asahi417 commited on
Commit
fdfc0ef
·
1 Parent(s): 7f90e24
Files changed (2) hide show
  1. fetch_dataset_s2s.py +1 -4
  2. main_s2s.sh +1 -1
fetch_dataset_s2s.py CHANGED
@@ -135,13 +135,10 @@ def get_audio(dataframe: pd.DataFrame):
135
  wav = wav[:int(end/sampling_rate * sr) + sr]
136
  if len(wav) < min_seq_length or len(wav) > max_seq_length:
137
  print(f"invalid wave length: {len(wav)}")
 
138
  return None
139
  print(f"SAVING: {features[f'{side}.path']}")
140
  sf.write(features[f"{side}.path"], wav, sr)
141
- # if sr != sampling_rate:
142
- # print(f"RESAMPLING: {wav.shape} length audio")
143
- # wav = librosa.resample(wav, orig_sr=sr, target_sr=sampling_rate)
144
- # sf.write(features[f"{side}.path"], wav[start:end], sampling_rate)
145
 
146
  except Exception as e:
147
  print(f"\n#### ERROR ####\n {e}")
 
135
  wav = wav[:int(end/sampling_rate * sr) + sr]
136
  if len(wav) < min_seq_length or len(wav) > max_seq_length:
137
  print(f"invalid wave length: {len(wav)}")
138
+ cleanup(features, feature_file)
139
  return None
140
  print(f"SAVING: {features[f'{side}.path']}")
141
  sf.write(features[f"{side}.path"], wav, sr)
 
 
 
 
142
 
143
  except Exception as e:
144
  print(f"\n#### ERROR ####\n {e}")
main_s2s.sh CHANGED
@@ -312,8 +312,8 @@ done
312
  # speaker embedding
313
  export SE_MODEL="metavoice"
314
  export SE_MODEL="w2vbert-600m"
315
- export SE_MODEL="xlsr-2b"
316
  export SE_MODEL="hubert-xl"
 
317
  for i in $(seq 1 100);
318
  do
319
  export DATASET_ID=${i}
 
312
  # speaker embedding
313
  export SE_MODEL="metavoice"
314
  export SE_MODEL="w2vbert-600m"
 
315
  export SE_MODEL="hubert-xl"
316
+ export SE_MODEL="xlsr-2b"
317
  for i in $(seq 1 100);
318
  do
319
  export DATASET_ID=${i}