leaf_size=leaf_size,
float_type=float_type,
n_jobs=n_jobs,
plat_dev_ids=plat_dev_ids,
verbose=verbose)
start_time = time.time()
nbrs.fit(Xtrain)
end_time = time.time()
print("Fitting time: %f" % (end_time - start_time))
start_time = time.time()
_, _ = nbrs.kneighbors(Xtest)
end_time = time.time()
print("Testing time: %f" % (end_time - start_time))
print("Parsing data ...")
Xtrain, Ytrain, Xtest = generate.get_data_set(NUM_TRAIN=2000000, NUM_TEST=10000000)
print("-------------------------------- DATA --------------------------------")
print("Number of training patterns:\t %i" % Xtrain.shape[0])
print("Number of test patterns:\t %i" % Xtest.shape[0])
print("Dimensionality of patterns:\t%i" % Xtrain.shape[1])
print("----------------------------------------------------------------------")
print("\n\nRunning the GPU version ...")
run_algorithm(algorithm="buffer_kd_tree", tree_depth=9)
print("\n\nRunning the CPU version ...")
run_algorithm(algorithm="kd_tree", leaf_size=32)
# platform dependent parameters
# (might have to be adapted!)
plat_dev_ids = {0:[0, 1, 2, 3]}
n_jobs = 8
# parameters
ofilename = "benchmark.json"
n_test_range = [1000000, 2500000, 5000000, 7500000, 10000000]
algorithms = ["brute", "kd_tree", "buffer_kd_tree"]
verbose = 0
n_neighbors = 10
print("Parsing data ...")
Xtrain, Ytrain, Xtest = generate.get_data_set(NUM_TRAIN=2000000, NUM_TEST=10000000)
print("-------------------------------- DATA --------------------------------")
print("Number of training patterns:\t %i" % Xtrain.shape[0])
print("Number of test patterns:\t %i" % Xtest.shape[0])
print("Dimensionality of patterns:\t%i" % Xtrain.shape[1])
print("----------------------------------------------------------------------")
def compute_opt_tree_depth(algorithm, n_test_tree=2000000):
opt_tree_depth = None
if algorithm in ["buffer_kd_tree", "kd_tree"]:
# the different tree depths that shall
# be tested for this data set
if algorithm == "buffer_kd_tree":
# https://github.com/gieseke/bufferkdtree/tree/master/examples
import time
from bufferkdtree import NearestNeighbors
import generate
Xtrain, Ytrain, Xtest = generate.get_data_set(data_set="psf_model_mag",
NUM_TRAIN=1000000,
NUM_TEST=1000000)
n_jobs = 1
print("Using n_jobs=%i" % n_jobs)
nbrs = NearestNeighbors(n_neighbors=10,
algorithm="kd_tree",
leaf_size=32,
n_jobs=n_jobs)
nbrs.fit(Xtrain)
start_time = time.time()
_, _ = nbrs.kneighbors(Xtest)
end_time = time.time()
print("Testing time: %f" % (end_time - start_time))