make the pix2pix support multi gpu trainning (#884)

* make the pix2pix support multi gpu trainning

* Update Image2Image.py

* Update Image2Image.py

* Update Image2Image.py

examples/GAN/Image2Image.py

@@ -12,10 +12,11 @@ import argparse
 
 
 from tensorpack import *
+from tensorpack.utils.gpu import get_num_gpu
 from tensorpack.utils.viz import stack_patches
 from tensorpack.tfutils.summary import add_moving_summary
 from tensorpack.tfutils.scope_utils import auto_reuse_variable_scope
-from GAN import GANTrainer, GANModelDesc
+from GAN import GANTrainer, MultiGPUGANTrainer, GANModelDesc
 
 """
 To train Image-to-Image translation model with image pairs:
@@ -218,7 +219,13 @@ if __name__ == '__main__':
 
         data = QueueInput(get_data())
 
-        GANTrainer(data, Model()).train_with_defaults(
+        nr_tower = max(get_num_gpu(), 1)
+        if nr_tower == 1:
+            trainer = GANTrainer(data, Model())
+        else:
+            trainer = MultiGPUGANTrainer(nr_tower, data, Model())
+
+        trainer.train_with_defaults(
             callbacks=[
                 PeriodicTrigger(ModelSaver(), every_k_epochs=3),
                 ScheduledHyperParamSetter('learning_rate', [(200, 1e-4)])