1# Copyright 2019 The TensorFlow Authors. All Rights Reserved. 2# 3# Licensed under the Apache License, Version 2.0 (the "License"); 4# you may not use this file except in compliance with the License. 5# You may obtain a copy of the License at 6# 7# http://www.apache.org/licenses/LICENSE-2.0 8# 9# Unless required by applicable law or agreed to in writing, software 10# distributed under the License is distributed on an "AS IS" BASIS, 11# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. 12# See the License for the specific language governing permissions and 13# limitations under the License. 14# ============================================================================== 15"""Test configs for fused_batch_norm.""" 16from __future__ import absolute_import 17from __future__ import division 18from __future__ import print_function 19 20import tensorflow.compat.v1 as tf 21from tensorflow.lite.testing.zip_test_utils import create_tensor_data 22from tensorflow.lite.testing.zip_test_utils import make_zip_of_tests 23from tensorflow.lite.testing.zip_test_utils import register_make_test_function 24 25 26@register_make_test_function() 27def make_fused_batch_norm_tests(options): 28 """Make a set of tests to do fused_batch_norm.""" 29 30 test_parameters = [{ 31 "dtype": [tf.float32], 32 "input_shape": [[1, 1, 6, 2]], 33 "epsilon": [0.001, 0.1], 34 "is_training": [False], 35 }] 36 37 # Training support in MLIR converter. 38 if options.use_experimental_converter: 39 test_parameters = test_parameters + [{ 40 "dtype": [tf.float32], 41 "input_shape": [[1, 1, 6, 2]], 42 "epsilon": [0.001, 0.1], 43 "is_training": [True], 44 }] 45 46 def build_graph(parameters): 47 """Build the testing graph for fused batch normalization.""" 48 input_shape = parameters["input_shape"] 49 scale_shape = input_shape[3] 50 51 scale = create_tensor_data(parameters["dtype"], scale_shape) 52 offset = create_tensor_data(parameters["dtype"], scale_shape) 53 mean = create_tensor_data(parameters["dtype"], scale_shape) 54 variance = create_tensor_data(parameters["dtype"], scale_shape) 55 56 x = tf.compat.v1.placeholder( 57 dtype=parameters["dtype"], name="x", shape=parameters["input_shape"]) 58 [x_norm, _, _] = tf.compat.v1.nn.fused_batch_norm( 59 x, 60 scale, 61 offset, 62 mean, 63 variance, 64 parameters["epsilon"], 65 data_format="NHWC", 66 is_training=parameters["is_training"]) 67 68 input_tensor = tf.compat.v1.placeholder( 69 dtype=parameters["dtype"], 70 name="input", 71 shape=parameters["input_shape"]) 72 out = tf.add(input_tensor, x_norm) 73 return [x, input_tensor], [out] 74 75 def build_inputs(parameters, sess, inputs, outputs): 76 input_values = [ 77 create_tensor_data(parameters["dtype"], parameters["input_shape"]), 78 create_tensor_data(parameters["dtype"], parameters["input_shape"]) 79 ] 80 81 return input_values, sess.run( 82 outputs, feed_dict=dict(zip(inputs, input_values))) 83 84 make_zip_of_tests(options, test_parameters, build_graph, build_inputs) 85