This is an automated email from the ASF dual-hosted git repository.
krickert pushed a change to branch feat/embeddings-timings-tests
in repository https://gitbox.apache.org/repos/asf/opennlp.git
from c9c27d992 GH-1471: Add a JMH benchmark for SentenceVectorsDL and keep
the tensors off the heap
add e650e7ea6 GH-1471: Address review on the embedding benchmark
No new revisions were added by this update.
Summary of changes:
opennlp-core/opennlp-ml/opennlp-dl/BENCHMARKS.md | 49 ++++-
opennlp-core/opennlp-ml/opennlp-dl/pom.xml | 69 +------
.../opennlp/dl/vectors/NativeMemoryProfiler.java | 108 +++++++++++
.../dl/vectors/SentenceVectorsDLBenchmark.java | 90 +++++++--
.../java/opennlp/dl/vectors/SentenceVectorsDL.java | 205 +++++++++++++--------
.../vectors/SentenceVectorsDLDirectMemoryTest.java | 166 +++++++++++++++++
...iny_vectors_model.py => gen_tiny_wide_model.py} | 25 ++-
.../resources/opennlp/dl/vectors/tiny-wide.onnx | Bin 0 -> 4463 bytes
opennlp-core/opennlp-runtime/pom.xml | 66 -------
pom.xml | 65 +++++++
10 files changed, 592 insertions(+), 251 deletions(-)
create mode 100644
opennlp-core/opennlp-ml/opennlp-dl/src/jmh/java/opennlp/dl/vectors/NativeMemoryProfiler.java
create mode 100644
opennlp-core/opennlp-ml/opennlp-dl/src/test/java/opennlp/dl/vectors/SentenceVectorsDLDirectMemoryTest.java
copy
opennlp-core/opennlp-ml/opennlp-dl/src/test/resources/opennlp/dl/vectors/{gen_tiny_vectors_model.py
=> gen_tiny_wide_model.py} (67%)
create mode 100644
opennlp-core/opennlp-ml/opennlp-dl/src/test/resources/opennlp/dl/vectors/tiny-wide.onnx
