nsynth_generate fails with Abseil error (audio generation docs need updating)

[9600]

Followed audio generation instructions and everything went fine until I got to executing nsynth_generate which, with or without arguments, fails with:

andrew@blast:/data/Liles/nsynth/open-nsynth-super/audio/workdir$ nsynth_generate WARNING:tensorflow:From /usr/local/lib/python2.7/dist-packages/magenta/models/nsynth/wavenet/masked.py:116: init (from tensorflow.python.ops.init_ops) is deprecated and will be removed in a future version. Instructions for updating: Use tf.initializers.variance_scaling instead with distribution=uniform to get equivalent behavior. Traceback (most recent call last): File "/usr/local/bin/nsynth_generate", line 7, in from magenta.models.nsynth.wavenet.nsynth_generate import console_entry_point File "/usr/local/lib/python2.7/dist-packages/magenta/models/nsynth/wavenet/nsynth_generate.py", line 28, in tf.app.flags.DEFINE_string("npy_only", False, "If True, use only .npy files.") File "/usr/local/lib/python2.7/dist-packages/tensorflow/python/platform/flags.py", line 58, in wrapper return original_function(*args, **kwargs) File "/usr/local/lib/python2.7/dist-packages/absl/flags/_defines.py", line 241, in DEFINE_string DEFINE(parser, name, default, help, flag_values, serializer, **args) File "/usr/local/lib/python2.7/dist-packages/absl/flags/_defines.py", line 81, in DEFINE DEFINE_flag(_flag.Flag(parser, serializer, name, default, help, **args), File "/usr/local/lib/python2.7/dist-packages/absl/flags/_flag.py", line 107, in init self._set_default(default) File "/usr/local/lib/python2.7/dist-packages/absl/flags/_flag.py", line 196, in _set_default self.default = self._parse(value) File "/usr/local/lib/python2.7/dist-packages/absl/flags/_flag.py", line 169, in _parse 'flag --%s=%s: %s' % (self.name, argument, e)) absl.flags._exceptions.IllegalFlagValueError: flag --npy_only=False: flag value must be a string, found "<type 'bool'>"

Any suggestions as to a remedy would be much appreciated!

[9600]

@jesseengel any thoughts on this?

[9600]

@nsynthsuper ?

[jesseengel]

Sorry it took so long to respond. I just tested with head of magenta and nsynth_generate seems to work fine.

I think the problem is that directions for setting up the environment in https://github.com/googlecreativelab/open-nsynth-super/tree/master/audio are old now. It's actually pulling from a stale branch of my personal repo.

It should be sufficient to just follow the install istructions at https://github.com/tensorflow/magenta and do pip install magenta-gpu or pip install magenta and then pip install tensorflow-gpu.

@nsynthsuper, want to verify and update the docs?

[9600]

Many thanks, this did the trick!