search

andabi / music-source-separation

Deep neural networks for separating singing voice from music written in TensorFlow
796 stars 149 forks source link

Im trying to create colab file but when I run train.py I get this error which i dont get on my machine #33

Open archmord opened 5 years ago

archmord commented 5 years ago

Traceback (most recent call last): File "train.py", line 98, in <module> train() File "train.py", line 47, in train mixed_wav, src1_wav, src2_wav, _ = data.next_wavs(TrainConfig.SECONDS, TrainConfig.NUM_WAVFILE) File "/content/music-source-separation/data.py", line 22, in next_wavs wavfiles = random.sample(wavfiles, size) File "/usr/lib/python3.6/random.py", line 320, in sample raise ValueError("Sample larger than population or is negative") ValueError: Sample larger than population or is negative

appernetic commented 5 years ago

Oracle Filtering! https://colab.research.google.com/drive/1Zo6iSPIi6SjOAL7wg8yzVWkS9mjLgjI-