saymrwulf/onnxruntime
1
0
Fork 0
mirror of https://github.com/saymrwulf/onnxruntime.git synced 2026-07-02 03:55:34 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions

CUDA/CPU test for NegativeLogLikelihoodLoss(12) function based loss operator. (#3793)

Browse source
This commit is contained in:
M. Zeeshan Siddiqui 2020-05-01 21:36:29 -07:00 committed by GitHub
parent b1c4d6ff4e
commit 4f9f6aedea
No known key found for this signature in database
GPG key ID: 4AEE18F83AFDEB23
1 changed files with 110 additions and 0 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

110
orttraining/orttraining/test/training_ops/cuda/negativeloglikelihood_test.cc Normal file
View file

@ -0,0 +1,110 @@
// Copyright (c) Microsoft Corporation. All rights reserved.
// Licensed under the MIT License.
#include "test/providers/compare_provider_test_utils.h"
using namespace std;
namespace onnxruntime {
namespace test {
static void TestNegativeLogLikelihoodLoss(const std::vector<int64_t>* X_dims,
const std::vector<int64_t>* index_dims,
const std::vector<int64_t>* weight_dims,
const std::vector<int64_t>* Y_dims,
const std::string& reduction,
const std::int64_t ignore_index = -1) {
CompareOpTester test("NegativeLogLikelihoodLoss", 12, onnxruntime::kOnnxDomain);
test.AddAttribute("reduction", reduction);
test.AddAttribute("ignore_index", ignore_index);
// create rand inputs
RandomValueGenerator random{};
std::vector<float> X_data = random.Uniform<float>(*X_dims, -200.0f, 200.0f);
std::vector<int64_t> index_data = random.Uniform<int64_t>(*index_dims, 0, (*X_dims)[1]);
//Add one data point that has ignore_index.
if (index_data.size() > 0) {
index_data[0] = ignore_index;
}
test.AddInput<float>("X", *X_dims, X_data);
test.AddInput<int64_t>("index", *index_dims, index_data);
if (weight_dims) {
std::vector<float> weight_data = random.Uniform<float>(*weight_dims, 0.0f, 1.0f);
test.AddInput<float>("weight", *weight_dims, weight_data);
}
std::vector<float> Y_data = FillZeros<float>(*Y_dims);
test.AddOutput<float>("output", *Y_dims, Y_data);
test.CompareWithCPU(kCudaExecutionProvider);
}
TEST(CudaKernelTest, NegativeLogLikelihoodLoss_TinySizeTensor) {
std::vector<int64_t> X_dims{8, 2};
std::vector<int64_t> index_dims{8};
std::vector<int64_t> weight_dims{2};
std::vector<int64_t> Y_dims{};
std::vector<int64_t> Y_dims_none{8};
TestNegativeLogLikelihoodLoss(&X_dims, &index_dims, &weight_dims, &Y_dims, "mean");
TestNegativeLogLikelihoodLoss(&X_dims, &index_dims, nullptr, &Y_dims, "mean");
TestNegativeLogLikelihoodLoss(&X_dims, &index_dims, &weight_dims, &Y_dims, "sum");
TestNegativeLogLikelihoodLoss(&X_dims, &index_dims, nullptr, &Y_dims, "sum");
TestNegativeLogLikelihoodLoss(&X_dims, &index_dims, &weight_dims, &Y_dims_none, "none");
TestNegativeLogLikelihoodLoss(&X_dims, &index_dims, nullptr, &Y_dims_none, "none");
// Just test ignore_index for small tensor because it will increase test time a lot with little verification gain.
TestNegativeLogLikelihoodLoss(&X_dims, &index_dims, &weight_dims, &Y_dims, "mean", 0);
TestNegativeLogLikelihoodLoss(&X_dims, &index_dims, nullptr, &Y_dims, "mean", 0);
TestNegativeLogLikelihoodLoss(&X_dims, &index_dims, &weight_dims, &Y_dims, "sum", 0);
TestNegativeLogLikelihoodLoss(&X_dims, &index_dims, nullptr, &Y_dims, "sum", 0);
TestNegativeLogLikelihoodLoss(&X_dims, &index_dims, &weight_dims, &Y_dims_none, "none", 0);
TestNegativeLogLikelihoodLoss(&X_dims, &index_dims, nullptr, &Y_dims_none, "none", 0);
}
TEST(CudaKernelTest, NegativeLogLikelihoodLoss_SmallSizeTensor) {
std::vector<int64_t> X_dims{8, 20, 10};
std::vector<int64_t> index_dims{8, 10};
std::vector<int64_t> weight_dims{20};
std::vector<int64_t> Y_dims{};
std::vector<int64_t> Y_dims_none{8, 10};
TestNegativeLogLikelihoodLoss(&X_dims, &index_dims, &weight_dims, &Y_dims, "mean");
TestNegativeLogLikelihoodLoss(&X_dims, &index_dims, nullptr, &Y_dims, "mean");
TestNegativeLogLikelihoodLoss(&X_dims, &index_dims, &weight_dims, &Y_dims, "sum");
TestNegativeLogLikelihoodLoss(&X_dims, &index_dims, nullptr, &Y_dims, "sum");
TestNegativeLogLikelihoodLoss(&X_dims, &index_dims, &weight_dims, &Y_dims_none, "none");
TestNegativeLogLikelihoodLoss(&X_dims, &index_dims, nullptr, &Y_dims_none, "none");
}
TEST(CudaKernelTest, NegativeLogLikelihoodLoss_MediumSizeTensor) {
std::vector<int64_t> X_dims{8, 1024};
std::vector<int64_t> index_dims{8};
std::vector<int64_t> weight_dims{1024};
std::vector<int64_t> Y_dims{};
std::vector<int64_t> Y_dims_none{8};
TestNegativeLogLikelihoodLoss(&X_dims, &index_dims, &weight_dims, &Y_dims, "mean");
TestNegativeLogLikelihoodLoss(&X_dims, &index_dims, nullptr, &Y_dims, "mean");
TestNegativeLogLikelihoodLoss(&X_dims, &index_dims, &weight_dims, &Y_dims, "sum");
TestNegativeLogLikelihoodLoss(&X_dims, &index_dims, nullptr, &Y_dims, "sum");
TestNegativeLogLikelihoodLoss(&X_dims, &index_dims, &weight_dims, &Y_dims_none, "none");
TestNegativeLogLikelihoodLoss(&X_dims, &index_dims, nullptr, &Y_dims_none, "none");
}
TEST(CudaKernelTest, NegativeLogLikelihoodLoss_LargeSizeTensor) {
std::vector<int64_t> X_dims{4, 512, 30528};
std::vector<int64_t> index_dims{4, 30528};
std::vector<int64_t> weight_dims{512};
std::vector<int64_t> Y_dims{};
std::vector<int64_t> Y_dims_none{4, 30528};
TestNegativeLogLikelihoodLoss(&X_dims, &index_dims, &weight_dims, &Y_dims, "mean");
TestNegativeLogLikelihoodLoss(&X_dims, &index_dims, nullptr, &Y_dims, "mean");
TestNegativeLogLikelihoodLoss(&X_dims, &index_dims, &weight_dims, &Y_dims, "sum");
TestNegativeLogLikelihoodLoss(&X_dims, &index_dims, nullptr, &Y_dims, "sum");
TestNegativeLogLikelihoodLoss(&X_dims, &index_dims, &weight_dims, &Y_dims_none, "none");
TestNegativeLogLikelihoodLoss(&X_dims, &index_dims, nullptr, &Y_dims_none, "none");
}
} // namespace test
} // namespace onnxruntime