modeling_hook_registry move

Summary:
Pull Request resolved: https://github.com/facebookresearch/d2go/pull/324

Fixes bug introduced in D37600026 (https://github.com/facebookresearch/d2go/commit/1a18ba3420e3c823accf731a11c0a91dc3babd85) -- forgot to fix imports after moving modelinghook to registry/builtin.py

Differential Revision: D37646330

fbshipit-source-id: cb763d65e7bbfd07eea6eff61727a42a6fcfbc88

d2go/modeling/distillation.py

@@ -23,6 +23,7 @@ from d2go.modeling import modeling_hook as mh
 from d2go.registry.builtin import (
     DISTILLATION_ALGORITHM_REGISTRY,
     DISTILLATION_HELPER_REGISTRY,
+    MODELING_HOOK_REGISTRY,
 )
 from detectron2.utils.file_io import PathManager
 from mobile_cv.common.misc.mixin import dynamic_mixin, remove_dynamic_mixin
@@ -225,7 +226,7 @@ class LabelDistillation(BaseDistillationAlgorithm):
         return super().forward(new_batched_inputs)
 
 
-@mh.MODELING_HOOK_REGISTRY.register()
+@MODELING_HOOK_REGISTRY.register()
 class DistillationModelingHook(mh.ModelingHook):
     """Wrapper hook that allows us to apply different distillation algorithms
     based on config

tests/modeling/test_modeling_meta_arch_modeling_hook.py

@@ -10,7 +10,7 @@ import torch
 from d2go.config import CfgNode
 from d2go.modeling import modeling_hook as mh
 from d2go.modeling.api import build_d2go_model
-from d2go.registry.builtin import META_ARCH_REGISTRY
+from d2go.registry.builtin import META_ARCH_REGISTRY, MODELING_HOOK_REGISTRY
 
 
 @META_ARCH_REGISTRY.register()
@@ -32,7 +32,7 @@ class PlusOneWrapper(torch.nn.Module):
         return self.model(x) + 1
 
 
-@mh.MODELING_HOOK_REGISTRY.register()
+@MODELING_HOOK_REGISTRY.register()
 class PlusOneHook(mh.ModelingHook):
     def __init__(self, cfg):
         super().__init__(cfg)
@@ -55,7 +55,7 @@ class TimesTwoWrapper(torch.nn.Module):
         return self.model(x) * 2
 
 
-@mh.MODELING_HOOK_REGISTRY.register()
+@MODELING_HOOK_REGISTRY.register()
 class TimesTwoHook(mh.ModelingHook):
     def __init__(self, cfg):
         super().__init__(cfg)