Skip to content

[Fix/Inference] Remove unused and non-functional functions #5543

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
yuanheng-zhao merged 2 commits into from
Apr 2, 2024
Merged

[Fix/Inference] Remove unused and non-functional functions #5543

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
f02c9d2
[fix] remove unused func
yuanheng-zhao Apr 2, 2024
d45bac0
rm non-functional partial
yuanheng-zhao Apr 2, 2024
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
29 changes: 8 additions & 21 deletions colossalai/inference/modeling/policy/nopadding_llama.py
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,5 +1,3 @@
from functools import partial

from torch.nn import Parameter
from transformers.models.llama.modeling_llama import LlamaDecoderLayer, LlamaForCausalLM, LlamaModel, LlamaRMSNorm

Expand All @@ -13,8 +11,6 @@
)
from colossalai.inference.utils import init_to_get_rotary
from colossalai.shardformer.policies.base_policy import ModulePolicyDescription, SubModuleReplacementDescription

# import colossalai
from colossalai.shardformer.policies.llama import LlamaForCausalLMPolicy


Expand Down Expand Up @@ -45,27 +41,18 @@ def module_policy(self):
]
)

self.shard_config._infer()

infer_forward = llama_causal_lm_forward
method_replacement = {"forward": partial(infer_forward)}
self.append_or_create_method_replacement(
description=method_replacement, policy=policy, target_key=LlamaForCausalLM
description={"forward": llama_causal_lm_forward}, policy=policy, target_key=LlamaForCausalLM
)

infer_forward = llama_model_forward
method_replacement = {"forward": partial(infer_forward)}
self.append_or_create_method_replacement(description=method_replacement, policy=policy, target_key=LlamaModel)

infer_forward = llama_decoder_layer_forward
method_replacement = {"forward": partial(infer_forward)}
self.append_or_create_method_replacement(
description=method_replacement, policy=policy, target_key=LlamaDecoderLayer
description={"forward": llama_model_forward}, policy=policy, target_key=LlamaModel
)
self.append_or_create_method_replacement(
description={"forward": llama_decoder_layer_forward}, policy=policy, target_key=LlamaDecoderLayer
)
self.append_or_create_method_replacement(
description={"forward": llama_rmsnorm_forward}, policy=policy, target_key=LlamaRMSNorm
)

infer_forward = llama_rmsnorm_forward
method_replacement = {"forward": partial(infer_forward)}
self.append_or_create_method_replacement(description=method_replacement, policy=policy, target_key=LlamaRMSNorm)

return policy

Expand Down
8 changes: 0 additions & 8 deletions colossalai/shardformer/shard/shard_config.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -36,8 +36,6 @@ class ShardConfig:
enable_sequence_overlap: bool = False
parallel_output = True
extra_kwargs: Dict[str, Any] = field(default_factory=dict)
# pipeline_parallel_size: int
# data_parallel_size: int
# tensor_parallel_mode: Literal['1d', '2d', '2.5d', '3d']

@property
Expand Down Expand Up @@ -70,9 +68,3 @@ def _turn_on_all_optimization(self):
self.enable_jit_fused = True
self.enable_sequence_parallelism = True
self.enable_sequence_overlap = True

def _infer(self):
"""
Set default params for inference.
"""
# assert self.pipeline_stage_manager is None, "pipeline parallelism is not supported in inference for now"