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

[major] Training: SmoothL1Loss update beta from 1.0 to 0.3 #1461

Merged
merged 9 commits into from
Nov 6, 2023
Merged

[major] Training: SmoothL1Loss update beta from 1.0 to 0.3 #1461

Changes from all commits
Commits
Show all changes
9 commits
Select commit Hold shift + click to select a range
41f83d1
replace false mentions of Huber loss
ourownstory Oct 18, 2023
8e6ef6e
set SmoothL1Loss beta to 0.1
ourownstory Oct 18, 2023
db12020
Merge branch 'main' into smoothl1loss-update-beta-01
ourownstory Oct 18, 2023
79adb9f
increase beta to 0.3
ourownstory Oct 19, 2023
c480093
reset beta to 1.0
ourownstory Oct 19, 2023
384867a
set beta to 0.3
ourownstory Oct 19, 2023
fd3062f
Merge branch 'main' into smoothl1loss-update-beta-03
ourownstory Oct 19, 2023
2367bb1
Merge branch 'main' into smoothl1loss-update-beta-03
ourownstory Oct 19, 2023
c0df76b
Merge branch 'main' into smoothl1loss-update-beta-03
ourownstory Nov 1, 2023
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
2 changes: 1 addition & 1 deletion neuralprophet/configure.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -119,7 +119,7 @@ def set_loss_func(self):
if isinstance(self.loss_func, str):
if self.loss_func.lower() in ["smoothl1", "smoothl1loss", "huber"]:
# keeping 'huber' for backwards compatiblility, though not identical
self.loss_func = torch.nn.SmoothL1Loss(reduction="none", beta=1.0)
self.loss_func = torch.nn.SmoothL1Loss(reduction="none", beta=0.3)
elif self.loss_func.lower() in ["mae", "maeloss", "l1", "l1loss"]:
self.loss_func = torch.nn.L1Loss(reduction="none")
elif self.loss_func.lower() in ["mse", "mseloss", "l2", "l2loss"]:
Expand Down