You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
ColossalAI/colossalai/shardformer/policies/opt.py

385 lines
14 KiB

import warnings
from functools import partial
from typing import Callable, Dict, List
import torch.nn as nn
from torch import Tensor, nn
from colossalai.shardformer.layer import (
FusedLayerNorm,
LayerNorm,
Linear1D_Col,
Linear1D_Row,
PaddingEmbedding,
PaddingLMHead,
VocabParallelEmbedding1D,
VocabParallelLMHead1D,
)
from .._utils import getattr_
from ..modeling.jit import get_jit_fused_dropout_add_func
from ..modeling.opt import (
OPTPipelineForwards,
get_jit_fused_opt_decoder_layer_forward,
get_lm_forward_with_dist_cross_entropy,
get_opt_decoder_forward_for_flash_attention,
get_opt_flash_attention_forward,
)
from .base_policy import ModulePolicyDescription, Policy, SubModuleReplacementDescription
__all__ = [
"OPTPolicy",
"OPTModelPolicy",
"OPTForCausalLMPolicy",
"OPTForSequenceClassificationPolicy",
"OPTForQuestionAnsweringPolicy",
]
class OPTPolicy(Policy):
def __init__(self) -> None:
super().__init__()
def config_sanity_check(self):
pass
def preprocess(self):
self.tie_weight = self.tie_weight_check()
self.origin_attn_implement = self.model.config._attn_implementation
return self.model
def module_policy(self):
from transformers.models.opt.modeling_opt import OPTAttention, OPTDecoder, OPTDecoderLayer, OptFlashAttention2
ATTN_IMPLEMENTATION = {
"eager": OPTAttention,
"flash_attention_2": OptFlashAttention2,
}
policy = {}
attn_cls = ATTN_IMPLEMENTATION[self.model.config._attn_implementation]
embedding_cls = None
if self.shard_config.enable_tensor_parallelism:
embedding_cls = VocabParallelEmbedding1D
else:
if self.tie_weight:
embedding_cls = PaddingEmbedding
if self.shard_config.enable_fused_normalization:
norm_cls = FusedLayerNorm
else:
norm_cls = LayerNorm
if self.shard_config.enable_sequence_parallelism:
self.shard_config.enable_sequence_parallelism = False
warnings.warn("OPT doesn't support sequence parallelism now, will ignore the sequence parallelism flag.")
if self.shard_config.enable_tensor_parallelism:
assert (
self.model.config.num_attention_heads % self.shard_config.tensor_parallel_size == 0
), f"The number of attention heads must be divisible by tensor parallel size."
policy[OPTDecoderLayer] = ModulePolicyDescription(
sub_module_replacement=[
SubModuleReplacementDescription(
suffix="fc1",
target_module=Linear1D_Col,
),
SubModuleReplacementDescription(
suffix="fc2",
target_module=Linear1D_Row,
),
]
)
policy[attn_cls] = ModulePolicyDescription(
attribute_replacement={
"embed_dim": self.model.config.hidden_size // self.shard_config.tensor_parallel_size,
"num_heads": self.model.config.num_attention_heads // self.shard_config.tensor_parallel_size,
},
sub_module_replacement=[
SubModuleReplacementDescription(
suffix="q_proj",
target_module=Linear1D_Col,
),
SubModuleReplacementDescription(
suffix="k_proj",
target_module=Linear1D_Col,
),
SubModuleReplacementDescription(
suffix="v_proj",
target_module=Linear1D_Col,
),
SubModuleReplacementDescription(
suffix="out_proj",
target_module=Linear1D_Row,
),
],
)
if embedding_cls is not None:
self.append_or_create_submodule_replacement(
description=SubModuleReplacementDescription(
suffix="embed_tokens",
target_module=embedding_cls,
kwargs={"make_vocab_size_divisible_by": self.shard_config.make_vocab_size_divisible_by},
),
policy=policy,
target_key=OPTDecoder,
)
# optimization configuration
self.append_or_create_submodule_replacement(
description=SubModuleReplacementDescription(
suffix="final_layer_norm",
target_module=norm_cls,
ignore_if_not_exist=True,
),
policy=policy,
target_key=OPTDecoder,
)
self.append_or_create_submodule_replacement(
description=[
SubModuleReplacementDescription(
suffix="self_attn_layer_norm",
target_module=norm_cls,
ignore_if_not_exist=True,
),
SubModuleReplacementDescription(
suffix="final_layer_norm",
target_module=norm_cls,
ignore_if_not_exist=True,
),
],
policy=policy,
target_key=OPTDecoderLayer,
)
# use flash attention
if self.shard_config.enable_flash_attention:
self.append_or_create_method_replacement(
description={
"forward": get_opt_flash_attention_forward(self.shard_config),
},
policy=policy,
target_key=attn_cls,
)
if not self.shard_config.pipeline_stage_manager:
self.append_or_create_method_replacement(
description={
"forward": get_opt_decoder_forward_for_flash_attention(self.shard_config),
},
policy=policy,
target_key=OPTDecoder,
)
# use jit fused operator
if self.shard_config.enable_jit_fused:
self.append_or_create_method_replacement(
description={
"forward": get_jit_fused_opt_decoder_layer_forward(),
"dropout_add": get_jit_fused_dropout_add_func(),
},
policy=policy,
target_key=OPTDecoderLayer,
)
return policy
def postprocess(self):
return self.model
def get_held_layers(self) -> List[nn.Module]:
"""Get pipeline layers for current stage."""
assert self.pipeline_stage_manager is not None
if self.model.__class__.__name__ == "OPTModel":
module = self.model.decoder
else:
module = self.model.model.decoder
stage_manager = self.pipeline_stage_manager
held_layers = []
layers_per_stage = stage_manager.distribute_layers(len(module.layers))
if stage_manager.is_first_stage():
held_layers.append(module.embed_tokens)
held_layers.append(module.embed_positions)
held_layers.append(module.project_in)
start_idx, end_idx = stage_manager.get_stage_index(layers_per_stage)
held_layers.extend(module.layers[start_idx:end_idx])
if stage_manager.is_last_stage():
held_layers.append(module.final_layer_norm)
held_layers.append(module.project_out)
return held_layers
def set_pipeline_forward(self, model_cls: nn.Module, new_forward: Callable, policy: Dict) -> None:
"""If under pipeline parallel setting, replacing the original forward method of huggingface
to customized forward method, and add this changing to policy."""
if self.pipeline_stage_manager:
stage_manager = self.pipeline_stage_manager
if self.model.__class__.__name__ == "OPTModel":
module = self.model.decoder
else:
module = self.model.model.decoder
layers_per_stage = stage_manager.distribute_layers(len(module.layers))
stage_index = stage_manager.get_stage_index(layers_per_stage)
method_replacement = {
"forward": partial(
new_forward,
stage_manager=stage_manager,
stage_index=stage_index,
shard_config=self.shard_config,
)
}
self.append_or_create_method_replacement(
description=method_replacement, policy=policy, target_key=model_cls
)
class OPTModelPolicy(OPTPolicy):
def module_policy(self):
from transformers.models.opt.modeling_opt import OPTModel
policy = super().module_policy()
if self.pipeline_stage_manager:
self.set_pipeline_forward(
model_cls=OPTModel,
new_forward=OPTPipelineForwards.opt_model_forward,
policy=policy,
)
return policy
def get_held_layers(self) -> List[nn.Module]:
return super().get_held_layers()
def get_shared_params(self) -> List[Dict[int, Tensor]]:
"""No shared params in OPTModel."""
return []
class OPTForCausalLMPolicy(OPTPolicy):
def module_policy(self):
from transformers.models.opt.modeling_opt import OPTForCausalLM
policy = super().module_policy()
if self.shard_config.enable_tensor_parallelism:
self.append_or_create_submodule_replacement(
description=SubModuleReplacementDescription(
suffix="lm_head",
target_module=VocabParallelLMHead1D,
kwargs=dict(
gather_output=not self.shard_config.parallel_output,
make_vocab_size_divisible_by=self.shard_config.make_vocab_size_divisible_by,
),
),
policy=policy,
target_key=OPTForCausalLM,
)
if self.shard_config.parallel_output:
method_replacement = {"forward": get_lm_forward_with_dist_cross_entropy(self.shard_config)}
self.append_or_create_method_replacement(
description=method_replacement, policy=policy, target_key=OPTForCausalLM
)
else:
self.append_or_create_submodule_replacement(
description=SubModuleReplacementDescription(
suffix="lm_head",
target_module=PaddingLMHead,
kwargs=dict(make_vocab_size_divisible_by=self.shard_config.make_vocab_size_divisible_by),
),
policy=policy,
target_key=OPTForCausalLM,
)
if self.pipeline_stage_manager:
self.set_pipeline_forward(
model_cls=OPTForCausalLM,
new_forward=OPTPipelineForwards.opt_for_causal_lm_forward,
policy=policy,
)
return policy
def get_held_layers(self) -> List[nn.Module]:
held_layers = super().get_held_layers()
if self.pipeline_stage_manager.is_last_stage():
held_layers.append(self.model.lm_head)
return held_layers
def get_shared_params(self) -> List[Dict[int, Tensor]]:
opt_model = self.model
if self.pipeline_stage_manager and self.pipeline_stage_manager.num_stages > 1:
num_stages = self.pipeline_stage_manager.num_stages
if id(opt_model.model.decoder.embed_tokens.weight) == id(opt_model.lm_head.weight):
return [
{
0: opt_model.model.decoder.embed_tokens.weight,
num_stages - 1: opt_model.lm_head.weight,
}
]
return []
def postprocess(self):
if self.shard_config.enable_tensor_parallelism and self.pipeline_stage_manager is None:
binding_map = {
"model.decoder.embed_tokens": "lm_head",
}
for k, v in binding_map.items():
src_mod = getattr_(self.model, k)
dst_mod = getattr_(self.model, v)
dst_mod.weight = src_mod.weight
return self.model
class OPTForSequenceClassificationPolicy(OPTPolicy):
def module_policy(self):
from transformers.models.opt.modeling_opt import OPTForSequenceClassification
policy = super().module_policy()
if self.pipeline_stage_manager:
self.set_pipeline_forward(
model_cls=OPTForSequenceClassification,
new_forward=OPTPipelineForwards.opt_for_sequence_classification_forward,
policy=policy,
)
return policy
def get_held_layers(self) -> List[nn.Module]:
held_layers = super().get_held_layers()
if self.pipeline_stage_manager.is_last_stage():
held_layers.append(self.model.score)
return held_layers
def get_shared_params(self) -> List[Dict[int, Tensor]]:
"no shared params in OPTForSequenceClassification"
return []
class OPTForQuestionAnsweringPolicy(OPTPolicy):
def module_policy(self):
from transformers.models.opt.modeling_opt import OPTForQuestionAnswering
policy = super().module_policy()
if self.pipeline_stage_manager:
self.set_pipeline_forward(
model_cls=OPTForQuestionAnswering,
new_forward=OPTPipelineForwards.opt_for_question_answering_forward,
policy=policy,
)
return policy
def get_held_layers(self) -> List[nn.Module]:
held_layers = super().get_held_layers()
if self.pipeline_stage_manager.is_last_stage():
held_layers.append(self.model.qa_outputs)
return held_layers
def get_shared_params(self) -> List[Dict[int, Tensor]]:
"no shared params in OPTForSequenceClassification"
return []