Update generator_model.py

[Uha661]

No description provided.

[Uha661]

StopIteration Traceback (most recent call last)
in ()
9 generator_network, discriminator_network, Gs_network = pickle.load(f)
10
---> 11 generator = Generator(Gs_network, batch_size=1, randomize_noise=False)
12
13

/content/stylegan-encoder/encoder/generator_model.py in init(self, model, batch_size, randomize_noise)
30 self.graph = tf.get_default_graph()
31
---> 32 self.dlatent_variable = next(v for v in tf.global_variables() if 'learnable_dlatents' in v.name)
33 self.set_dlatents(self.initial_dlatents)
34

StopIteration:

[progmars]

The fix leads to error:

 File "encode_images.py", line 84, in <module>
    main()
  File "encode_images.py", line 57, in main
    generator = Generator(Gs_network, args.batch_size, randomize_noise=args.randomize_noise)
  File "D:\GAN\stylegan\encoder\generator_model.py", line 35, in __init__
    self.set_dlatents(self.initial_dlatents)
  File "D:\GAN\stylegan\encoder\generator_model.py", line 46, in set_dlatents
    self.sess.run(tf.assign(self.dlatent_variable, dlatents))
  File "D:\GAN\Anaconda\lib\site-packages\tensorflow\python\ops\state_ops.py", line 224, in assign
    if ref.dtype._is_ref_dtype:
AttributeError: 'generator' object has no attribute 'dtype'

And it seems to be unnecessary at all. I found that StopIteration error occurs only if you use Nvlabs dmlib instead of this repository dmlib because this one has modifications to pass custom arguments to the run method.