def testExecuteMultipleNonListOutput(self): x = tensor.Tensor([1, 2, 3, 4, 5, 6]) y = tensor.Tensor([1, 3, 5]) result = array_ops.listdiff(x, y) out, idx = result self.assertTrue(out is result.out) self.assertTrue(idx is result.idx) self.assertAllEqual([2, 4, 6], out.numpy()) self.assertAllEqual([1, 3, 5], idx.numpy())
def testExecuteMultipleNonListOutput(self): x = constant_op.constant([1, 2, 3, 4, 5, 6]) y = constant_op.constant([1, 3, 5]) result = array_ops.listdiff(x, y) out, idx = result self.assertTrue(out is result.out) self.assertTrue(idx is result.idx) self.assertAllEqual([2, 4, 6], out) self.assertAllEqual([1, 3, 5], idx)
def _testListDiff(self, x, y, out, idx): for dtype in [dtypes.int32, dtypes.int64]: for index_dtype in [dtypes.int32, dtypes.int64]: with self.cached_session(): x_tensor = ops.convert_to_tensor(x, dtype=dtype) y_tensor = ops.convert_to_tensor(y, dtype=dtype) with self.test_scope(): out_tensor, idx_tensor = array_ops.listdiff( x_tensor, y_tensor, out_idx=index_dtype) tf_out, tf_idx = self.evaluate([out_tensor, idx_tensor]) self.assertAllEqual(out, tf_out) self.assertAllEqual(idx, tf_idx) self.assertEqual(1, out_tensor.get_shape().ndims) self.assertEqual(1, idx_tensor.get_shape().ndims)
def _ProdGrad(op, grad): """Gradient for Prod.""" # The gradient can be expressed by dividing the product by each entry of the # input tensor, but this approach can't deal with zeros in the input. # Here, we avoid this problem by composing the output as a product of two # cumprod operations. input_shape = array_ops.shape(op.inputs[0]) # Reshape reduction indices for the case where the parameter is a scalar reduction_indices = array_ops.reshape(op.inputs[1], [-1]) # Expand grad to full input shape output_shape_kept_dims = math_ops.reduced_shape(input_shape, op.inputs[1]) tile_scaling = _safe_shape_div(input_shape, output_shape_kept_dims) grad = array_ops.reshape(grad, output_shape_kept_dims) grad = array_ops.tile(grad, tile_scaling) # Pack all reduced dimensions into a single one, so we can perform the # cumprod ops. If the reduction dims list is empty, it defaults to float32, # so we need to cast here. We put all the shape-related ops on CPU to avoid # copying back and forth, and since listdiff is CPU only. with ops.device("/cpu:0"): reduced = math_ops.cast(reduction_indices, dtypes.int32) idx = math_ops.range(0, array_ops.rank(op.inputs[0])) other, _ = array_ops.listdiff(idx, reduced) perm = array_ops.concat(0, [reduced, other]) reduced_num = math_ops.reduce_prod( array_ops.gather(input_shape, reduced)) other_num = math_ops.reduce_prod(array_ops.gather(input_shape, other)) permuted = array_ops.transpose(op.inputs[0], perm) permuted_shape = array_ops.shape(permuted) reshaped = array_ops.reshape(permuted, (reduced_num, other_num)) # Calculate product, leaving out the current entry left = math_ops.cumprod(reshaped, axis=0, exclusive=True) right = math_ops.cumprod(reshaped, axis=0, exclusive=True, reverse=True) y = array_ops.reshape(left * right, permuted_shape) # Invert the transpose and reshape operations. # Make sure to set the statically known shape information through a reshape. out = grad * array_ops.transpose(y, array_ops.invert_permutation(perm)) return array_ops.reshape(out, input_shape), None
def _ProdGrad(op, grad): """Gradient for Prod.""" # The gradient can be expressed by dividing the product by each entry of the # input tensor, but this approach can't deal with zeros in the input. # Here, we avoid this problem by composing the output as a product of two # cumprod operations. input_shape = array_ops.shape(op.inputs[0]) # Reshape reduction indices for the case where the parameter is a scalar reduction_indices = array_ops.reshape(op.inputs[1], [-1]) # Expand grad to full input shape output_shape_kept_dims = math_ops.reduced_shape(input_shape, op.inputs[1]) tile_scaling = _safe_shape_div(input_shape, output_shape_kept_dims) grad = array_ops.reshape(grad, output_shape_kept_dims) grad = array_ops.tile(grad, tile_scaling) # Pack all reduced dimensions into a single one, so we can perform the # cumprod ops. If the reduction dims list is empty, it defaults to float32, # so we need to cast here. We put all the shape-related ops on CPU to avoid # copying back and forth, and since listdiff is CPU only. with ops.device("/cpu:0"): reduced = math_ops.cast(reduction_indices, dtypes.int32) idx = math_ops.range(0, array_ops.rank(op.inputs[0])) other, _ = array_ops.listdiff(idx, reduced) perm = array_ops.concat(0, [reduced, other]) reduced_num = math_ops.reduce_prod(array_ops.gather(input_shape, reduced)) other_num = math_ops.reduce_prod(array_ops.gather(input_shape, other)) permuted = array_ops.transpose(op.inputs[0], perm) permuted_shape = array_ops.shape(permuted) reshaped = array_ops.reshape(permuted, (reduced_num, other_num)) # Calculate product, leaving out the current entry left = math_ops.cumprod(reshaped, axis=0, exclusive=True) right = math_ops.cumprod(reshaped, axis=0, exclusive=True, reverse=True) y = array_ops.reshape(left * right, permuted_shape) # Invert the transpose and reshape operations. # Make sure to set the statically known shape information through a reshape. out = grad * array_ops.transpose(y, array_ops.invert_permutation(perm)) return array_ops.reshape(out, input_shape), None