hyper-connections 0.3.9__py3-none-any.whl → 0.3.10__py3-none-any.whl

hyper_connections/hyper_connections.py

@@ -175,7 +175,6 @@ class HyperConnections(Module):
         tanh = True,
         channel_first = False,
         dropout = 0.,
-        residual_transform: Module | None = None, # to support resnet blocks where dimension in not equal to dimension out - usually a residual conv
         add_branch_out_to_residual = True,  # will disable depth connections (weighted residual sum with beta) if set False
         num_input_views = 1,                # allow for the branch module to receive multiple input views, dimension placed on the very left (before batch)
         depth_residual_fn = add,
@@ -255,10 +254,6 @@ class HyperConnections(Module):
 
         self.channel_first = channel_first
 
-        # maybe residual transform
-
-        self.residual_transform = default(residual_transform, nn.Identity())
-
         # maybe custom depth connection residual function
         # this is to prepare for gating the addition of the branch outputs to the residual streams
         # needed for memory lanes a la RMT / LMM
@@ -271,8 +266,6 @@ class HyperConnections(Module):
     ):
         streams = self.num_residual_streams
 
-        maybe_transformed_residuals = self.residual_transform(residuals)
-
         # width connection
 
         # handle channel first
@@ -334,7 +327,14 @@ class HyperConnections(Module):
 
         branch_input = self.merge_fracs(branch_input)
 
-        return branch_input, maybe_transformed_residuals, dict(beta = beta)
+        # reshape residuals back
+
+        if self.channel_first:
+            residuals = rearrange(residuals, 'b ... f s d -> (b s) (f d) ...')
+        else:
+            residuals = rearrange(residuals, 'b ... f s d -> (b s) ... (f d)')
+
+        return branch_input, residuals, dict(beta = beta)
 
     def depth_connection(
         self,

hyper_connections/hyper_connections_channel_first.py

@@ -84,7 +84,6 @@ class HyperConnections(Module):
         tanh = True,
         channel_first = True,
         dropout = 0.,
-        residual_transform: Module | None = None, # to support resnet blocks where dimension in not equal to dimension out - usually a residual conv
     ):
         """
         Appendix J, Algorithm2 in - https://arxiv.org/abs/2409.19606
@@ -124,19 +123,12 @@ class HyperConnections(Module):
         self.dynamic_alpha_scale = nn.Parameter(torch.ones(()) * 1e-2)
         self.dynamic_beta_scale = nn.Parameter(torch.ones(()) * 1e-2)
 
-
         # dropouts
 
         self.dropout = nn.Dropout(dropout)
 
-        # maybe residual transform
-
-        self.residual_transform = default(residual_transform, nn.Identity())
-
     def width_connection(self, residuals):
 
-        maybe_transformed_residuals = self.residual_transform(residuals)
-
         # width connection
 
         normed = self.norm(residuals)
@@ -161,7 +153,9 @@ class HyperConnections(Module):
 
         branch_input, residuals = mix_h[:, 0, ...], mix_h[:, 1:, ...]
 
-        return branch_input, maybe_transformed_residuals, dict(beta = beta)
+        residuals = rearrange(residuals, 'b s d ... -> (b s) d ...')
+
+        return branch_input, residuals, dict(beta = beta)
 
     def depth_connection(self, branch_output, residuals, *, beta):
         # 'depth' connection

hyper_connections/manifold_constrained_hyper_connections.py

@@ -307,7 +307,7 @@ class ManifoldConstrainedHyperConnections(Module):
     ):
         streams = self.num_residual_streams
 
-        maybe_transformed_residuals = self.residual_transform(residuals)
+        residuals = self.residual_transform(residuals)
 
         # width connection
 
@@ -397,13 +397,14 @@ class ManifoldConstrainedHyperConnections(Module):
 
         branch_input = self.merge_fracs(branch_input)
 
-        branch_input = branch_input.to(dtype)
-        residuals = residuals.to(dtype)
+        residuals = rearrange(residuals, 'b ... f s d -> (b s) ... (f d)')
+
+        branch_input, residuals = tuple(t.to(dtype) for t in (branch_input, residuals))
 
         if exists(beta):
             beta = beta.to(dtype)
 
-        return branch_input, maybe_transformed_residuals, dict(beta = beta)
+        return branch_input, residuals, dict(beta = beta)
 
     def depth_connection(
         self,

{hyper_connections-0.3.9.dist-info → hyper_connections-0.3.10.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: hyper-connections
-Version: 0.3.9
+Version: 0.3.10
 Summary: Hyper-Connections
 Project-URL: Homepage, https://pypi.org/project/hyper-connections/
 Project-URL: Repository, https://github.com/lucidrains/hyper-connections

hyper_connections-0.3.10.dist-info/RECORD

@@ -0,0 +1,12 @@
+hyper_connections/__init__.py,sha256=BAGwi53ozXcnfPJAGur0RHA4vcolF1ORBhbZ9a8SkrE,602
+hyper_connections/hyper_connections.py,sha256=rqFJj3U0LF3uDKNKNPBpRrmf0oa2BGWVbD6S-xdZdLo,14904
+hyper_connections/hyper_connections_channel_first.py,sha256=Mh_hzhTi96ZoOPmhSKwUaF4TbHpNqhs83wNe5hNuL7o,6532
+hyper_connections/hyper_connections_with_multi_branch_inputs.py,sha256=6BXKdSwyx6wdQVseebKG2EQkhVaVLrrepOlL8lLnex4,7855
+hyper_connections/hyper_connections_with_multi_input_streams.py,sha256=ueT3CJPHrt5hRU7q1bFF0rANWJh_pXqclt6HiUu1gBY,11331
+hyper_connections/manifold_constrained_hyper_connections.py,sha256=SkGAWpBHnrOlIcixb0iIGej9StO82O7KXrFjYuSKx7I,17424
+hyper_connections/residuals.py,sha256=JVSFJj_H7xQ3_Fd-pZH5Hdv9SveAQu29jQNvMyom5ek,921
+hyper_connections/vit.py,sha256=fTC8hAYkD4qm-KURAj8SJ66C6ZWtsBdHf_kS-4rJZGQ,5049
+hyper_connections-0.3.10.dist-info/METADATA,sha256=tEYVvFTVY_13gYQbflz-mjWMQEwH4DvPvQk76X9Iq2E,6705
+hyper_connections-0.3.10.dist-info/WHEEL,sha256=WLgqFyCfm_KASv4WHyYy0P3pM_m7J5L9k2skdKLirC8,87
+hyper_connections-0.3.10.dist-info/licenses/LICENSE,sha256=E7RGS7kpJIStk5za_-4DVhWEAamf65EU0CNML25mq4c,1066
+hyper_connections-0.3.10.dist-info/RECORD,,

hyper_connections-0.3.9.dist-info/RECORD

@@ -1,12 +0,0 @@
-hyper_connections/__init__.py,sha256=BAGwi53ozXcnfPJAGur0RHA4vcolF1ORBhbZ9a8SkrE,602
-hyper_connections/hyper_connections.py,sha256=UHxZhyRwx89GRgmQVt53Gv6JeNhX8UCjjETlydMZjTk,15021
-hyper_connections/hyper_connections_channel_first.py,sha256=_1PM4LRcPpDqfCiHlBMc2nLV08sXM2nuyZGSKTiuqbE,6818
-hyper_connections/hyper_connections_with_multi_branch_inputs.py,sha256=6BXKdSwyx6wdQVseebKG2EQkhVaVLrrepOlL8lLnex4,7855
-hyper_connections/hyper_connections_with_multi_input_streams.py,sha256=ueT3CJPHrt5hRU7q1bFF0rANWJh_pXqclt6HiUu1gBY,11331
-hyper_connections/manifold_constrained_hyper_connections.py,sha256=uF9WALGLeEBdfUm_p8O8ZTmmsk3L44gg-G1GW1SCMO0,17382
-hyper_connections/residuals.py,sha256=JVSFJj_H7xQ3_Fd-pZH5Hdv9SveAQu29jQNvMyom5ek,921
-hyper_connections/vit.py,sha256=fTC8hAYkD4qm-KURAj8SJ66C6ZWtsBdHf_kS-4rJZGQ,5049
-hyper_connections-0.3.9.dist-info/METADATA,sha256=mAciMU5pRr1oxP1OvjUFnwJAxqU9RTOKjPs-I7xn1ns,6704
-hyper_connections-0.3.9.dist-info/WHEEL,sha256=WLgqFyCfm_KASv4WHyYy0P3pM_m7J5L9k2skdKLirC8,87
-hyper_connections-0.3.9.dist-info/licenses/LICENSE,sha256=E7RGS7kpJIStk5za_-4DVhWEAamf65EU0CNML25mq4c,1066
-hyper_connections-0.3.9.dist-info/RECORD,,