diff options
| author | Gertjan van den Burg <gertjanvandenburg@gmail.com> | 2019-11-20 19:14:50 +0000 |
|---|---|---|
| committer | Gertjan van den Burg <gertjanvandenburg@gmail.com> | 2019-11-20 19:14:50 +0000 |
| commit | b907bec6a3927b16016f30bc13da8155a70dd406 (patch) | |
| tree | 456cae1dcf79f581766e240b709b8e89f0f93652 | |
| parent | bugfix for kernel rbf (diff) | |
| download | gensvm-b907bec6a3927b16016f30bc13da8155a70dd406.tar.gz gensvm-b907bec6a3927b16016f30bc13da8155a70dd406.zip | |
Increase tolerance for training result
This is just too dependent on architecture/blas implementation, etc.
| -rw-r--r-- | tests/src/test_gensvm_train.c | 2 |
1 files changed, 1 insertions, 1 deletions
diff --git a/tests/src/test_gensvm_train.c b/tests/src/test_gensvm_train.c index f17cfef..118f21f 100644 --- a/tests/src/test_gensvm_train.c +++ b/tests/src/test_gensvm_train.c @@ -125,7 +125,7 @@ char *test_gensvm_train_seed_linear() mu_assert(model->m == data->m, "Incorrect model m"); mu_assert(model->K == data->K, "Incorrect model K"); - double eps = 1e-13; + double eps = 1e-7; mu_assert(fabs(matrix_get(model->V, model->K-1, 0, 0) - -1.1907736868272805) < eps, "Incorrect model->V at 0, 0"); |