keithito / tacotron

A TensorFlow implementation of Google's Tacotron speech synthesis with pre-trained model (unofficial)
MIT License
2.96k stars 958 forks source link

Cannot convert a symbolic Tensor Error #359

Closed thegeekywanderer closed 3 years ago

thegeekywanderer commented 3 years ago

Traceback (most recent call last): File "demo_server.py", line 91, in synthesizer.load(args.checkpoint) File "/home/wolf/Desktop/Project Alpha/tacotron/synthesizer.py", line 19, in load self.wav_output = audio.inv_spectrogram_tensorflow(self.model.linear_outputs[0]) File "/home/wolf/Desktop/Project Alpha/tacotron/util/audio.py", line 46, in inv_spectrogram_tensorflow return _griffin_lim_tensorflow(tf.pow(S, hparams.power)) File "/home/wolf/Desktop/Project Alpha/tacotron/util/audio.py", line 86, in _griffin_lim_tensorflow y = _istft_tensorflow(S_complex) File "/home/wolf/Desktop/Project Alpha/tacotron/util/audio.py", line 111, in _istft_tensorflow return tf.contrib.signal.inverse_stft(stfts, win_length, hop_length, n_fft) File "/home/wolf/anaconda3/envs/project_alpha/lib/python3.7/site-packages/tensorflow_core/python/ops/signal/spectral_ops.py", line 275, in inverse_stft return reconstruction_ops.overlap_and_add(real_frames, frame_step) File "/home/wolf/anaconda3/envs/project_alpha/lib/python3.7/site-packages/tensorflow_core/python/ops/signal/reconstruction_ops.py", line 107, in overlap_and_add outer_paddings = array_ops.zeros([outer_rank, 2], dtypes.int32) File "/home/wolf/anaconda3/envs/project_alpha/lib/python3.7/site-packages/tensorflow_core/python/ops/array_ops.py", line 2338, in zeros output = _constant_if_small(zero, shape, dtype, name) File "/home/wolf/anaconda3/envs/project_alpha/lib/python3.7/site-packages/tensorflow_core/python/ops/array_ops.py", line 2295, in _constant_if_small if np.prod(shape) < 1000: File "<__array_function__ internals>", line 6, in prod File "/home/wolf/anaconda3/envs/project_alpha/lib/python3.7/site-packages/numpy/core/fromnumeric.py", line 3031, in prod keepdims=keepdims, initial=initial, where=where) File "/home/wolf/anaconda3/envs/project_alpha/lib/python3.7/site-packages/numpy/core/fromnumeric.py", line 87, in _wrapreduction return ufunc.reduce(obj, axis, dtype, out, **passkwargs) File "/home/wolf/anaconda3/envs/project_alpha/lib/python3.7/site-packages/tensorflow_core/python/framework/ops.py", line 736, in array " array.".format(self.name)) NotImplementedError: Cannot convert a symbolic Tensor (model/griffinlim/inverse_stft/overlap_and_add/Size:0) to a numpy array.

vinamramunot-tech commented 3 years ago

@wolf-hash I am getting this error too. Were you able to find a solution?

thegeekywanderer commented 3 years ago

@wolf-hash I am getting this error too. Were you able to find a solution?

Yes the solution was quite simple. I downgraded numpy to 1.19.5 pip install numpy==1.19.5 and it seemed to fix the error. Apparently it automatically got upgraded to 1.20.0 while I was installing tensorflow 2.x Sorry I forgot to close this issue.