1 /* Copyright 2016 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 #include "tensorflow/core/kernels/fractional_pool_common.h" 16 17 #include "tensorflow/core/lib/random/simple_philox.h" 18 19 namespace tensorflow { 20 static std::vector<int64> GeneratePoolingSequencePseudoRandom( 21 int input_length, int output_length, GuardedPhiloxRandom* generator) { 22 std::vector<int64> cum_seq(output_length + 1, 0); 23 std::vector<int64> diff(output_length, 0); 24 25 double alpha = static_cast<double>(input_length) / output_length; 26 int k = input_length / output_length; 27 28 // In the paper [1], author proposes the following procedure to generate a 29 // pseudo random pooling region: 30 // 1) Set a_0 = 1, a_Nout = Nin; 31 // 2) a_i = ceil(alpha*(u+i)) 32 // in which, i = 1, 2, ... , Nout-1 33 // u is a random number in (0,1) for all i 34 // alpha = Nin/Nout in (1,2) 35 // The sequence {a_i} should satisfy a_i-a_{i-1} = 1 or 2 36 // Note: for step 1), it makes more sense to make a_Nout = Nin+1, that way, 37 // a_i-a_{i-1} = 1 or 2 is also true for i = Nout. 38 // 39 // However, there are choices of alpha and u that will make 40 // a_i - a_{i-1} > 2. This happens at the left boundary. For example, with 41 // alpha = 1.732, u = 0.8, then a_1 = 4, a_1-a_0 = 3. 42 // This is why u_max1 is needed, i.e. u is a random number in (0,u_max1) 43 // instead of (0,1). 44 // Define k = ceil(alpha)-1, then we require: 45 // a_1 = alpha*(u+1) <= a_0+(k+1) 46 // ===> This gives u_max1 = (k+2)/alpha - 1. 47 // 48 // In addition, when extending the pooling sequence generation process for 49 // alpha beyond (1,2), e.g. (k,k+1); a check on the right boundary is also 50 // needed to make sure the last gap a_Nout-a_{Nout-1} >= k. Solving it gives 51 // u_max2 = (Nin+1-k)/alpha - (Nout-1) 52 // Here is an example where u > u_max2, alpha = 2.3, u = 0.7, u_max2 = 0.565, 53 // Nin = 23, Nout = 10; the sequence 54 // from a_0 to a_10 is: 55 // [1, 4, 7, 9, 11, 14, 16, 18, 21, 23, 24] 56 // The last gap is only 1. 57 // 58 // [1]: https://arxiv.org/abs/1412.6071 59 double u_max1 = (k + 2) / alpha - 1; 60 double u_max2 = (input_length + 1 - k) / alpha - (output_length - 1); 61 double max_u = std::min(u_max1, u_max2); 62 63 // Generate random number in parallel. 64 auto local_gen = generator->ReserveSamples32(2); 65 random::SimplePhilox random(&local_gen); 66 const double u = random.RandDouble() * max_u; 67 68 cum_seq[0] = 1; 69 cum_seq[output_length] = input_length + 1; 70 for (int i = 1; i < output_length; ++i) { 71 cum_seq[i] = static_cast<int>(ceil(alpha * (i + u))); 72 } 73 74 for (int i = 0; i < output_length; ++i) { 75 diff[i] = cum_seq[i + 1] - cum_seq[i]; 76 } 77 78 return diff; 79 } 80 81 static std::vector<int64> GeneratePoolingSequenceRandom( 82 int input_length, int output_length, GuardedPhiloxRandom* generator) { 83 int k = input_length / output_length; 84 int num_random_spot = input_length % output_length; 85 std::vector<int64> diff(output_length, k); 86 87 for (int i = 0; i < num_random_spot; ++i) { 88 diff[i] += 1; 89 } 90 91 // Randomly shuffle this vector. 92 auto local_gen = generator->ReserveSamples32(diff.size()); 93 random::SingleSampleAdapter<random::PhiloxRandom> single(&local_gen); 94 const auto uniform = [&single](uint32 n) { return single() % n; }; 95 RandomShuffle(diff.begin(), diff.end(), uniform); 96 97 return diff; 98 } 99 100 std::vector<int64> GeneratePoolingSequence(int input_length, int output_length, 101 GuardedPhiloxRandom* generator, 102 bool pseudo_random) { 103 std::vector<int64> diff; 104 // This is a case that regular pooling can handle, just return diff with 105 // each element input_length/output_length. 106 if (input_length % output_length == 0) { 107 diff = std::vector<int64>(output_length, input_length / output_length); 108 } 109 110 if (pseudo_random) { 111 diff = GeneratePoolingSequencePseudoRandom(input_length, output_length, 112 generator); 113 } else { 114 diff = 115 GeneratePoolingSequenceRandom(input_length, output_length, generator); 116 } 117 118 // Sanity check. 119 int k = input_length / output_length; 120 for (int i = 0; i < output_length; ++i) { 121 // k<= diff[i] <= k+1. 122 DCHECK_GE(diff[i], k); 123 DCHECK_LE(diff[i], k + 1); 124 } 125 126 // Return cumulative sequence. 127 std::vector<int64> cum_seq(output_length + 1, 0); 128 for (int i = 1; i < cum_seq.size(); ++i) { 129 cum_seq[i] = cum_seq[i - 1] + diff[i - 1]; 130 } 131 return cum_seq; 132 } 133 134 } // namespace tensorflow 135
