This function should declare the variables 'bucket_batch_sizes' and 'bucket_boundaries'. However, line 77 of 'extract_durations.py' has omitted 'bucket_boundaries' variable.
I tried insert 'bucket_boundaries=config['bucket_boundaries'],' in the code like this:
Traceback (most recent call last): File "/home/aiuser8/PycharmProjects/TransformerTTS/extract_durations.py", line 114, in <module> pred_mel = tf.expand_dims(1 - tf.squeeze(create_mel_padding_mask(mel_batch[:, 1:, :])), -1) * pred_mel File "/home/aiuser8/PycharmProjects/TransformerTTS/venv/lib/python3.6/site-packages/tensorflow/python/ops/math_ops.py", line 1164, in binary_op_wrapper return func(x, y, name=name) File "/home/aiuser8/PycharmProjects/TransformerTTS/venv/lib/python3.6/site-packages/tensorflow/python/ops/math_ops.py", line 1496, in _mul_dispatch return multiply(x, y, name=name) File "/home/aiuser8/PycharmProjects/TransformerTTS/venv/lib/python3.6/site-packages/tensorflow/python/util/dispatch.py", line 201, in wrapper return target(*args, **kwargs) File "/home/aiuser8/PycharmProjects/TransformerTTS/venv/lib/python3.6/site-packages/tensorflow/python/ops/math_ops.py", line 518, in multiply return gen_math_ops.mul(x, y, name) File "/home/aiuser8/PycharmProjects/TransformerTTS/venv/lib/python3.6/site-packages/tensorflow/python/ops/gen_math_ops.py", line 6068, in mul _ops.raise_from_not_ok_status(e, name) File "/home/aiuser8/PycharmProjects/TransformerTTS/venv/lib/python3.6/site-packages/tensorflow/python/framework/ops.py", line 6862, in raise_from_not_ok_status six.raise_from(core._status_to_exception(e.code, message), None) File "<string>", line 3, in raise_from tensorflow.python.framework.errors_impl.InvalidArgumentError: Incompatible shapes: [21,593,1] vs. [21,600,80] [Op:Mul] 2020-12-23 13:44:32.902541: W tensorflow/core/kernels/data/generator_dataset_op.cc:107] Error occurred when finalizing GeneratorDataset iterator: Failed precondition: Python interpreter state is not initialized. The process may be terminated. [[{{node PyFunc}}]]
Incompatible shapes: [21,593,1] vs. [21,600,80] message, It seems that the wrong array shape was used.
What is best declares of 'bucket_boundaries' variables?
Hello, I have tried Compute alignment dataset part, but I cannot run extract_durations.py . I had run following command
python extract_durations.py --config config/wavernn --binary --fix_jumps --fill_mode_nextBut, I got the following error:
ERROR: model's reduction factor is greater than 1, check config. (r=10 Traceback (most recent call last): File "/home/aiuser8/PycharmProjects/TransformerTTS/extract_durations.py", line 77, in <module> dataset = data_handler.get_dataset(bucket_batch_sizes=[64, 42, 32, 25, 21, 18, 16, 14, 12, 11, 1], shuffle=False, drop_remainder=False) TypeError: get_dataset() missing 1 required positional argument: 'bucket_boundaries'The 'data_handler.get_dataset()' function is declared at line 80 in preprocessing/datasets.py .
def get_dataset(self, bucket_batch_sizes, bucket_boundaries, shuffle=True, drop_remainder=False):This function should declare the variables 'bucket_batch_sizes' and 'bucket_boundaries'. However, line 77 of 'extract_durations.py' has omitted 'bucket_boundaries' variable.
I tried insert 'bucket_boundaries=config['bucket_boundaries'],' in the code like this:
dataset = data_handler.get_dataset(bucket_batch_sizes=[64, 42, 32, 25, 21, 18, 16, 14, 12, 11, 1], bucket_boundaries=config['bucket_boundaries'], shuffle=False, drop_remainder=False)But I faced the following error.
Traceback (most recent call last): File "/home/aiuser8/PycharmProjects/TransformerTTS/extract_durations.py", line 114, in <module> pred_mel = tf.expand_dims(1 - tf.squeeze(create_mel_padding_mask(mel_batch[:, 1:, :])), -1) * pred_mel File "/home/aiuser8/PycharmProjects/TransformerTTS/venv/lib/python3.6/site-packages/tensorflow/python/ops/math_ops.py", line 1164, in binary_op_wrapper return func(x, y, name=name) File "/home/aiuser8/PycharmProjects/TransformerTTS/venv/lib/python3.6/site-packages/tensorflow/python/ops/math_ops.py", line 1496, in _mul_dispatch return multiply(x, y, name=name) File "/home/aiuser8/PycharmProjects/TransformerTTS/venv/lib/python3.6/site-packages/tensorflow/python/util/dispatch.py", line 201, in wrapper return target(*args, **kwargs) File "/home/aiuser8/PycharmProjects/TransformerTTS/venv/lib/python3.6/site-packages/tensorflow/python/ops/math_ops.py", line 518, in multiply return gen_math_ops.mul(x, y, name) File "/home/aiuser8/PycharmProjects/TransformerTTS/venv/lib/python3.6/site-packages/tensorflow/python/ops/gen_math_ops.py", line 6068, in mul _ops.raise_from_not_ok_status(e, name) File "/home/aiuser8/PycharmProjects/TransformerTTS/venv/lib/python3.6/site-packages/tensorflow/python/framework/ops.py", line 6862, in raise_from_not_ok_status six.raise_from(core._status_to_exception(e.code, message), None) File "<string>", line 3, in raise_from tensorflow.python.framework.errors_impl.InvalidArgumentError: Incompatible shapes: [21,593,1] vs. [21,600,80] [Op:Mul] 2020-12-23 13:44:32.902541: W tensorflow/core/kernels/data/generator_dataset_op.cc:107] Error occurred when finalizing GeneratorDataset iterator: Failed precondition: Python interpreter state is not initialized. The process may be terminated. [[{{node PyFunc}}]]Incompatible shapes: [21,593,1] vs. [21,600,80]message, It seems that the wrong array shape was used. What is best declares of 'bucket_boundaries' variables?