Spaces:
Configuration error
Configuration error
Commit ·
c9ddadb
1
Parent(s): b3b3c3c
Delete mdx.py
Browse files
mdx.py
DELETED
|
@@ -1,228 +0,0 @@
|
|
| 1 |
-
import torch
|
| 2 |
-
import onnxruntime as ort
|
| 3 |
-
from tqdm import tqdm
|
| 4 |
-
import warnings
|
| 5 |
-
import numpy as np
|
| 6 |
-
import hashlib
|
| 7 |
-
import queue
|
| 8 |
-
import threading
|
| 9 |
-
|
| 10 |
-
warnings.filterwarnings("ignore")
|
| 11 |
-
|
| 12 |
-
class MDX_Model:
|
| 13 |
-
def __init__(self, device, dim_f, dim_t, n_fft, hop=1024, stem_name=None, compensation=1.000):
|
| 14 |
-
self.dim_f = dim_f
|
| 15 |
-
self.dim_t = dim_t
|
| 16 |
-
self.dim_c = 4
|
| 17 |
-
self.n_fft = n_fft
|
| 18 |
-
self.hop = hop
|
| 19 |
-
self.stem_name = stem_name
|
| 20 |
-
self.compensation = compensation
|
| 21 |
-
|
| 22 |
-
self.n_bins = self.n_fft//2+1
|
| 23 |
-
self.chunk_size = hop * (self.dim_t-1)
|
| 24 |
-
self.window = torch.hann_window(window_length=self.n_fft, periodic=True).to(device)
|
| 25 |
-
|
| 26 |
-
out_c = self.dim_c
|
| 27 |
-
|
| 28 |
-
self.freq_pad = torch.zeros([1, out_c, self.n_bins-self.dim_f, self.dim_t]).to(device)
|
| 29 |
-
|
| 30 |
-
def stft(self, x):
|
| 31 |
-
x = x.reshape([-1, self.chunk_size])
|
| 32 |
-
x = torch.stft(x, n_fft=self.n_fft, hop_length=self.hop, window=self.window, center=True, return_complex=True)
|
| 33 |
-
x = torch.view_as_real(x)
|
| 34 |
-
x = x.permute([0,3,1,2])
|
| 35 |
-
x = x.reshape([-1,2,2,self.n_bins,self.dim_t]).reshape([-1,4,self.n_bins,self.dim_t])
|
| 36 |
-
return x[:,:,:self.dim_f]
|
| 37 |
-
|
| 38 |
-
def istft(self, x, freq_pad=None):
|
| 39 |
-
freq_pad = self.freq_pad.repeat([x.shape[0],1,1,1]) if freq_pad is None else freq_pad
|
| 40 |
-
x = torch.cat([x, freq_pad], -2)
|
| 41 |
-
# c = 4*2 if self.target_name=='*' else 2
|
| 42 |
-
x = x.reshape([-1,2,2,self.n_bins,self.dim_t]).reshape([-1,2,self.n_bins,self.dim_t])
|
| 43 |
-
x = x.permute([0,2,3,1])
|
| 44 |
-
x = x.contiguous()
|
| 45 |
-
x = torch.view_as_complex(x)
|
| 46 |
-
x = torch.istft(x, n_fft=self.n_fft, hop_length=self.hop, window=self.window, center=True)
|
| 47 |
-
return x.reshape([-1,2,self.chunk_size])
|
| 48 |
-
|
| 49 |
-
|
| 50 |
-
class MDX:
|
| 51 |
-
|
| 52 |
-
DEFAULT_SR = 44100
|
| 53 |
-
# Unit: seconds
|
| 54 |
-
DEFAULT_CHUNK_SIZE = 0 * DEFAULT_SR
|
| 55 |
-
DEFAULT_MARGIN_SIZE = 1 * DEFAULT_SR
|
| 56 |
-
|
| 57 |
-
DEFAULT_PROCESSOR = 0
|
| 58 |
-
|
| 59 |
-
def __init__(self, model_path:str, params:MDX_Model, processor=DEFAULT_PROCESSOR):
|
| 60 |
-
|
| 61 |
-
# Set the device and the provider (CPU or CUDA)
|
| 62 |
-
self.device = torch.device(f'cuda:{processor}') if processor >= 0 else torch.device('cpu')
|
| 63 |
-
self.provider = ['CUDAExecutionProvider'] if processor >= 0 else ['CPUExecutionProvider']
|
| 64 |
-
|
| 65 |
-
self.model = params
|
| 66 |
-
|
| 67 |
-
# Load the ONNX model using ONNX Runtime
|
| 68 |
-
self.ort = ort.InferenceSession(model_path, providers=self.provider)
|
| 69 |
-
# Preload the model for faster performance
|
| 70 |
-
self.ort.run(None, {'input':torch.rand(1, 4, params.dim_f, params.dim_t).numpy()})
|
| 71 |
-
self.process = lambda spec:self.ort.run(None, {'input': spec.cpu().numpy()})[0]
|
| 72 |
-
|
| 73 |
-
self.prog = None
|
| 74 |
-
|
| 75 |
-
@staticmethod
|
| 76 |
-
def get_hash(model_path):
|
| 77 |
-
try:
|
| 78 |
-
with open(model_path, 'rb') as f:
|
| 79 |
-
f.seek(- 10000 * 1024, 2)
|
| 80 |
-
model_hash = hashlib.md5(f.read()).hexdigest()
|
| 81 |
-
except:
|
| 82 |
-
model_hash = hashlib.md5(open(model_path,'rb').read()).hexdigest()
|
| 83 |
-
|
| 84 |
-
return model_hash
|
| 85 |
-
|
| 86 |
-
@staticmethod
|
| 87 |
-
def segment(wave, combine=True, chunk_size=DEFAULT_CHUNK_SIZE, margin_size=DEFAULT_MARGIN_SIZE):
|
| 88 |
-
"""
|
| 89 |
-
Segment or join segmented wave array
|
| 90 |
-
|
| 91 |
-
Args:
|
| 92 |
-
wave: (np.array) Wave array to be segmented or joined
|
| 93 |
-
combine: (bool) If True, combines segmented wave array. If False, segments wave array.
|
| 94 |
-
chunk_size: (int) Size of each segment (in samples)
|
| 95 |
-
margin_size: (int) Size of margin between segments (in samples)
|
| 96 |
-
|
| 97 |
-
Returns:
|
| 98 |
-
numpy array: Segmented or joined wave array
|
| 99 |
-
"""
|
| 100 |
-
|
| 101 |
-
if combine:
|
| 102 |
-
processed_wave = None # Initializing as None instead of [] for later numpy array concatenation
|
| 103 |
-
for segment_count, segment in enumerate(wave):
|
| 104 |
-
start = 0 if segment_count == 0 else margin_size
|
| 105 |
-
end = None if segment_count == len(wave)-1 else -margin_size
|
| 106 |
-
if margin_size == 0:
|
| 107 |
-
end = None
|
| 108 |
-
if processed_wave is None: # Create array for first segment
|
| 109 |
-
processed_wave = segment[:, start:end]
|
| 110 |
-
else: # Concatenate to existing array for subsequent segments
|
| 111 |
-
processed_wave = np.concatenate((processed_wave, segment[:, start:end]), axis=-1)
|
| 112 |
-
|
| 113 |
-
else:
|
| 114 |
-
processed_wave = []
|
| 115 |
-
sample_count = wave.shape[-1]
|
| 116 |
-
|
| 117 |
-
if chunk_size <= 0 or chunk_size > sample_count:
|
| 118 |
-
chunk_size = sample_count
|
| 119 |
-
|
| 120 |
-
if margin_size > chunk_size:
|
| 121 |
-
margin_size = chunk_size
|
| 122 |
-
|
| 123 |
-
for segment_count, skip in enumerate(range(0, sample_count, chunk_size)):
|
| 124 |
-
|
| 125 |
-
margin = 0 if segment_count == 0 else margin_size
|
| 126 |
-
end = min(skip+chunk_size+margin_size, sample_count)
|
| 127 |
-
start = skip-margin
|
| 128 |
-
|
| 129 |
-
cut = wave[:,start:end].copy()
|
| 130 |
-
processed_wave.append(cut)
|
| 131 |
-
|
| 132 |
-
if end == sample_count:
|
| 133 |
-
break
|
| 134 |
-
|
| 135 |
-
return processed_wave
|
| 136 |
-
|
| 137 |
-
def pad_wave(self, wave):
|
| 138 |
-
"""
|
| 139 |
-
Pad the wave array to match the required chunk size
|
| 140 |
-
|
| 141 |
-
Args:
|
| 142 |
-
wave: (np.array) Wave array to be padded
|
| 143 |
-
|
| 144 |
-
Returns:
|
| 145 |
-
tuple: (padded_wave, pad, trim)
|
| 146 |
-
- padded_wave: Padded wave array
|
| 147 |
-
- pad: Number of samples that were padded
|
| 148 |
-
- trim: Number of samples that were trimmed
|
| 149 |
-
"""
|
| 150 |
-
n_sample = wave.shape[1]
|
| 151 |
-
trim = self.model.n_fft//2
|
| 152 |
-
gen_size = self.model.chunk_size-2*trim
|
| 153 |
-
pad = gen_size - n_sample%gen_size
|
| 154 |
-
|
| 155 |
-
# Padded wave
|
| 156 |
-
wave_p = np.concatenate((np.zeros((2,trim)), wave, np.zeros((2,pad)), np.zeros((2,trim))), 1)
|
| 157 |
-
|
| 158 |
-
mix_waves = []
|
| 159 |
-
for i in range(0, n_sample+pad, gen_size):
|
| 160 |
-
waves = np.array(wave_p[:, i:i+self.model.chunk_size])
|
| 161 |
-
mix_waves.append(waves)
|
| 162 |
-
|
| 163 |
-
mix_waves = torch.tensor(mix_waves, dtype=torch.float32).to(self.device)
|
| 164 |
-
|
| 165 |
-
return mix_waves, pad, trim
|
| 166 |
-
|
| 167 |
-
def _process_wave(self, mix_waves, trim, pad, q:queue.Queue, _id:int):
|
| 168 |
-
"""
|
| 169 |
-
Process each wave segment in a multi-threaded environment
|
| 170 |
-
|
| 171 |
-
Args:
|
| 172 |
-
mix_waves: (torch.Tensor) Wave segments to be processed
|
| 173 |
-
trim: (int) Number of samples trimmed during padding
|
| 174 |
-
pad: (int) Number of samples padded during padding
|
| 175 |
-
q: (queue.Queue) Queue to hold the processed wave segments
|
| 176 |
-
_id: (int) Identifier of the processed wave segment
|
| 177 |
-
|
| 178 |
-
Returns:
|
| 179 |
-
numpy array: Processed wave segment
|
| 180 |
-
"""
|
| 181 |
-
mix_waves = mix_waves.split(1)
|
| 182 |
-
with torch.no_grad():
|
| 183 |
-
pw = []
|
| 184 |
-
for mix_wave in mix_waves:
|
| 185 |
-
self.prog.update()
|
| 186 |
-
spec = self.model.stft(mix_wave)
|
| 187 |
-
processed_spec = torch.tensor(self.process(spec))
|
| 188 |
-
processed_wav = self.model.istft(processed_spec.to(self.device))
|
| 189 |
-
processed_wav = processed_wav[:,:,trim:-trim].transpose(0,1).reshape(2, -1).cpu().numpy()
|
| 190 |
-
pw.append(processed_wav)
|
| 191 |
-
processed_signal = np.concatenate(pw, axis=-1)[:, :-pad]
|
| 192 |
-
q.put({_id:processed_signal})
|
| 193 |
-
return processed_signal
|
| 194 |
-
|
| 195 |
-
def process_wave(self, wave:np.array, mt_threads=1):
|
| 196 |
-
"""
|
| 197 |
-
Process the wave array in a multi-threaded environment
|
| 198 |
-
|
| 199 |
-
Args:
|
| 200 |
-
wave: (np.array) Wave array to be processed
|
| 201 |
-
mt_threads: (int) Number of threads to be used for processing
|
| 202 |
-
|
| 203 |
-
Returns:
|
| 204 |
-
numpy array: Processed wave array
|
| 205 |
-
"""
|
| 206 |
-
self.prog = tqdm(total=0)
|
| 207 |
-
chunk = wave.shape[-1]//mt_threads
|
| 208 |
-
waves = self.segment(wave, False, chunk)
|
| 209 |
-
|
| 210 |
-
# Create a queue to hold the processed wave segments
|
| 211 |
-
q = queue.Queue()
|
| 212 |
-
threads = []
|
| 213 |
-
for c, batch in enumerate(waves):
|
| 214 |
-
mix_waves, pad, trim = self.pad_wave(batch)
|
| 215 |
-
self.prog.total = len(mix_waves)*mt_threads
|
| 216 |
-
thread = threading.Thread(target=self._process_wave, args=(mix_waves, trim, pad, q, c))
|
| 217 |
-
thread.start()
|
| 218 |
-
threads.append(thread)
|
| 219 |
-
for thread in threads:
|
| 220 |
-
thread.join()
|
| 221 |
-
self.prog.close()
|
| 222 |
-
|
| 223 |
-
processed_batches = []
|
| 224 |
-
while not q.empty():
|
| 225 |
-
processed_batches.append(q.get())
|
| 226 |
-
processed_batches = [list(wave.values())[0] for wave in sorted(processed_batches, key=lambda d: list(d.keys())[0])]
|
| 227 |
-
assert len(processed_batches) == len(waves), 'Incomplete processed batches, please reduce batch size!'
|
| 228 |
-
return self.segment(processed_batches, True, chunk)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|