add some began samples.

examples/GAN/BEGAN.py

@@ -17,6 +17,8 @@ from GAN import GANModelDesc, GANTrainer
 """
 Boundary Equilibrium GAN.
 See the docstring in DCGAN.py for usage.
+
+A pretrained model on CelebA is at https://drive.google.com/open?id=0B5uDfUQ1JTglUmgyZV8zQmNOTVU
 """
 
 
@@ -89,7 +91,6 @@ class Model(GANModelDesc):
         def summary_image(name, x):
             x = (x + 1.0) * 128.0
             x = tf.clip_by_value(x, 0, 255)
-            x = tf.cast(x, tf.uint8)
             tf.summary.image(name, x, max_outputs=30)
 
         with argscope([Conv2D, FullyConnected],
@@ -153,7 +154,7 @@ def get_config():
 if __name__ == '__main__':
     args = DCGAN.get_args()
     if args.sample:
-        DCGAN.sample(args.load)
+        DCGAN.sample(args.load, 'gen/conv4.3/output')
     else:
         assert args.data
         logger.auto_set_dir()

examples/GAN/DCGAN.py

@@ -4,6 +4,7 @@
 # Author: Yuxin Wu <ppwwyyxxc@gmail.com>
 
 import glob
+import numpy as np
 import os, sys
 import argparse
 
@@ -128,16 +129,17 @@ def get_config():
     )
 
 
-def sample(model_path):
+def sample(model_path, output_name='gen/gen'):
     pred = PredictConfig(
         session_init=get_model_loader(model_path),
         model=Model(),
         input_names=['z'],
-        output_names=['gen/gen', 'z'])
-    pred = SimpleDatasetPredictor(pred, RandomZData((100, 100)))
+        output_names=[output_name, 'z'])
+    pred = SimpleDatasetPredictor(pred, RandomZData((100, opt.Z_DIM)))
     for o in pred.get_result():
         o, zs = o[0] + 1, o[1]
         o = o * 128.0
+        o = np.clip(o, 0, 255)
         o = o[:, :, :, ::-1]
         viz = stack_patches(o, nr_row=10, nr_col=10, viz=True)
 

examples/GAN/README.md

@@ -26,11 +26,11 @@ Reproduce DCGAN following the setup in [dcgan.torch](https://github.com/soumith/
 
 + Generated samples
 
-![sample](demo/CelebA-samples.jpg)
+![sample](demo/DCGAN-CelebA-samples.jpg)
 
 + Vector arithmetic: smiling woman - neutral woman + neutral man = smiling man
 
-![vec](demo/CelebA-vec.jpg)
+![vec](demo/DCGAN-CelebA-vec.jpg)
 
 ## Image2Image.py
 
@@ -61,6 +61,9 @@ Train a simple GAN on mnist, conditioned on the class labels.
 ## WGAN.py, Improved-WGAN.py, BEGAN.py
 
 These variants are implemented by some small modifications on top of DCGAN.py.
+Some BEGAN samples:
+
+![began-sample](demo/BEGAN-CelebA-samples.jpg)
 
 ## DiscoGAN-CelebA.py
 

tensorpack/models/pool.py

@@ -209,7 +209,7 @@ class TestPool(TestModel):
         self.assertTrue((res == 0).all())
 
     def test_BilinearUpSample(self):
-        h, w = 5, 5
+        h, w = 12, 12
         scale = 2
 
         mat = np.random.rand(h, w).astype('float32')
@@ -224,11 +224,6 @@ class TestPool(TestModel):
 
         diff = np.abs(res2 - res)
 
-        # TODO not equivalent to rescale on edge?
-        #diff[0, :] = 0
-        #diff[-1, :] = 0
-        #diff[:, 0] = 0
-        #diff[:, -1] = 0
         # if not diff.max() < 1e-4:
         #     import IPython
         #     IPython.embed(config=IPython.terminal.ipapp.load_default_config())