update example to use new functionals (#5330)

PaddleAudio/examples/audioset_training/dataset.py

@@ -26,7 +26,7 @@ import paddle
 import paddleaudio
 import yaml
 from paddle.io import DataLoader, Dataset, IterableDataset
-from paddleaudio import augment
+from paddleaudio.utils import augments
 from utils import get_labels, get_ytid_clsidx_mapping
 
 
@@ -108,12 +108,14 @@ class H5AudioSet(Dataset):
         x = x[:, :target_len]
 
         if self.training and self.augment:
-            x = augment.random_crop2d(x,
-                                      self.config['mel_crop_len'],
-                                      tempo_axis=1)
+            x = augments.random_crop2d(x,
+                                       self.config['mel_crop_len'],
+                                       tempo_axis=1)
             x = spect_permute(x, tempo_axis=1, nblocks=random_choice([0, 2, 3]))
             aug_level = random_choice([0.2, 0.1, 0])
-            x = augment.adaptive_spect_augment(x, tempo_axis=1, level=aug_level)
+            x = augments.adaptive_spect_augment(x,
+                                                tempo_axis=1,
+                                                level=aug_level)
         return x.T
 
     def __getitem__(self, idx):

PaddleAudio/examples/audioset_training/evaluate.py

@@ -24,10 +24,11 @@ from dataset import get_val_loader
 from model import resnet50
 from paddle.utils import download
 from sklearn.metrics import average_precision_score, roc_auc_score
-from utils import compute_dprime,download_assets
+from utils import compute_dprime, download_assets
 
 checkpoint_url = 'https://bj.bcebos.com/paddleaudio/paddleaudio/resnet50_weight_averaging_mAP0.416.pdparams'
 
+
 def evaluate(epoch, val_loader, model, loss_fn):
     model.eval()
     avg_loss = 0.0

PaddleAudio/examples/audioset_training/inference.py

@@ -22,6 +22,7 @@ import paddleaudio as pa
 import yaml
 from model import resnet50
 from paddle.utils import download
+from paddleaudio.functional import melspectrogram
 from utils import (download_assets, get_label_name_mapping, get_labels,
                    get_metrics)
 
@@ -32,22 +33,22 @@ checkpoint_url = 'https://bj.bcebos.com/paddleaudio/paddleaudio/resnet50_weight_
 
 def load_and_extract_feature(file, c):
     s, _ = pa.load(file, sr=c['sample_rate'])
-    x = pa.features.melspectrogram(s,
-                                   sr=c['sample_rate'],
-                                   window_size=c['window_size'],
-                                   hop_length=c['hop_size'],
-                                   n_mels=c['mel_bins'],
-                                   fmin=c['fmin'],
-                                   fmax=c['fmax'],
-                                   window='hann',
-                                   center=True,
-                                   pad_mode='reflect',
-                                   ref=1.0,
-                                   amin=1e-10,
-                                   top_db=None)
-
-    x = x.T  # !!
-    x = paddle.Tensor(x).unsqueeze((0, 1))
+    x = melspectrogram(paddle.to_tensor(s),
+                       sr=c['sample_rate'],
+                       win_length=c['window_size'],
+                       n_fft=c['window_size'],
+                       hop_length=c['hop_size'],
+                       n_mels=c['mel_bins'],
+                       f_min=c['fmin'],
+                       f_max=c['fmax'],
+                       window='hann',
+                       center=True,
+                       pad_mode='reflect',
+                       to_db=True,
+                       amin=1e-3,
+                       top_db=None)
+    x = x.transpose((0, 2, 1))
+    x = x.unsqueeze((0, ))
     return x
 
 

PaddleAudio/examples/audioset_training/train.py

@@ -129,7 +129,7 @@ if __name__ == '__main__':
         model.train()
         model.clear_gradients()
         t0 = time.time()
-        for batch_id, (x,y) in enumerate(train_loader()):
+        for batch_id, (x, y) in enumerate(train_loader()):
             if step < warm_steps:
                 optimizer.set_lr(lrs[step])
             x.stop_gradient = False
@@ -215,4 +215,4 @@ if __name__ == '__main__':
                 else:
                     factor = 0.8
                 optimizer.set_lr(optimizer.get_lr() * factor)
-                print('decreased lr to {}'.format(optimizer.get_lr()))
\ No newline at end of file
+                print('decreased lr to {}'.format(optimizer.get_lr()))

PaddleAudio/examples/audioset_training/wav2mel.py

@@ -4,8 +4,10 @@ import os
 
 import h5py
 import numpy as np
+import paddle
 import paddleaudio as pa
 import tqdm
+from paddleaudio.functional import melspectrogram
 
 parser = argparse.ArgumentParser(description='wave2mel')
 parser.add_argument('--wav_file', type=str, required=False, default='')
@@ -64,20 +66,23 @@ if len(h5_files) > 0:
             s = src_h5[key][:]
             s = pa.depth_convert(s, 'float32')
             # s = pa.resample(s,32000,args.sample_rate)
-            x = pa.features.melspectrogram(s,
-                                           sr=args.sample_rate,
-                                           window_size=args.window_size,
-                                           hop_length=args.hop_length,
-                                           n_mels=args.mel_bins,
-                                           fmin=args.fmin,
-                                           fmax=args.fmax,
-                                           window='hann',
-                                           center=True,
-                                           pad_mode='reflect',
-                                           ref=1.0,
-                                           amin=1e-10,
-                                           top_db=None)
-            dst_h5.create_dataset(key, data=x)
+
+            x = melspectrogram(paddle.to_tensor(s),
+                               sr=args.sample_rate,
+                               win_length=args.window_size,
+                               n_fft=args.window_size,
+                               hop_length=args.hop_length,
+                               n_mels=args.mel_bins,
+                               f_min=args.fmin,
+                               f_max=args.fmax,
+                               window='hann',
+                               center=True,
+                               pad_mode='reflect',
+                               to_db=True,
+                               amin=1e-3,
+                               top_db=None)
+
+            dst_h5.create_dataset(key, data=x[0].numpy())
         src_h5.close()
         dst_h5.close()
 
@@ -91,23 +96,24 @@ if len(wav_files) > 0:
     print(f'{len(wav_files)} wav files listed')
     for f in tqdm.tqdm(wav_files):
         s, _ = pa.load(f, sr=args.sample_rate)
-        x = pa.melspectrogram(s,
-                              sr=args.sample_rate,
-                              window_size=args.window_size,
-                              hop_length=args.hop_length,
-                              n_mels=args.mel_bins,
-                              fmin=args.fmin,
-                              fmax=args.fmax,
-                              window='hann',
-                              center=True,
-                              pad_mode='reflect',
-                              ref=1.0,
-                              amin=1e-10,
-                              top_db=None)
+        x = melspectrogram(paddle.to_tensor(s),
+                           sr=args.sample_rate,
+                           win_length=args.window_size,
+                           n_fft=args.window_size,
+                           hop_length=args.hop_length,
+                           n_mels=args.mel_bins,
+                           f_min=args.fmin,
+                           f_max=args.fmax,
+                           window='hann',
+                           center=True,
+                           pad_mode='reflect',
+                           to_db=True,
+                           amin=1e-3,
+                           top_db=None)
         #         figure(figsize=(8,8))
         #         imshow(x)
         #         show()
         #         print(x.shape)
         key = f.split('/')[-1][:11]
-        dst_h5.create_dataset(key, data=x)
+        dst_h5.create_dataset(key, data=x[0].numpy())
     dst_h5.close()