Fixed typos in various files

README.md

@@ -104,7 +104,7 @@ Train AudioSep from scratch:
   python train.py --workspace workspace/AudioSep --config_yaml config/audiosep_base.yaml --resume_checkpoint_path checkpoint/ ''
   ```
 
-Finetune AudioSep from pretrained checkpoint:
+Finetune AudioSep from pre-trained checkpoint:
   ```python
   python train.py --workspace workspace/AudioSep --config_yaml config/audiosep_base.yaml --resume_checkpoint_path path_to_checkpoint
   ```

data/datamodules.py

@@ -47,7 +47,7 @@ def setup(self, stage: Optional[str] = None) -> NoReturn:
         # called on every process in DDP
 
         # SegmentSampler is used for selecting segments for training.
-        # On multiple devices, each SegmentSampler samples a part of mini-batch
+        # On multiple devices, each SegmentSampler samples a part of a mini-batch
         # data.
         self.train_dataset = self._train_dataset
 
@@ -119,4 +119,4 @@ def collate_fn(list_data_dict):
         'audio_text': at_data_dict
     }
 
-    return data_dict
+    return data_dict

data/waveform_mixers.py

@@ -43,10 +43,10 @@ def __call__(self, waveforms):
             # randomly normalize background noise
             noise = dynamic_loudnorm(audio=noise, reference=segment, **self.loudness_param)
 
-            # create audio mixyure
+            # create an audio mixture
             mixture = segment + noise
 
-            # declipping if need be
+            # Declipping if need be
             max_value = torch.max(torch.abs(mixture))
             if max_value > 1:
                 segment *= 0.9 / max_value

evaluation/evaluate_audiocaps.py

@@ -51,7 +51,7 @@ def __call__(
         self,
         pl_model: pl.LightningModule
     ) -> Dict:
-        r"""Evalute."""
+        r"""Evaluate."""
 
         print(f'Evaluation on AudioCaps with [{self.query}] queries.')
 
@@ -107,4 +107,4 @@ def __call__(
         mean_sisdr = np.mean(sisdrs_list)
         mean_sdri = np.mean(sdris_list)
 
-        return mean_sisdr, mean_sdri
+        return mean_sisdr, mean_sdri

evaluation/evaluate_audioset.py

@@ -62,7 +62,7 @@ def __call__(
         self,
         pl_model: pl.LightningModule
     ) -> Dict:
-        r"""Evalute."""
+        r"""Evaluate."""
 
         pl_model.eval()
 

evaluation/evaluate_clotho.py

@@ -46,7 +46,7 @@ def __call__(
         self,
         pl_model: pl.LightningModule
     ) -> Dict:
-        r"""Evalute."""
+        r"""Evaluate."""
 
         print(f'Evaluation on Clotho Evaluation with [caption] queries.')
 
@@ -99,4 +99,4 @@ def __call__(
         mean_sisdr = np.mean(sisdrs_list)
         mean_sdri = np.mean(sdris_list)
 
-        return mean_sisdr, mean_sdri
+        return mean_sisdr, mean_sdri

evaluation/evaluate_esc50.py

@@ -47,7 +47,7 @@ def __call__(
         self,
         pl_model: pl.LightningModule
     ) -> Dict:
-        r"""Evalute."""
+        r"""Evaluate."""
 
         print(f'Evaluation on ESC-50 with [text label] queries.')
 

evaluation/evaluate_music.py

@@ -54,7 +54,7 @@ def __call__(
         self,
         pl_model: pl.LightningModule
     ) -> Dict:
-        r"""Evalute."""
+        r"""Evaluate."""
 
         print(f'Evaluation on MUSIC Test with [text label] queries.')
 
@@ -115,4 +115,4 @@ def __call__(
         mean_sdri = np.mean(mean_sdri_list)
         mean_sisdr = np.mean(mean_sisdr_list)
 
-        return mean_sisdr, mean_sdri
+        return mean_sisdr, mean_sdri

evaluation/evaluate_vggsound.py

@@ -49,7 +49,7 @@ def __call__(
         self,
         pl_model: pl.LightningModule
     ) -> Dict:
-        r"""Evalute."""
+        r"""Evaluate."""
 
         print(f'Evaluation on VGGSound+ with [text label] queries.')
 
@@ -111,4 +111,4 @@ def __call__(
         mean_sisdr = np.mean(sisdrs_list)
         mean_sdri = np.mean(sdris_list)
 
-        return mean_sisdr, mean_sdri
+        return mean_sisdr, mean_sdri