Closed caigun closed 8 months ago
Hi, what version of librosa are you using?
0.10.1
Would you please give the log of the wav_path
and sampling_rate
in your experiment?
Thanks~ Problem solved: Amphion/preprocessors/ljspeech.py, line 139, in prepare_align: add a "sr=" before "samplingrate": wav, = librosa.load(wav_path, sr=sampling_rate)
Describe the bug
I followed the tutorials for the example recipe of FastSpeech2 and didn't pass the first stage. This problem also occurs on my Windows laptop.
How To Reproduce
Steps to reproduce the behavior:
Expected behavior
Data Preparation failed and was interrupted.
Screenshots
See error: (Amphion) harrywang@Harrys-MacBook-Air Amphion % sh egs/tts/FastSpeech2/run.sh --stage 1 /Users/harrywang/Amphion/mfa Exprimental Configuration File: /Users/harrywang/Amphion/egs/tts/FastSpeech2/exp_config.json Preprocess LJSpeech... Prepare alignment LJSpeech... 0it [00:00, ?it/s] Traceback (most recent call last): File "/Users/harrywang/Amphion/bins/tts/preprocess.py", line 244, in
main()
File "/Users/harrywang/Amphion/bins/tts/preprocess.py", line 240, in main
preprocess(cfg, args)
File "/Users/harrywang/Amphion/bins/tts/preprocess.py", line 112, in preprocess
prepare_align(
File "/Users/harrywang/Amphion/preprocessors/processor.py", line 104, in prepare_align
ljspeech.prepare_align(dataset, dataset_path, cfg, output_path)
File "/Users/harrywang/Amphion/preprocessors/ljspeech.py", line 139, in preparealign
wav, = librosa.load(wav_path, sampling_rate)
TypeError: load() takes 1 positional argument but 2 were given
Environment Information