gradboard 2.3.3__py3-none-any.whl → 2.5.0__py3-none-any.whl

gradboard/optimiser.py

@@ -138,15 +138,7 @@ def get_optimiser(model, optimiser=AdamW, lr=1e-3, weight_decay=1e-2):
     """
     # TODO: print a warning when params are excluded from weight decay IFF wd is set
     weight_decay_exclude = []
-    for keyword in [
-        "nondecay",
-        "bias",
-        "norm",
-        "embedding",
-        "beta",
-        "input_query",
-        "reentrant_query",
-    ]:
+    for keyword in ["nondecay", "bias", "norm", "embedding", "beta"]:
         weight_decay_exclude += [
             p for name, p in model.named_parameters() if keyword in name.lower()
         ]

gradboard/scheduler.py

@@ -165,13 +165,11 @@ class PASS:
         self.max_lr = None
         self.cool_point = None
         for p in sorted(points_left_of_min, key=lambda x: x[0]):
-            if (self.cool_point is None) and (p[1] < minimum[1] + 0.66 * difference):
-                self.cool_point = p[0]
-            elif (self.max_lr is None) and (p[1] < minimum[1] + 0.33 * difference):
+            if (self.max_lr is None) and (p[1] < minimum[1] + 0.2 * difference):
                 self.max_lr = p[0]
             else:
                 continue
-        self.cool_point = min(self.cool_point, self.max_lr / 4)
+        self.cool_point = self.max_lr / 60
         print("High LR", self.max_lr)
         print("Cool point", self.cool_point)
 

{gradboard-2.3.3.dist-info → gradboard-2.5.0.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.3
 Name: gradboard
-Version: 2.3.3
+Version: 2.5.0
 Summary: Easily snowboard down gnarly loss gradients
 License: MIT
 Author: Nicholas Bailey

gradboard-2.5.0.dist-info/RECORD

@@ -0,0 +1,8 @@
+gradboard/__init__.py,sha256=57AkHusYwLCsusiVnajH5pMFKioRCj-3IjF9qpdOzE0,69
+gradboard/cycles.py,sha256=iGEW3Rlp-JNyQZLfpaDXxUCAcEV01ANjLF-Fnhug-qA,10120
+gradboard/optimiser.py,sha256=Br7MNiziiwQZhq6UWkBy1zndRuDFznNfrE91usguZHI,6168
+gradboard/scheduler.py,sha256=rptLeQtFDia69e0_4J92EiflkMIQ2rMM0yvV-Lc56dQ,6621
+gradboard-2.5.0.dist-info/LICENSE,sha256=0BAzJE5BqQ7Iixp_AFdB2W1uO-HCRX-Qfun8PHt6yVM,1073
+gradboard-2.5.0.dist-info/METADATA,sha256=bb6HkJWlXaKwa5un0UYkC379aAm1e2x4JYhyXLVqSEA,2173
+gradboard-2.5.0.dist-info/WHEEL,sha256=b4K_helf-jlQoXBBETfwnf4B04YC67LOev0jo4fX5m8,88
+gradboard-2.5.0.dist-info/RECORD,,

gradboard-2.3.3.dist-info/RECORD

@@ -1,8 +0,0 @@
-gradboard/__init__.py,sha256=57AkHusYwLCsusiVnajH5pMFKioRCj-3IjF9qpdOzE0,69
-gradboard/cycles.py,sha256=iGEW3Rlp-JNyQZLfpaDXxUCAcEV01ANjLF-Fnhug-qA,10120
-gradboard/optimiser.py,sha256=OPWaFSludziX5083NolZOg9BclQFjDbzNz5zsxbERtU,6265
-gradboard/scheduler.py,sha256=TlfdIGx5ufb5dXrhWbDkpePfqLMoHJSGZbdauPgd2FU,6770
-gradboard-2.3.3.dist-info/LICENSE,sha256=0BAzJE5BqQ7Iixp_AFdB2W1uO-HCRX-Qfun8PHt6yVM,1073
-gradboard-2.3.3.dist-info/METADATA,sha256=MSn2XRpfiSRzOCkHcH_FYZate5ONKF1PM8GeJYhJ4XI,2173
-gradboard-2.3.3.dist-info/WHEEL,sha256=b4K_helf-jlQoXBBETfwnf4B04YC67LOev0jo4fX5m8,88
-gradboard-2.3.3.dist-info/RECORD,,