def _test_result_1d(self, dtype):
A = np.array([[1, 2, 3, 4, 5]],dtype=dtype)
tensor = ops.median3x3_downsample(A)
B = tensor.eval()
correct = np.array([[1, 3, 5]],dtype=dtype)
self.assertAllEqual(B, correct)
tensor = ops.median3x3_downsample(A.transpose())
B = tensor.eval()
self.assertAllEqual(B, correct.transpose())
def test_gpu_equals_cpu(self):
if not tf.test.is_gpu_available():
return
for dtype in TYPES:
A = np.random.rand(10,13).astype(dtype)
with self.test_session(use_gpu=False, force_gpu=False):
tensor_cpu = ops.median3x3_downsample(A)
result_cpu = tensor_cpu.eval()
with self.test_session(use_gpu=True, force_gpu=True):
tensor_gpu = ops.median3x3_downsample(A)
result_gpu = tensor_cpu.eval()
self.assertAllEqual(result_cpu, result_gpu)
def _test_single_element(self, dtype):
A = np.array([[1]], dtype=dtype)
tensor = ops.median3x3_downsample(A)
B = tensor.eval()
#print(A,B)
self.assertAllEqual(B, A)