MeTA v2.1.0 Release Notes
Release Date: 2016-02-13 // about 8 years ago-
๐ New features
- โ Add the GloVe algorithm for
training word embeddings and a library classword_embeddings
for loading and
querying trained embeddings. To facilitate returning word embeddings, a simple
util::array_view
class was added. - โ Add simple vector math library (and move
fastapprox
into themath
namespace).
๐ Bug fixes
- Fix
probe_map::extract()
forinline_key_value_storage
type; old
implementation forgot to delete all sentinel values before returning the
vector. - ๐ Fix incorrect definition of
l1norm()
insgd_model
. - ๐ Fix
gmap
calculation where 0 average precision was ignored - ๐ Fix progress output in
multiway_merge
.
โจ Enhancements
- ๐ Improve performance of
printing::progress
. Before,progress::operator()
in
๐ tight loops could dramatically hurt performance, particularly due to frequent
calls tostd::chrono::steady_clock::now()
. Now,progress::operator()
simply sets an atomic iteration counter and a background thread periodically
โก๏ธ wakes to update the progress output. - ๐ Allow full text storage in index as metadata field. If
store-full-text = true
(default false) in the corpus config, the string metadata field
"content" will be added. This is to simplify the creation of full text
๐ metadata: the user doesn't have to duplicate their dataset inmetadata.dat
,
๐ andmetadata.dat
will still be somewhat human-readable without large strings
of full text added. - ๐ Allow
make_index
to take a user-supplied corpus object.
Miscellaneous
- ZLIB is now a required dependency.
- โ
Switch to just using the standalone
./unit-test
instead ofctest
. There
โ aren't really many advantages for us to using CTest at this point with the new
โ unit test framework, so just use our unit test executable.
- โ Add the GloVe algorithm for