forked from hyungkwonko/progressive-umap
-
Notifications
You must be signed in to change notification settings - Fork 0
/
test.py
33 lines (24 loc) · 2.35 KB
/
test.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
from umap import umap_panene
from utils import load_merge_mnist, load_merge_cifar, load_merge_kuzushiji, load_coil
if __name__ == "__main__":
x, y = load_merge_mnist(data="fashion")
umap_panene.UMAP(n_neighbors=5, random_state=2, first_ops=15000, ops=1000).fit_transform(X=x, y=None, label=y, dname='save_dir', progressive=True)
# umap_panene.UMAP(n_neighbors=5, random_state=2, first_ops=40000, ops=300).fit_transform(X=x, y=None, label=y, dname='fashion', progressive=True)
# umap_panene.UMAP(n_neighbors=5, random_state=2).fit_transform(X=x, y=None, label=y, dname='fashion', progressive=False)
# x, y = load_merge_mnist(data="mnist")
# umap_panene.UMAP(n_neighbors=5, random_state=4, first_ops=40000, ops=300).fit_transform(X=x, y=None, label=y, dname='mnist', progressive=True)
# umap_panene.UMAP(n_neighbors=5, random_state=4).fit_transform(X=x, y=None, label=y, dname='mnist', progressive=False)
# x2, y2 = load_merge_cifar()
# umap_panene.UMAP(n_neighbors=5).fit_transform(X=x2, y=None, label=y2, dname='cifar-10', progressive=True)
# x3, y3 = load_merge_kuzushiji(data="kuzushiji", seed=12)
# umap_panene.UMAP(n_neighbors=5, random_state=12, first_ops=100000, ops=300).fit_transform(X=x3, y=None, label=y3, dname='kuzushiji', progressive=True)
# umap_panene.UMAP(n_neighbors=5, random_state=12).fit_transform(X=x3, y=None, label=y3, dname='kuzushiji', progressive=False)
# x4, y4 = load_merge_kuzushiji(data="kuzushiji49", seed=30)
# umap_panene.UMAP(n_neighbors=5, random_state=30, first_ops=400000, ops=100).fit_transform(X=x4, y=None, label=y4, dname='kuzushiji49', progressive=True)
# umap_panene.UMAP(n_neighbors=5, random_state=30).fit_transform(X=x4, y=None, label=y4, dname='kuzushiji49', progressive=False)
# x5, y5 = load_coil(data="coil20", seed=3)
# umap_panene.UMAP(n_neighbors=50, random_state=3, first_ops=2500, ops=10).fit_transform(X=x5, y=None, label=y5, dname='coil20', progressive=True)
# umap_panene.UMAP(n_neighbors=50, random_state=3).fit_transform(X=x5, y=None, label=y5, dname='coil20', progressive=False)
# x6, y6 = load_coil(data="coil100", seed=3)
# umap_panene.UMAP(n_neighbors=10, random_state=3, first_ops=10000, ops=10).fit_transform(X=x6, y=None, label=y6, dname='coil100', progressive=True)
# umap_panene.UMAP(n_neighbors=10, random_state=3).fit_transform(X=x6, y=None, label=y6, dname='coil100', progressive=False)