We read every piece of feedback, and take your input very seriously.
To see all available qualifiers, see our documentation.
There was an error while loading. Please reload this page.
1 parent f902bcd commit 2a7d256Copy full SHA for 2a7d256
timm/models/layers/activations.py
@@ -12,7 +12,7 @@
12
from torch.nn import functional as F
13
14
15
-_USE_MEM_EFFICIENT_ISH = False
+_USE_MEM_EFFICIENT_ISH = True
16
if _USE_MEM_EFFICIENT_ISH:
17
# This version reduces memory overhead of Swish during training by
18
# recomputing torch.sigmoid(x) in backward instead of saving it.
0 commit comments