def testImagenetMultiResolutionPreprocessExample(self, resize_method): example = {"inputs": tf.random_uniform([64, 64, 3], minval=-1.)} mode = tf.estimator.ModeKeys.TRAIN hparams = HParams(resolutions=[8, 16, 32]) if resize_method is not None: hparams.resize_method = resize_method problem = imagenet.ImageImagenetMultiResolutionGen() preprocessed_example = problem.preprocess_example( example, mode, hparams) self.assertLen(preprocessed_example, 1) self.assertEqual(preprocessed_example["inputs"].shape, (42, 32, 3))
def testCelebaMultiResolutionPreprocessExample(self, resize_method): example = {"inputs": tf.random_uniform([218, 178, 3], minval=-1.)} mode = tf.estimator.ModeKeys.TRAIN hparams = HParams(resolutions=[8, 16, 32]) if resize_method is not None: hparams.resize_method = resize_method problem = celeba.ImageCelebaMultiResolution() preprocessed_example = problem.preprocess_example( example, mode, hparams) self.assertLen(preprocessed_example, 2) self.assertEqual(preprocessed_example["inputs"].shape, (138, 138, 3)) self.assertEqual(preprocessed_example["targets"].shape, (42, 32, 3))