Closed lihuxiao1 closed 1 year ago
I adjusted the version of numpy and forcibly generated a txt file,However, when I run GestreKNN, this phase will report an error, and its scale is not consistent with (n len 4) in the downloaded file
Dear lihuxiao1,
Thank you for catching this problem!
I changed this line (you need to change it too), plus I updated this file a bit (should not affect your running code), and I upload a bash to run quick start; then the detailed process of rebuilding the dataset on speaker 1
is shown here.
Best Regards
Dear lihuxiao1,
Thank you for catching this problem!
I changed this line (you need to change it too), plus I updated this file a bit (should not affect your running code), and I upload a bash to run quick start; then the detailed process of rebuilding the dataset on speaker
1
is shown here.Best Regards
Thank you very much,I have successfully reproduced it
Traceback (most recent call last): File "make_beat_dataset.py", line 609, in
make_txt_dataset(save_dir, prefix, mode='noduplication')
File "make_beat_dataset.py", line 577, in make_txt_dataset
make_duplication_dataset(key, stride=n_frames)
File "make_beat_dataset.py", line 571, in make_duplication_dataset
txt=np.array(txt_data), aux=np.array(aux_data), energy=np.array(energy_data),
ValueError: setting an array element with a sequence. The requested array has an inhomogeneous shape after 1 dimensions. The detected shape was (1424,) + inhomogeneous part.
when I generate the speaker_1_state_0_train_240_txt.npz , i can't solve this problem.I guess it's because there was an error generating the file before