projects / platform / upstream / tvm.git / commit
? search:
summary | shortlog | log | commit | commitdiff | tree
(parent: 518c3fd) | patch
[Quantization] Make calibration faster and more memory usage friendly (#4589)
authormasahi <masahi129@gmail.com>
Fri, 3 Jan 2020 12:23:09 +0000 (21:23 +0900)
committerWuwei Lin <wuwei@apache.org>
Fri, 3 Jan 2020 12:23:09 +0000 (21:23 +0900)
commit2440c9ced823b449dcb5a718bc485d34f80191d3
treee2aa1ba3d9891c4c8c048b66a0f75e6557b7f468tree | snapshot
parent518c3fd0cf5f5a98245d4f4566bd0280b2a304c5commit | diff
[Quantization] Make calibration faster and more memory usage friendly (#4589)

* Use memory efficient calibrate

* Fixed indexing

* add cpp kl stub

* ported KL cpp from mxnet

* Fixed std::distance arguments order

* remove python implementation

* fix lint and indent

* fix indent

* refactoring

* fix lint

* fix for i386
python/tvm/relay/quantize/_calibrate.py diff | blob | history
python/tvm/relay/quantize/kl_divergence.py diff | blob | history
python/tvm/relay/quantize/quantize.py diff | blob | history
src/relay/pass/quantize/calibrate.cc diff | blob | history
src/relay/pass/quantize/quantize.h diff | blob | history
tests/python/relay/test_pass_auto_quantize.py diff | blob | history
Domain: Machine Learning / ML Framework;