gradboard 2.3.2__py3-none-any.whl → 2.4.0__py3-none-any.whl
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
Potentially problematic release.
This version of gradboard might be problematic. Click here for more details.
- gradboard/optimiser.py +1 -8
- {gradboard-2.3.2.dist-info → gradboard-2.4.0.dist-info}/METADATA +1 -1
- gradboard-2.4.0.dist-info/RECORD +8 -0
- gradboard-2.3.2.dist-info/RECORD +0 -8
- {gradboard-2.3.2.dist-info → gradboard-2.4.0.dist-info}/LICENSE +0 -0
- {gradboard-2.3.2.dist-info → gradboard-2.4.0.dist-info}/WHEEL +0 -0
gradboard/optimiser.py
CHANGED
|
@@ -138,14 +138,7 @@ def get_optimiser(model, optimiser=AdamW, lr=1e-3, weight_decay=1e-2):
|
|
|
138
138
|
"""
|
|
139
139
|
# TODO: print a warning when params are excluded from weight decay IFF wd is set
|
|
140
140
|
weight_decay_exclude = []
|
|
141
|
-
for keyword in [
|
|
142
|
-
"bias",
|
|
143
|
-
"norm",
|
|
144
|
-
"embedding",
|
|
145
|
-
"beta",
|
|
146
|
-
"input_query",
|
|
147
|
-
"reentrant_query",
|
|
148
|
-
]:
|
|
141
|
+
for keyword in ["nondecay", "bias", "norm", "embedding", "beta"]:
|
|
149
142
|
weight_decay_exclude += [
|
|
150
143
|
p for name, p in model.named_parameters() if keyword in name.lower()
|
|
151
144
|
]
|
|
@@ -0,0 +1,8 @@
|
|
|
1
|
+
gradboard/__init__.py,sha256=57AkHusYwLCsusiVnajH5pMFKioRCj-3IjF9qpdOzE0,69
|
|
2
|
+
gradboard/cycles.py,sha256=iGEW3Rlp-JNyQZLfpaDXxUCAcEV01ANjLF-Fnhug-qA,10120
|
|
3
|
+
gradboard/optimiser.py,sha256=Br7MNiziiwQZhq6UWkBy1zndRuDFznNfrE91usguZHI,6168
|
|
4
|
+
gradboard/scheduler.py,sha256=TlfdIGx5ufb5dXrhWbDkpePfqLMoHJSGZbdauPgd2FU,6770
|
|
5
|
+
gradboard-2.4.0.dist-info/LICENSE,sha256=0BAzJE5BqQ7Iixp_AFdB2W1uO-HCRX-Qfun8PHt6yVM,1073
|
|
6
|
+
gradboard-2.4.0.dist-info/METADATA,sha256=s1oRCEIdinQyols8xqtXOZgn7ogUxaPGssq8RB7u950,2173
|
|
7
|
+
gradboard-2.4.0.dist-info/WHEEL,sha256=b4K_helf-jlQoXBBETfwnf4B04YC67LOev0jo4fX5m8,88
|
|
8
|
+
gradboard-2.4.0.dist-info/RECORD,,
|
gradboard-2.3.2.dist-info/RECORD
DELETED
|
@@ -1,8 +0,0 @@
|
|
|
1
|
-
gradboard/__init__.py,sha256=57AkHusYwLCsusiVnajH5pMFKioRCj-3IjF9qpdOzE0,69
|
|
2
|
-
gradboard/cycles.py,sha256=iGEW3Rlp-JNyQZLfpaDXxUCAcEV01ANjLF-Fnhug-qA,10120
|
|
3
|
-
gradboard/optimiser.py,sha256=8an_TlTENDyLt3qeh8HG3wvxBZc_YkvzWGu_OAoQuCc,6245
|
|
4
|
-
gradboard/scheduler.py,sha256=TlfdIGx5ufb5dXrhWbDkpePfqLMoHJSGZbdauPgd2FU,6770
|
|
5
|
-
gradboard-2.3.2.dist-info/LICENSE,sha256=0BAzJE5BqQ7Iixp_AFdB2W1uO-HCRX-Qfun8PHt6yVM,1073
|
|
6
|
-
gradboard-2.3.2.dist-info/METADATA,sha256=oIhDEsPL34QIOYk6TrGDtOmIq0cv6YEMxidjnPmAsoU,2173
|
|
7
|
-
gradboard-2.3.2.dist-info/WHEEL,sha256=b4K_helf-jlQoXBBETfwnf4B04YC67LOev0jo4fX5m8,88
|
|
8
|
-
gradboard-2.3.2.dist-info/RECORD,,
|
|
File without changes
|
|
File without changes
|