SONG
SONG copied to clipboard
random_seed doesn't help to get the same result
I am launching the same script two times in a row:
sng = SONG(n_components=2, n_neighbors=10, random_seed=42)
emb = sng.fit(data)
tr = emb.tranform(data)
and I get different values for tr, despite using fixed random seed
Can this be fixed somehow?
I think this is caused by the fact that you don't pass random_seed to UMAP instance here:
self.disp_model = UMAP(learning_rate=self.um_lr, n_components=self.dim, n_epochs=self.um_epochs, init=Y_init, min_dist=self.um_min_dist) but maybe there are other reasons as well
Hi Cristian, I'm like two years late, but this should be fixed in the newer versions.
On Tue, 24 Aug 2021 at 23:05, Cristian Padureac @.***> wrote:
I think this is caused by the fact that you don't pass random_seed to UMAP instance here:
self.disp_model = UMAP(learning_rate=self.um_lr, n_components=self.dim, n_epochs=self.um_epochs, init=Y_init, min_dist=self.um_min_dist) but maybe there are other reasons as well
— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/damithsenanayake/SONG/issues/3#issuecomment-904623151, or unsubscribe https://github.com/notifications/unsubscribe-auth/AAMEQMGOZNLWT43OSYZTTHDT6OKIPANCNFSM5CWYTORA .
-- *- Damith Senanayake - *