search

Music-and-Culture-Technology-Lab / omnizart

Omniscient Mozart, being able to transcribe everything in the music, including vocal, drum, chord, beat, instruments, and more.
https://music-and-culture-technology-lab.github.io/omnizart-doc/
MIT License
1.63k stars 99 forks source link

chord@v2 model RuntimeError #90

Open liufeigit opened 1 year ago

liufeigit commented 1 year ago

Hello, omnizart chord transcribe use chord@v1 model is suscess, but use chord@v2 model is fail, 

Traceback (most recent call last):
  File "python3.8/site-packages/tensorflow/python/training/py_checkpoint_reader.py", line 69, in get_tensor
    return CheckpointReader.CheckpointReader_GetTensor(
RuntimeError: The length checksum does not match: expected 2557634344 but actual is 2753687263
tensorflow.python.framework.errors_impl.OpError: The length checksum does not match: expected 2557634344 but actual is 2753687263