Performance: upgrade to latest ann-benchmarks (alexklibisz#652)

.gitmodules

@@ -1,3 +1,4 @@
 [submodule "ann-benchmarks/ann-benchmarks"]
 	path = ann-benchmarks/ann-benchmarks
 	url = https://github.com/alexklibisz/ann-benchmarks.git
+	branch = fix-get-algorithm-definitions

Taskfile.yaml

@@ -53,7 +53,7 @@ tasks:
       OBJC_DISABLE_INITIALIZE_FORK_SAFETY: "YES"
       RESULTS_DIR: ../../docs/pages/performance/fashion-mnist
     cmds:
-      - venv/bin/python run.py --dataset fashion-mnist-784-euclidean --algorithm elastiknn-l2lsh --runs 3 --count 100 --parallelism 1 --force --local
+      - venv/bin/python run.py --dataset fashion-mnist-784-euclidean --algorithm elastiknn-l2lsh --run-disabled --runs 3 --count 100 --parallelism 1 --force --local
       - mkdir -p $RESULTS_DIR
       - venv/bin/python plot.py --dataset fashion-mnist-784-euclidean --count 100 --output $RESULTS_DIR/plot.png | venv/bin/python ../parse_results.py > $RESULTS_DIR/results.md
       - base64 -w 0 -i $RESULTS_DIR/plot.png > $RESULTS_DIR/plot.b64
@@ -68,7 +68,7 @@ tasks:
       RESULTS_DIR: ../../docs/pages/performance/fashion-mnist
     cmds:
       - venv/bin/python install.py --algorithm elastiknn
-      - venv/bin/python run.py --dataset fashion-mnist-784-euclidean --algorithm elastiknn-l2lsh --runs 3 --count 100 --parallelism 3 --force
+      - venv/bin/python run.py --dataset fashion-mnist-784-euclidean --algorithm elastiknn-l2lsh --run-disabled --runs 3 --count 100 --parallelism 3 --force
       - mkdir -p $RESULTS_DIR
       - venv/bin/python plot.py --dataset fashion-mnist-784-euclidean --count 100 --output $RESULTS_DIR/plot.png | venv/bin/python ../parse_results.py > $RESULTS_DIR/results.md
       - base64 -i $RESULTS_DIR/plot.png > $RESULTS_DIR/plot.b64

docs/pages/performance/fashion-mnist/results.md

@@ -1,10 +1,10 @@
 |Model|Parameters|Recall|Queries per Second|
 |---|---|---|---|
-|eknn-l2lsh|L=100 k=4 w=1024 candidates=500 probes=0|0.378|380.590|
-|eknn-l2lsh|L=100 k=4 w=1024 candidates=1000 probes=0|0.446|315.649|
-|eknn-l2lsh|L=100 k=4 w=1024 candidates=500 probes=3|0.635|296.996|
-|eknn-l2lsh|L=100 k=4 w=1024 candidates=1000 probes=3|0.716|254.062|
-|eknn-l2lsh|L=100 k=4 w=2048 candidates=500 probes=0|0.767|324.453|
-|eknn-l2lsh|L=100 k=4 w=2048 candidates=1000 probes=0|0.847|285.500|
-|eknn-l2lsh|L=100 k=4 w=2048 candidates=500 probes=3|0.921|227.240|
-|eknn-l2lsh|L=100 k=4 w=2048 candidates=1000 probes=3|0.960|202.961|
+|eknn-l2lsh|L=100 k=4 w=1024 candidates=500 probes=0|0.378|385.618|
+|eknn-l2lsh|L=100 k=4 w=1024 candidates=1000 probes=0|0.446|321.420|
+|eknn-l2lsh|L=100 k=4 w=1024 candidates=500 probes=3|0.635|301.004|
+|eknn-l2lsh|L=100 k=4 w=1024 candidates=1000 probes=3|0.716|260.304|
+|eknn-l2lsh|L=100 k=4 w=2048 candidates=500 probes=0|0.767|328.881|
+|eknn-l2lsh|L=100 k=4 w=2048 candidates=1000 probes=0|0.847|287.976|
+|eknn-l2lsh|L=100 k=4 w=2048 candidates=500 probes=3|0.922|229.053|
+|eknn-l2lsh|L=100 k=4 w=2048 candidates=1000 probes=3|0.960|204.299|