[shardformer] rename policy file name

colossalai/shardformer/policies/autopolicy.py → colossalai/shardformer/policies/auto_policy.py

@@ -3,7 +3,7 @@ from dataclasses import dataclass
 
 import torch.nn as nn
 
-from .basepolicy import Policy
+from .base_policy import Policy
 
 __all__ = ["PolicyLocation", "get_autopolicy", "import_policy"]
 

colossalai/shardformer/policies/bert.py

@@ -3,7 +3,7 @@ import torch.nn as nn
 import colossalai.shardformer.layer as col_nn
 
 from .._utils import getattr_, setattr_
-from .basepolicy import ModulePolicyDescription, Policy, SubModuleReplacementDescription
+from .base_policy import ModulePolicyDescription, Policy, SubModuleReplacementDescription
 
 __all__ = [
     'BertPolicy', 'BertModelPolicy', 'BertForPretrainingPolicy', 'BertLMHeadModelPolicy', 'BertForMaskedLMPolicy',

colossalai/shardformer/policies/bloom.py

@@ -4,7 +4,7 @@ import colossalai.shardformer.layer as col_nn
 
 from .._utils import getattr_, setattr_
 from ..modeling.bloom import build_bloom_alibi_tensor_fn
-from .basepolicy import ModulePolicyDescription, Policy, SubModuleReplacementDescription
+from .base_policy import ModulePolicyDescription, Policy, SubModuleReplacementDescription
 
 
 class BloomPolicy(Policy):

colossalai/shardformer/policies/gpt2.py

@@ -3,7 +3,7 @@ import torch.nn as nn
 import colossalai.shardformer.layer as col_nn
 
 from .._utils import getattr_, setattr_
-from .basepolicy import ModulePolicyDescription, Policy, SubModuleReplacementDescription
+from .base_policy import ModulePolicyDescription, Policy, SubModuleReplacementDescription
 
 __all__ = [
     'GPT2Policy', 'GPT2ModelPolicy', 'GPT2LMHeadModelPolicy', 'GPT2DoubleHeadsModelPolicy',

colossalai/shardformer/policies/llama.py

@@ -4,7 +4,7 @@ import torch.nn as nn
 
 from colossalai.shardformer.layer import FusedRMSNorm, Linear1D_Col, Linear1D_Row, VocabParallelEmbedding1D
 
-from .basepolicy import ModulePolicyDescription, Policy, SubModuleReplacementDescription
+from .base_policy import ModulePolicyDescription, Policy, SubModuleReplacementDescription
 
 __all__ = ['LlamaPolicy', 'LlamaForCausalLMPolicy', 'LlamaForSequenceClassificationPolicy']
 

colossalai/shardformer/policies/opt.py

 from colossalai.shardformer.layer import FusedLayerNorm, Linear1D_Col, Linear1D_Row, VocabParallelEmbedding1D
 
 from .._utils import getattr_, setattr_
-from .basepolicy import ModulePolicyDescription, Policy, SubModuleReplacementDescription
+from .base_policy import ModulePolicyDescription, Policy, SubModuleReplacementDescription
 
 __all__ = [
     'OPTPolicy', 'OPTModelPolicy', 'OPTForCausalLMPolicy', 'OPTForSequenceClassificationPolicy',

colossalai/shardformer/policies/t5.py

@@ -6,10 +6,10 @@ from colossalai.shardformer.layer import (
     Linear1D_Row,
     VocabParallelEmbedding1D,
 )
-from colossalai.shardformer.policies.basepolicy import ModulePolicyDescription
+from colossalai.shardformer.policies.base_policy import ModulePolicyDescription
 
 from .._utils import getattr_, setattr_
-from .basepolicy import ModulePolicyDescription, Policy, SubModuleReplacementDescription
+from .base_policy import ModulePolicyDescription, Policy, SubModuleReplacementDescription
 
 __all__ = ["T5ModelPolicy", "T5ForConditionalGenerationPolicy", "T5EncoderPolicy"]
 

colossalai/shardformer/policies/vit.py

@@ -4,7 +4,7 @@ import torch.nn as nn
 
 from colossalai.shardformer.layer import DropoutForReplicatedInput, FusedLayerNorm, Linear1D_Col, Linear1D_Row
 
-from .basepolicy import ModulePolicyDescription, Policy, SubModuleReplacementDescription
+from .base_policy import ModulePolicyDescription, Policy, SubModuleReplacementDescription
 
 __all__ = ['ViTPolicy']
 

colossalai/shardformer/shard/sharder.py

@@ -6,8 +6,8 @@ from torch import Tensor
 from colossalai.lazy import LazyTensor
 
 from .._utils import getattr_, setattr_
-from ..policies.autopolicy import get_autopolicy
-from ..policies.basepolicy import Policy, SubModuleReplacementDescription
+from ..policies.auto_policy import get_autopolicy
+from ..policies.base_policy import Policy, SubModuleReplacementDescription
 from .shard_config import ShardConfig
 from .utils import set_tensors_to_none
 

colossalai/shardformer/shard/shardformer.py

@@ -5,7 +5,7 @@ from torch import Tensor
 
 from colossalai.cluster import DistCoordinator
 
-from ..policies.basepolicy import Policy
+from ..policies.base_policy import Policy
 from .shard_config import ShardConfig
 from .sharder import ModelSharder