_time = time.time()
|
||||
|
||||
if args.algorithm.startswith('invert') and len(args.input) != 2:
|
||||
raise ValueError('There should be two input files.')
|
||||
|
||||
if not args.algorithm.startswith('invert') and len(args.input) < 2:
|
||||
raise ValueError('There must be at least two input files.')
|
||||
|
||||
wave, specs = {}, {}
|
||||
mp = ModelParameters(args.model_params)
|
||||
|
||||
for i in range(len(args.input)):
|
||||
spec = {}
|
||||
|
||||
for d in range(len(mp.param['band']), 0, -1):
|
||||
bp = mp.param['band'][d]
|
||||
|
||||
if d == len(mp.param['band']): # high-end band
|
||||
wave[d], _ = librosa.load(
|
||||
args.input[i], bp['sr'], False, dtype=np.float32, res_type=bp['res_type'])
|
||||
|
||||
if len(wave[d].shape) == 1: # mono to stereo
|
||||
wave[d] = np.array([wave[d], wave[d]])
|
||||
else: # lower bands
|
||||
wave[d] = librosa.resample(wave[d+1], mp.param['band'][d+1]['sr'], bp['sr'], res_type=bp['res_type'])
|
||||
|
||||
spec[d] = wave_to_spectrogram(wave[d], bp['hl'], bp['n_fft'], mp.param['mid_side'], mp.param['mid_side_b2'], mp.param['reverse'])
|
||||
|
||||
specs[i] = combine_spectrograms(spec, mp)
|
||||
|
||||
del wave
|
||||
|
||||
if args.algorithm == 'deep':
|
||||
d_spec = np.where(np.abs(specs[0]) <= np.abs(spec[1]), specs[0], spec[1])
|
||||
v_spec = d_spec - specs[1]
|
||||
sf.write(os.path.join('{}.wav'.format(args.output_name)), cmb_spectrogram_to_wave(v_spec, mp), mp.param['sr'])
|
||||
|
||||
if args.algorithm.startswith('invert'):
|
||||
ln = min([specs[0].shape[2], specs[1].shape[2]])
|
||||
specs[0] = specs[0][:,:,:ln]
|
||||
specs[1] = specs[1][:,:,:ln]
|
||||
|
||||
if 'invert_p' == args.algorithm:
|
||||
X_mag = np.abs(specs[0])
|
||||
y_mag = np.abs(specs[1])
|
||||
max_mag = np.where(X_mag >= y_mag, X_mag, y_mag)
|
||||
v_spec = specs[1] - max_mag * np.exp(1.j * np.angle(specs[0]))
|
||||
else:
|
||||
specs[1] = reduce_vocal_aggressively(specs[0], specs[1], 0.2)
|
||||
v_spec = specs[0] - specs[1]
|
||||
|
||||
if not args.vocals_only:
|
||||
X_mag = np.abs(specs[0])
|
||||
y_mag = np.abs(specs[1])
|
||||
v_mag = np.abs(v_spec)
|
||||
|
||||
X_image = spectrogram_to_image(X_mag)
|
||||
y_image = spectrogram_to_image(y_mag)
|
||||
v_image = spectrogram_to_image(v_mag)
|
||||
|
||||
cv2.imwrite('{}_X.png'.format(args.output_name), X_image)
|
||||
cv2.imwrite('{}_y.png'.format(args.output_name), y_image)
|
||||
cv2.imwrite('{}_v.png'.format(args.output_name), v_image)
|
||||
|
||||
sf.write('{}_X.wav'.format(args.output_name), cmb_spectrogram_to_wave(specs[0], mp), mp.param['sr'])
|
||||
sf.write('{}_y.wav'.format(args.output_name), cmb_spectrogram_to_wave(specs[1], mp), mp.param['sr'])
|
||||
|
||||
sf.write('{}_v.wav'.format(args.output_name), cmb_spectrogram_to_wave(v_spec, mp), mp.param['sr'])
|
||||
else:
|
||||
if not args.algorithm == 'deep':
|
||||
sf.write(os.path.join('ensembled','{}.wav'.format(args.output_name)), cmb_spectrogram_to_wave(ensembling(args.algorithm, specs), mp), mp.param['sr'])
|
||||
|
||||
if args.algorithm == 'align':
|
||||
|
||||
trackalignment = [
|
||||
{
|
||||
'file1':'"{}"'.format(args.input[0]),
|
||||
'file2':'"{}"'.format(args.input[1])
|
||||
}
|
||||
]
|
||||
|
||||
for i,e in tqdm(enumerate(trackalignment), desc="Performing Alignment..."):
|
||||
os.system(f"python lib/align_tracks.py {e['file1']} {e['file2']}")
|
||||
|
||||
#print('Total time: {0:.{1}f}s'.format(time.time() - start_time, 1))
|
||||
@ -1,61 +0,0 @@
|
||||
from pathlib import Path
|
||||
|
||||
inited = False
|
||||
root = None
|
||||
|
||||
|
||||
def init(func):
|
||||
def wrapper(*args, **kwargs):
|
||||
global inited
|
||||
global root
|
||||
|
||||
if not inited:
|
||||
from tkinter import _default_root
|
||||
|
||||
path = (Path(__file__).parent / "sun-valley.tcl").resolve()
|
||||
|
||||
try:
|
||||
_default_root.tk.call("source", str(path))
|
||||
except AttributeError:
|
||||
raise RuntimeError(
|
||||
"can't set theme. "
|
||||
"Tk is not initialized. "
|
||||
"Please first create a tkinter.Tk instance, then set the theme."
|
||||
) from None
|
||||
else:
|
||||
inited = True
|
||||
root = _default_root
|
||||
|
||||
return func(*args, **kwargs)
|
||||
|
||||
return wrapper
|
||||
|
||||
|
||||
@init
|
||||
def set_theme(theme):
|
||||
if theme not in {"dark", "light"}:
|
||||
raise RuntimeError(f"not a valid theme name: {theme}")
|
||||
|
||||
root.tk.call("set_theme", theme)
|
||||
|
||||
|
||||
@init
|
||||
def get_theme():
|
||||
theme = root.tk.call("ttk::style", "theme", "use")
|
||||
|
||||
try:
|
||||
return {"sun-valley-dark": "dark", "sun-valley-light": "light"}[theme]
|
||||
except KeyError:
|
||||
return theme
|
||||
|
||||
|
||||
@init
|
||||
def toggle_theme():
|
||||
if get_theme() == "dark":
|
||||
use_light_theme()
|
||||
else:
|
||||
use_dark_theme()
|
||||
|
||||
|
||||
use_dark_theme = lambda: set_theme("dark")
|
||||
use_light_theme = lambda: set_theme("light")
|
||||
Binary file not shown.
Binary file not shown.
@ -1,46 +0,0 @@
|
||||
source [file join [file dirname [info script]] theme dark.tcl]
|
||||
|
||||
option add *tearOff 0
|
||||
|
||||
proc set_theme {mode} {
|
||||
if {$mode == "dark"} {
|
||||
ttk::style theme use "sun-valley-dark"
|
||||
|
||||
array set colors {
|
||||
-fg "#F6F6F7"
|
||||
-bg "#0e0e0f"
|
||||
-disabledfg "#F6F6F7"
|
||||
-selectfg "#F6F6F7"
|
||||
-selectbg "#003b50"
|
||||
}
|
||||
|
||||
ttk::style configure . \
|
||||
-background $colors(-bg) \
|
||||
-foreground $colors(-fg) \
|
||||
-troughcolor $colors(-bg) \
|
||||
-focuscolor $colors(-selectbg) \
|
||||
-selectbackground $colors(-selectbg) \
|
||||
-selectforeground $colors(-selectfg) \
|
||||
-insertwidth 0 \
|
||||
-insertcolor $colors(-fg) \
|
||||
-fieldbackground $colors(-selectbg) \
|
||||
-font {"Century Gothic" 10} \
|
||||
-borderwidth 0 \
|
||||
-relief flat
|
||||
|
||||
tk_setPalette \
|
||||
background [ttk::style lookup . -background] \
|
||||
foreground [ttk::style lookup . -foreground] \
|
||||
highlightColor [ttk::style lookup . -focuscolor] \
|
||||
selectBackground [ttk::style lookup . -selectbackground] \
|
||||
selectForeground [ttk::style lookup . -selectforeground] \
|
||||
activeBackground [ttk::style lookup . -selectbackground] \
|
||||
activeForeground [ttk::style lookup . -selectforeground]
|
||||
|
||||
ttk::style map . -foreground [list disabled $colors(-disabledfg)]
|
||||
|
||||
option add *font [ttk::style lookup . -font]
|
||||
option add *Menu.selectcolor $colors(-fg)
|
||||
option add *Menu.background #0e0e0f
|
||||
}
|
||||
}
|
||||
@ -1,534 +0,0 @@
|
||||
# Copyright © 2021 rdbende <rdbende@gmail.com>
|
||||
|
||||
# A stunning dark theme for ttk based on Microsoft's Sun Valley visual style
|
||||
|
||||
package require Tk 8.6
|
||||
|
||||
namespace eval ttk::theme::sun-valley-dark {
|
||||
variable version 1.0
|
||||
package provide ttk::theme::sun-valley-dark $version
|
||||
|
||||
ttk::style theme create sun-valley-dark -parent clam -settings {
|
||||
proc load_images {imgdir} {
|
||||
variable images
|
||||
foreach file [glob -directory $imgdir *.png] {
|
||||
set images([file tail [file rootname $file]]) \
|
||||
[image create photo -file $file -format png]
|
||||
}
|
||||
}
|
||||
|
||||
load_images [file join [file dirname [info script]] dark]
|
||||
|
||||
array set colors {
|
||||
-fg "#F6F6F7"
|
||||
-bg "#0e0e0f"
|
||||
-disabledfg "#F6F6F7"
|
||||
-selectfg "#ffffff"
|
||||
-selectbg "#2f60d8"
|
||||
}
|
||||
|
||||
ttk::style layout TButton {
|
||||
Button.button -children {
|
||||
Button.padding -children {
|
||||
Button.label -side left -expand 1
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
ttk::style layout Toolbutton {
|
||||
Toolbutton.button -children {
|
||||
Toolbutton.padding -children {
|
||||
Toolbutton.label -side left -expand 1
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
ttk::style layout TMenubutton {
|
||||
Menubutton.button -children {
|
||||
Menubutton.padding -children {
|
||||
Menubutton.label -side left -expand 1
|
||||
Menubutton.indicator -side right -sticky nsew
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
ttk::style layout TOptionMenu {
|
||||
OptionMenu.button -children {
|
||||
OptionMenu.padding -children {
|
||||
OptionMenu.label -side left -expand 0
|
||||
OptionMenu.indicator -side right -sticky nsew
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
ttk::style layout Accent.TButton {
|
||||
AccentButton.button -children {
|
||||
AccentButton.padding -children {
|
||||
AccentButton.label -side left -expand 1
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
ttk::style layout Titlebar.TButton {
|
||||
TitlebarButton.button -children {
|
||||
TitlebarButton.padding -children {
|
||||
TitlebarButton.label -side left -expand 1
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
ttk::style layout Close.Titlebar.TButton {
|
||||
CloseButton.button -children {
|
||||
CloseButton.padding -children {
|
||||
CloseButton.label -side left -expand 1
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
ttk::style layout TCheckbutton {
|
||||
Checkbutton.button -children {
|
||||
Checkbutton.padding -children {
|
||||
Checkbutton.indicator -side left
|
||||
Checkbutton.label -side right -expand 1
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
ttk::style layout Switch.TCheckbutton {
|
||||
Switch.button -children {
|
||||
Switch.padding -children {
|
||||
Switch.indicator -side left
|
||||
Switch.label -side right -expand 1
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
ttk::style layout Toggle.TButton {
|
||||
ToggleButton.button -children {
|
||||
ToggleButton.padding -children {
|
||||
ToggleButton.label -side left -expand 1
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
ttk::style layout TRadiobutton {
|
||||
Radiobutton.button -children {
|
||||
Radiobutton.padding -children {
|
||||
Radiobutton.indicator -side left
|
||||
Radiobutton.label -side right -expand 1
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
ttk::style layout Vertical.TScrollbar {
|
||||
Vertical.Scrollbar.trough -sticky ns -children {
|
||||
Vertical.Scrollbar.uparrow -side top
|
||||
Vertical.Scrollbar.downarrow -side bottom
|
||||
Vertical.Scrollbar.thumb -expand 1
|
||||
}
|
||||
}
|
||||
|
||||
ttk::style layout Horizontal.TScrollbar {
|
||||
Horizontal.Scrollbar.trough -sticky ew -children {
|
||||
Horizontal.Scrollbar.leftarrow -side left
|
||||
Horizontal.Scrollbar.rightarrow -side right
|
||||
Horizontal.Scrollbar.thumb -expand 1
|
||||
}
|
||||
}
|
||||
|
||||
ttk::style layout TSeparator {
|
||||
TSeparator.separator -sticky nsew
|
||||
}
|
||||
|
||||
ttk::style layout TCombobox {
|
||||
Combobox.field -sticky nsew -children {
|
||||
Combobox.padding -expand 1 -sticky nsew -children {
|
||||
Combobox.textarea -sticky nsew
|
||||
}
|
||||
}
|
||||
null -side right -sticky ns -children {
|
||||
Combobox.arrow -sticky nsew
|
||||
}
|
||||
}
|
||||
|
||||
ttk::style layout TSpinbox {
|
||||
Spinbox.field -sticky nsew -children {
|
||||
Spinbox.padding -expand 1 -sticky nsew -children {
|
||||
Spinbox.textarea -sticky nsew
|
||||
}
|
||||
|
||||
}
|
||||
null -side right -sticky nsew -children {
|
||||
Spinbox.uparrow -side left -sticky nsew
|
||||
Spinbox.downarrow -side right -sticky nsew
|
||||
}
|
||||
}
|
||||
|
||||
ttk::style layout Card.TFrame {
|
||||
Card.field {
|
||||
Card.padding -expand 1
|
||||
}
|
||||
}
|
||||
|
||||
ttk::style layout TLabelframe {
|
||||
Labelframe.border {
|
||||
Labelframe.padding -expand 1 -children {
|
||||
Labelframe.label -side left
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
ttk::style layout TNotebook {
|
||||
Notebook.border -children {
|
||||
TNotebook.Tab -expand 1
|
||||
Notebook.client -sticky nsew
|
||||
}
|
||||
}
|
||||
|
||||
ttk::style layout Treeview.Item {
|
||||
Treeitem.padding -sticky nsew -children {
|
||||
Treeitem.image -side left -sticky {}
|
||||
Treeitem.indicator -side left -sticky {}
|
||||
Treeitem.text -side left -sticky {}
|
||||
}
|
||||
}
|
||||
|
||||
# Button
|
||||
ttk::style configure TButton -padding {8 4} -anchor center -foreground $colors(-fg)
|
||||
|
||||
ttk::style map TButton -foreground \
|
||||
[list disabled #7a7a7a \
|
||||
pressed #d0d0d0]
|
||||
|
||||
ttk::style element create Button.button image \
|
||||
[list $images(button-rest) \
|
||||
{selected disabled} $images(button-disabled) \
|
||||
disabled $images(button-disabled) \
|
||||
selected $images(button-rest) \
|
||||
pressed $images(button-pressed) \
|
||||
active $images(button-hover) \
|
||||
] -border 4 -sticky nsew
|
||||
|
||||
# Toolbutton
|
||||
ttk::style configure Toolbutton -padding {8 4} -anchor center
|
||||
|
||||
ttk::style element create Toolbutton.button image \
|
||||
[list $images(empty) \
|
||||
{selected disabled} $images(button-disabled) \
|
||||
selected $images(button-rest) \
|
||||
pressed $images(button-pressed) \
|
||||
active $images(button-hover) \
|
||||
] -border 4 -sticky nsew
|
||||
|
||||
# Menubutton
|
||||
ttk::style configure TMenubutton -padding {8 4 0 4}
|
||||
|
||||
ttk::style element create Menubutton.button \
|
||||
image [list $images(button-rest) \
|
||||
disabled $images(button-disabled) \
|
||||
pressed $images(button-pressed) \
|
||||
active $images(button-hover) \
|
||||
] -border 4 -sticky nsew
|
||||
|
||||
ttk::style element create Menubutton.indicator image $images(arrow-down) -width 28 -sticky {}
|
||||
|
||||
# OptionMenu
|
||||
ttk::style configure TOptionMenu -padding {8 4 0 4}
|
||||
|
||||
ttk::style element create OptionMenu.button \
|
||||
image [list $images(button-rest) \
|
||||
disabled $images(button-disabled) \
|
||||
pressed $images(button-pressed) \
|
||||
active $images(button-hover) \
|
||||
] -border 0 -sticky nsew
|
||||
|
||||
ttk::style element create OptionMenu.indicator image $images(arrow-down) -width 28 -sticky {}
|
||||
|
||||
# Accent.TButton
|
||||
ttk::style configure Accent.TButton -padding {8 4} -anchor center -foreground #ffffff
|
||||
|
||||
ttk::style map Accent.TButton -foreground \
|
||||
[list pressed #25536a \
|
||||
disabled #a5a5a5]
|
||||
|
||||
ttk::style element create AccentButton.button image \
|
||||
[list $images(button-accent-rest) \
|
||||
{selected disabled} $images(button-accent-disabled) \
|
||||
disabled $images(button-accent-disabled) \
|
||||
selected $images(button-accent-rest) \
|
||||
pressed $images(button-accent-pressed) \
|
||||
active $images(button-accent-hover) \
|
||||
] -border 4 -sticky nsew
|
||||
|
||||
# Titlebar.TButton
|
||||
ttk::style configure Titlebar.TButton -padding {8 4} -anchor center -foreground #ffffff
|
||||
|
||||
ttk::style map Titlebar.TButton -foreground \
|
||||
[list disabled #6f6f6f \
|
||||
pressed #d1d1d1 \
|
||||
active #ffffff]
|
||||
|
||||
ttk::style element create TitlebarButton.button image \
|
||||
[list $images(empty) \
|
||||
disabled $images(empty) \
|
||||
pressed $images(button-titlebar-pressed) \
|
||||
active $images(button-titlebar-hover) \
|
||||
] -border 4 -sticky nsew
|
||||
|
||||
# Close.Titlebar.TButton
|
||||
ttk::style configure Close.Titlebar.TButton -padding {8 4} -anchor center -foreground #ffffff
|
||||
|
||||
ttk::style map Close.Titlebar.TButton -foreground \
|
||||
[list disabled #6f6f6f \
|
||||
pressed #e8bfbb \
|
||||
active #ffffff]
|
||||
|
||||
ttk::style element create CloseButton.button image \
|
||||
[list $images(empty) \
|
||||
disabled $images(empty) \
|
||||
pressed $images(button-close-pressed) \
|
||||
active $images(button-close-hover) \
|
||||
] -border 4 -sticky nsew
|
||||
|
||||
# Checkbutton
|
||||
ttk::style configure TCheckbutton -padding 4
|
||||
|
||||
ttk::style element create Checkbutton.indicator image \
|
||||
[list $images(check-unsel-rest) \
|
||||
{alternate disabled} $images(check-tri-disabled) \
|
||||
{selected disabled} $images(check-disabled) \
|
||||
disabled $images(check-unsel-disabled) \
|
||||
{pressed alternate} $images(check-tri-hover) \
|
||||
{active alternate} $images(check-tri-hover) \
|
||||
alternate $images(check-tri-rest) \
|
||||
{pressed selected} $images(check-hover) \
|
||||
{active selected} $images(check-hover) \
|
||||
selected $images(check-rest) \
|
||||
{pressed !selected} $images(check-unsel-pressed) \
|
||||
active $images(check-unsel-hover) \
|
||||
] -width 26 -sticky w
|
||||
|
||||
# Switch.TCheckbutton
|
||||
ttk::style element create Switch.indicator image \
|
||||
[list $images(switch-off-rest) \
|
||||
{selected disabled} $images(switch-on-disabled) \
|
||||
disabled $images(switch-off-disabled) \
|
||||
{pressed selected} $images(switch-on-pressed) \
|
||||
{active selected} $images(switch-on-hover) \
|
||||
selected $images(switch-on-rest) \
|
||||
{pressed !selected} $images(switch-off-pressed) \
|
||||
active $images(switch-off-hover) \
|
||||
] -width 46 -sticky w
|
||||
|
||||
# Toggle.TButton
|
||||
ttk::style configure Toggle.TButton -padding {8 4 8 4} -anchor center -foreground $colors(-fg)
|
||||
|
||||
ttk::style map Toggle.TButton -foreground \
|
||||
[list {selected disabled} #a5a5a5 \
|
||||
{selected pressed} #d0d0d0 \
|
||||
selected #ffffff \
|
||||
pressed #25536a \
|
||||
disabled #7a7a7a
|
||||
]
|
||||
|
||||
ttk::style element create ToggleButton.button image \
|
||||
[list $images(button-rest) \
|
||||
{selected disabled} $images(button-accent-disabled) \
|
||||
disabled $images(button-disabled) \
|
||||
{pressed selected} $images(button-rest) \
|
||||
{active selected} $images(button-accent-hover) \
|
||||
selected $images(button-accent-rest) \
|
||||
{pressed !selected} $images(button-accent-rest) \
|
||||
active $images(button-hover) \
|
||||
] -border 4 -sticky nsew
|
||||
|
||||
# Radiobutton
|
||||
ttk::style configure TRadiobutton -padding 0
|
||||
|
||||
ttk::style element create Radiobutton.indicator image \
|
||||
[list $images(radio-unsel-rest) \
|
||||
{selected disabled} $images(radio-disabled) \
|
||||
disabled $images(radio-unsel-disabled) \
|
||||
{pressed selected} $images(radio-pressed) \
|
||||
{active selected} $images(radio-hover) \
|
||||
selected $images(radio-rest) \
|
||||
{pressed !selected} $images(radio-unsel-pressed) \
|
||||
active $images(radio-unsel-hover) \
|
||||
] -width 20 -sticky w
|
||||
|
||||
ttk::style configure Menu.TRadiobutton -padding 0
|
||||
|
||||
ttk::style element create Menu.Radiobutton.indicator image \
|
||||
[list $images(radio-unsel-rest) \
|
||||
{selected disabled} $images(radio-disabled) \
|
||||
disabled $images(radio-unsel-disabled) \
|
||||
{pressed selected} $images(radio-pressed) \
|
||||
{active selected} $images(radio-hover) \
|
||||
selected $images(radio-rest) \
|
||||
{pressed !selected} $images(radio-unsel-pressed) \
|
||||
active $images(radio-unsel-hover) \
|
||||
] -width 20 -sticky w
|
||||
|
||||
# Scrollbar
|
||||
ttk::style element create Horizontal.Scrollbar.trough image $images(scroll-hor-trough) -sticky ew -border 6
|
||||
ttk::style element create Horizontal.Scrollbar.thumb image $images(scroll-hor-thumb) -sticky ew -border 3
|
||||
|
||||
ttk::style element create Horizontal.Scrollbar.rightarrow image $images(scroll-right) -sticky {} -width 12
|
||||
ttk::style element create Horizontal.Scrollbar.leftarrow image $images(scroll-left) -sticky {} -width 12
|
||||
|
||||
ttk::style element create Vertical.Scrollbar.trough image $images(scroll-vert-trough) -sticky ns -border 6
|
||||
ttk::style element create Vertical.Scrollbar.thumb image $images(scroll-vert-thumb) -sticky ns -border 3
|
||||
|
||||
ttk::style element create Vertical.Scrollbar.uparrow image $images(scroll-up) -sticky {} -height 12
|
||||
ttk::style element create Vertical.Scrollbar.downarrow image $images(scroll-down) -sticky {} -height 12
|
||||
|
||||
# Scale
|
||||
ttk::style element create Horizontal.Scale.trough image $images(scale-trough-hor) \
|
||||
-border 5 -padding 0
|
||||
|
||||
ttk::style element create Vertical.Scale.trough image $images(scale-trough-vert) \
|
||||
-border 5 -padding 0
|
||||
|
||||
ttk::style element create Scale.slider \
|
||||
image [list $images(scale-thumb-rest) \
|
||||
disabled $images(scale-thumb-disabled) \
|
||||
pressed $images(scale-thumb-pressed) \
|
||||
active $images(scale-thumb-hover) \
|
||||
] -sticky {}
|
||||
|
||||
# Progressbar
|
||||
ttk::style element create Horizontal.Progressbar.trough image $images(progress-trough-hor) \
|
||||
-border 1 -sticky ew
|
||||
|
||||
ttk::style element create Horizontal.Progressbar.pbar image $images(progress-pbar-hor) \
|
||||
-border 2 -sticky ew
|
||||
|
||||
ttk::style element create Vertical.Progressbar.trough image $images(progress-trough-vert) \
|
||||
-border 1 -sticky ns
|
||||
|
||||
ttk::style element create Vertical.Progressbar.pbar image $images(progress-pbar-vert) \
|
||||
-border 2 -sticky ns
|
||||
|
||||
# Entry
|
||||
ttk::style configure TEntry -foreground $colors(-fg)
|
||||
|
||||
ttk::style map TEntry -foreground \
|
||||
[list disabled #757575 \
|
||||
pressed #cfcfcf
|
||||
]
|
||||
|
||||
ttk::style element create Entry.field \
|
||||
image [list $images(entry-rest) \
|
||||
{focus hover !invalid} $images(entry-focus) \
|
||||
invalid $images(entry-invalid) \
|
||||
disabled $images(entry-disabled) \
|
||||
{focus !invalid} $images(entry-focus) \
|
||||
hover $images(entry-hover) \
|
||||
] -border 5 -padding 8 -sticky nsew
|
||||
|
||||
# Combobox
|
||||
ttk::style configure TCombobox -foreground $colors(-fg)
|
||||
|
||||
ttk::style map TCombobox -foreground \
|
||||
[list disabled #757575 \
|
||||
pressed #cfcfcf
|
||||
]
|
||||
|
||||
ttk::style configure ComboboxPopdownFrame -borderwidth 0 -flat solid
|
||||
|
||||
ttk::style map TCombobox -selectbackground [list \
|
||||
{readonly hover} $colors(-selectbg) \
|
||||
{readonly focus} $colors(-selectbg) \
|
||||
] -selectforeground [list \
|
||||
{readonly hover} $colors(-selectfg) \
|
||||
{readonly focus} $colors(-selectfg) \
|
||||
]
|
||||
|
||||
ttk::style element create Combobox.field \
|
||||
image [list $images(entry-rest) \
|
||||
{readonly disabled} $images(button-disabled) \
|
||||
{readonly pressed} $images(button-pressed) \
|
||||
{readonly hover} $images(button-hover) \
|
||||
readonly $images(button-rest) \
|
||||
invalid $images(entry-invalid) \
|
||||
disabled $images(entry-disabled) \
|
||||
focus $images(entry-focus) \
|
||||
hover $images(entry-hover) \
|
||||
] -border 0 -padding {8 8 28 8}
|
||||
|
||||
ttk::style element create Combobox.arrow image $images(arrow-down) -width 35 -sticky {}
|
||||
|
||||
# Spinbox
|
||||
ttk::style configure TSpinbox -foreground $colors(-fg)
|
||||
|
||||
ttk::style map TSpinbox -foreground \
|
||||
[list disabled #757575 \
|
||||
pressed #cfcfcf
|
||||
]
|
||||
|
||||
ttk::style element create Spinbox.field \
|
||||
image [list $images(entry-rest) \
|
||||
invalid $images(entry-invalid) \
|
||||
disabled $images(entry-disabled) \
|
||||
focus $images(entry-focus) \
|
||||
hover $images(entry-hover) \
|
||||
] -border 5 -padding {8 8 54 8} -sticky nsew
|
||||
|
||||
ttk::style element create Spinbox.uparrow image $images(arrow-up) -width 35 -sticky {}
|
||||
ttk::style element create Spinbox.downarrow image $images(arrow-down) -width 35 -sticky {}
|
||||
|
||||
# Sizegrip
|
||||
ttk::style element create Sizegrip.sizegrip image $images(sizegrip) \
|
||||
-sticky nsew
|
||||
|
||||
# Separator
|
||||
ttk::style element create TSeparator.separator image $images(separator)
|
||||
|
||||
# Card
|
||||
ttk::style element create Card.field image $images(card) \
|
||||
-border 10 -padding 4 -sticky nsew
|
||||
|
||||
# Labelframe
|
||||
ttk::style element create Labelframe.border image $images(card) \
|
||||
-border 5 -padding 4 -sticky nsew
|
||||
|
||||
# Notebook
|
||||
ttk::style configure TNotebook -padding 1
|
||||
|
||||
ttk::style element create Notebook.border \
|
||||
image $images(notebook-border) -border 5 -padding 5
|
||||
|
||||
ttk::style element create Notebook.client image $images(notebook)
|
||||
|
||||
ttk::style element create Notebook.tab \
|
||||
image [list $images(tab-rest) \
|
||||
selected $images(tab-selected) \
|
||||
active $images(tab-hover) \
|
||||
] -border 13 -padding {16 14 16 6} -height 32
|
||||
|
||||
# Treeview
|
||||
ttk::style element create Treeview.field image $images(card) \
|
||||
-border 5
|
||||
|
||||
ttk::style element create Treeheading.cell \
|
||||
image [list $images(treeheading-rest) \
|
||||
pressed $images(treeheading-pressed) \
|
||||
active $images(treeheading-hover)
|
||||
] -border 5 -padding 15 -sticky nsew
|
||||
|
||||
ttk::style element create Treeitem.indicator \
|
||||
image [list $images(arrow-right) \
|
||||
user2 $images(empty) \
|
||||
user1 $images(arrow-down) \
|
||||
] -width 26 -sticky {}
|
||||
|
||||
ttk::style configure Treeview -background $colors(-bg) -rowheight [expr {[font metrics font -linespace] + 2}]
|
||||
ttk::style map Treeview \
|
||||
-background [list selected #292929] \
|
||||
-foreground [list selected $colors(-selectfg)]
|
||||
|
||||
# Panedwindow
|
||||
# Insane hack to remove clam's ugly sash
|
||||
ttk::style configure Sash -gripcount 0
|
||||
}
|
||||
}
|
||||
Binary file not shown.
|
Before Width: | Height: | Size: 2.8 KiB |
Binary file not shown.
|
Before Width: | Height: | Size: 261 B |
Binary file not shown.
|
Before Width: | Height: | Size: 274 B |
Binary file not shown.
|
Before Width: | Height: | Size: 262 B |
Binary file not shown.
|
Before Width: | Height: | Size: 373 B |
Binary file not shown.
|
Before Width: | Height: | Size: 363 B |
Binary file not shown.
|
Before Width: | Height: | Size: 377 B |
Binary file not shown.
|
Before Width: | Height: | Size: 274 B |
Binary file not shown.
|
Before Width: | Height: | Size: 274 B |
Binary file not shown.
|
Before Width: | Height: | Size: 2.7 KiB |
Binary file not shown.
|
Before Width: | Height: | Size: 2.8 KiB |
Binary file not shown.
|
Before Width: | Height: | Size: 2.8 KiB |
Binary file not shown.
|
Before Width: | Height: | Size: 2.9 KiB |
Binary file not shown.
|
Before Width: | Height: | Size: 2.7 KiB |
Binary file not shown.
|
Before Width: | Height: | Size: 245 B |
Binary file not shown.
|
Before Width: | Height: | Size: 238 B |
Binary file not shown.
|
Before Width: | Height: | Size: 2.9 KiB |
Binary file not shown.
|
Before Width: | Height: | Size: 383 B |
Binary file not shown.
|
Before Width: | Height: | Size: 2.9 KiB |
Binary file not shown.
|
Before Width: | Height: | Size: 2.9 KiB |
Binary file not shown.
|
Before Width: | Height: | Size: 3.0 KiB |
Binary file not shown.
|
Before Width: | Height: | Size: 294 B |
Binary file not shown.
|
Before Width: | Height: | Size: 362 B |
Binary file not shown.
|
Before Width: | Height: | Size: 358 B |
Binary file not shown.
|
Before Width: | Height: | Size: 363 B |
Some files were not shown because too many files have changed in this diff Show More
Loading…
Reference in New Issue
Block a user