dhgrs / chainer-waveglow Goto Github PK
View Code? Open in Web Editor NEWA Chainer implementation of WaveGlow.
A Chainer implementation of WaveGlow.
Hi!
I run the script with a single GPU, batch_size = 1 and iteration=1000000 but learning time will be 50 days.
How long have you been learning with the default parameters?
While trying to generate audio from tacotron2 generated mel, generate.py just throws error on line 30, and hangs.
Traceback (most recent call last):
File "generate.py", line 30, in <module>
By executing command python train.py -g 7 -r snapshot_iter_500000
it throws error you see below.
P.S. additionally, -g 7 runs the model on eighth gpu, how can I run it on all gpus?
Traceback (most recent call last):
File "train.py", line 133, in <module>
chainer.serializers.load_npz(args.resume, trainer)
File "/home/lchkhetianni/.conda/envs/TARS/lib/python3.6/site-packages/chainer/serializers/npz.py", line 190, in load_npz
d.load(obj)
File "/home/lchkhetianni/.conda/envs/TARS/lib/python3.6/site-packages/chainer/serializer.py", line 83, in load
obj.serialize(self)
File "/home/lchkhetianni/.conda/envs/TARS/lib/python3.6/site-packages/chainer/training/trainer.py", line 341, in serialize
self.updater.serialize(serializer['updater'])
File "/home/lchkhetianni/.conda/envs/TARS/lib/python3.6/site-packages/chainer/training/updaters/standard_updater.py", line 189, in serialize
iterator.serialize(serializer['iterator:' + name])
File "/home/lchkhetianni/.conda/envs/TARS/lib/python3.6/site-packages/chainer/iterators/multiprocess_iterator.py", line 216, in serialize
serializer('order', self._order)
File "/home/lchkhetianni/.conda/envs/TARS/lib/python3.6/site-packages/chainer/serializers/npz.py", line 150, in __call__
numpy.copyto(value, dataset)
ValueError: could not broadcast input array from shape (11790) into shape (0)
Exception ignored in: <bound method MultiprocessIterator.__del__ of <chainer.iterators.multiprocess_iterator.MultiprocessIterator object at 0x7f5f708567b8>>
Traceback (most recent call last):
File "/home/lchkhetianni/.conda/envs/TARS/lib/python3.6/site-packages/chainer/iterators/multiprocess_iterator.py", line 168, in __del__
File "/home/lchkhetianni/.conda/envs/TARS/lib/python3.6/site-packages/chainer/iterators/multiprocess_iterator.py", line 321, in terminate
File "/home/lchkhetianni/.conda/envs/TARS/lib/python3.6/threading.py", line 347, in notify
TypeError: 'NoneType' object is not callable
Exception ignored in: <bound method MultiprocessIterator.__del__ of <chainer.iterators.multiprocess_iterator.MultiprocessIterator object at 0x7f5f6d4ca9e8>>
Traceback (most recent call last):
File "/home/lchkhetianni/.conda/envs/TARS/lib/python3.6/site-packages/chainer/iterators/multiprocess_iterator.py", line 168, in __del__
File "/home/lchkhetianni/.conda/envs/TARS/lib/python3.6/site-packages/chainer/iterators/multiprocess_iterator.py", line 321, in terminate
File "/home/lchkhetianni/.conda/envs/TARS/lib/python3.6/threading.py", line 347, in notify
TypeError: 'NoneType' object is not callable
Training throws this error. @dhgrs could you please help me out?
Traceback (most recent call last):
File "train.py", line 133, in <module>
chainer.serializers.load_npz(args.resume, trainer)
File "/home/lchkhetianni/.local/lib/python3.5/site-packages/chainer/serializers/npz.py", line 221, in load_npz
d.load(obj)
File "/home/lchkhetianni/.local/lib/python3.5/site-packages/chainer/serializer.py", line 83, in load
obj.serialize(self)
File "/home/lchkhetianni/.local/lib/python3.5/site-packages/chainer/training/trainer.py", line 361, in serialize
self.updater.serialize(serializer['updater'])
File "/home/lchkhetianni/.local/lib/python3.5/site-packages/chainer/training/updaters/standard_updater.py", line 194, in serialize
iterator.serialize(serializer['iterator:' + name])
File "/home/lchkhetianni/.local/lib/python3.5/site-packages/chainer/iterators/multiprocess_iterator.py", line 221, in serialize
serializer('order', order)
File "/home/lchkhetianni/.local/lib/python3.5/site-packages/chainer/serializers/npz.py", line 172, in __call__
numpy.copyto(value, dataset)
ValueError: could not broadcast input array from shape (11790) into shape (4064)
I'm interested, what is the generation time in case of this implementation, and which mel generator is the model optimized on.
Additionally, trying to understand the difference, or improvement besides the probability density into probability mass convertation.
Installed all required dependencies, and training throws this error:
Exception in main training loop: Argument 'x' has incorrect type (expected cupy.core.core.ndarray, got Variable)
Traceback (most recent call last):
File "/home/lchkhetianni/.local/lib/python3.5/site-packages/chainer/training/trainer.py", line 316, in run
update()
File "/home/lchkhetianni/.local/lib/python3.5/site-packages/chainer/training/updaters/standard_updater.py", line 170, in update
self.update_core()
File "/home/lchkhetianni/.local/lib/python3.5/site-packages/chainer/training/updaters/parallel_updater.py", line 135, in update_core
loss = loss_func(*in_arrays)
File "/home/lchkhetianni/chainer-WaveGlow/WaveGlow/model.py", line 53, in __call__
z, gaussian_nll, sum_log_s, sum_log_det_W = self._forward(x, condition)
File "/home/lchkhetianni/chainer-WaveGlow/WaveGlow/model.py", line 81, in _forward
mean=self.xp.zeros_like(z, dtype=self.xp.float32),
File "/home/lchkhetianni/.conda/envs/chainer/lib/python3.5/site-packages/cupy/creation/basic.py", line 205, in zeros_like
order, strides, memptr = _new_like_order_and_strides(a, dtype, order)
File "/home/lchkhetianni/.conda/envs/chainer/lib/python3.5/site-packages/cupy/creation/basic.py", line 35, in _new_like_order_and_strides
order = chr(_update_order_char(a, ord(order)))
Will finalize trainer extensions and updater before reraising the exception.
Traceback (most recent call last):
File "train.py", line 144, in <module>
trainer.run()
File "/home/lchkhetianni/.local/lib/python3.5/site-packages/chainer/training/trainer.py", line 349, in run
six.reraise(*exc_info)
File "/home/lchkhetianni/.conda/envs/chainer/lib/python3.5/site-packages/six.py", line 693, in reraise
raise value
File "/home/lchkhetianni/.local/lib/python3.5/site-packages/chainer/training/trainer.py", line 316, in run
update()
File "/home/lchkhetianni/.local/lib/python3.5/site-packages/chainer/training/updaters/standard_updater.py", line 170, in update
self.update_core()
File "/home/lchkhetianni/.local/lib/python3.5/site-packages/chainer/training/updaters/parallel_updater.py", line 135, in update_core
loss = loss_func(*in_arrays)
File "/home/lchkhetianni/chainer-WaveGlow/WaveGlow/model.py", line 53, in __call__
z, gaussian_nll, sum_log_s, sum_log_det_W = self._forward(x, condition)
File "/home/lchkhetianni/chainer-WaveGlow/WaveGlow/model.py", line 81, in _forward
mean=self.xp.zeros_like(z, dtype=self.xp.float32),
File "/home/lchkhetianni/.conda/envs/chainer/lib/python3.5/site-packages/cupy/creation/basic.py", line 205, in zeros_like
order, strides, memptr = _new_like_order_and_strides(a, dtype, order)
File "/home/lchkhetianni/.conda/envs/chainer/lib/python3.5/site-packages/cupy/creation/basic.py", line 35, in _new_like_order_and_strides
order = chr(_update_order_char(a, ord(order)))
TypeError: Argument 'x' has incorrect type (expected cupy.core.core.ndarray, got Variable)
How to train for multiple GPUs ?
A declarative, efficient, and flexible JavaScript library for building user interfaces.
๐ Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
An Open Source Machine Learning Framework for Everyone
The Web framework for perfectionists with deadlines.
A PHP framework for web artisans
Bring data to life with SVG, Canvas and HTML. ๐๐๐
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
Some thing interesting about web. New door for the world.
A server is a program made to process requests and deliver data to clients.
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
Some thing interesting about visualization, use data art
Some thing interesting about game, make everyone happy.
We are working to build community through open source technology. NB: members must have two-factor auth.
Open source projects and samples from Microsoft.
Google โค๏ธ Open Source for everyone.
Alibaba Open Source for everyone
Data-Driven Documents codes.
China tencent open source team.