Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Updating the number of queries and the seed (also README) #3

Merged
merged 3 commits into from
Apr 24, 2016
Merged

Updating the number of queries and the seed (also README) #3

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
0fa27b4
Updating the number of queries and the seed.
searchivarius Apr 24, 2016
8fa121f
Glove queries will have to be recomputed.
searchivarius Apr 24, 2016
186619d
Update README.rst
searchivarius Apr 24, 2016
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
9 changes: 7 additions & 2 deletions README.rst
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -15,7 +15,7 @@ Evaluated
* `PANNS <https://github.com/ryanrhymes/panns>`__
* `NearPy <http://nearpy.io>`__
* `KGraph <https://github.com/aaalgo/kgraph>`__
* `NonMetricSpaceLib <https://github.com/searchivarius/NonMetricSpaceLib>`__
* `NMSLIB (Non-Metric Space Library) <https://github.com/searchivarius/nmslib>`__
* `RPForest <https://github.com/lyst/rpforest>`__
* `FALCONN <http://falconn-lib.org/>`__

Expand Down Expand Up @@ -70,7 +70,12 @@ This is very much a work in progress... more results coming later!
:align: center

Note that KGraph has a substantial performance regression in the latest version.
Once the author has confirmed and fixed, I will rerun the KGraph benchmarks.
Once the author has confirmed and fixed, I will rerun the KGraph benchmarks.

Also note that NMSLIB saves indices in the directory indices.
If the tests are re-run using a different seed and/or a different number of queries, the
content of this directory should be deleted.


Testing
-------
Expand Down
6 changes: 5 additions & 1 deletion ann_benchmarks/__init__.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -354,7 +354,11 @@ def get_dataset(which='glove', limit=-1):
X = numpy.vstack(X)
import sklearn.cross_validation

X_train, X_test = sklearn.cross_validation.train_test_split(X, test_size=1000, random_state=42)
# Here Eric is most welcome to use any other random_state
# Last time, Leo was testing using random_state==1
# However, it is best to use a new random seed for each major re-evaluation,
# so that we test on a trully bind data.
X_train, X_test = sklearn.cross_validation.train_test_split(X, test_size=10000, random_state=2016)
print(X_train.shape, X_test.shape)
return X_train, X_test

Expand Down
Loading