size mismatch error

[immonomono]

Hello,

I have problem when I run the train.py script with ASVspoof2017 data like below

I've edited extract_mfcc method in extract_feature.py because librosa.feature.delta(mfcc) shows librosa.util.exceptions.ParameterError: when mode='interp', width=9 cannot exceed data.shape[axis]=7 error

So, I changed from mfcc_delta = librosa.feature.delta(mfcc) mfcc_delta_delta = librosa.feature.delta(mfcc) to mfcc_delta = librosa.feature.delta(mfcc, mode='nearest') mfcc_delta_delta = librosa.feature.delta(mfcc, mode='nearest')

after that I got the another error like below

RuntimeError: size mismatch, m1: [256 x 429], m2: [990 x 512] at /opt/conda/conda-bld/pytorch_1544173631724/work/aten/src/THC/generic/THCTensorMathBlas.cu:266

Can you give me some help?

[azraelkuan]

may be u can try librosa 0.5.1??