biondizzle/vllm
1
0
Fork 0
Code Issues Pull Requests Actions 2 Packages Projects Releases Wiki Activity

Update Optional[x] -> x | None and Union[x, y] to x | y (#26633)

Browse Source 
Signed-off-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>
This commit is contained in:
Harry Mellor
2025-10-12 17:51:31 +01:00
committed by GitHub
parent 9bb38130cb
commit 8fcaaf6a16
944 changed files with 9490 additions and 10121 deletions
Download Patch File Download Diff File Expand all files Collapse all files

34
vllm/model_executor/models/ernie45_moe.py
View File

@@ -25,7 +25,7 @@
from collections.abc import Iterable
from itertools import islice
from typing import Any, Optional, Union
from typing import Any
import torch
from torch import nn
@@ -83,7 +83,7 @@ class Ernie4_5_MoeMLP(nn.Module):
intermediate_size: int,
hidden_act: str,
use_bias: bool = False,
quant_config: Optional[QuantizationConfig] = None,
quant_config: QuantizationConfig | None = None,
reduce_results: bool = True,
prefix: str = "",
) -> None:
@@ -120,7 +120,7 @@ class Ernie4_5_MoeMoE(nn.Module):
def __init__(
self,
config: PretrainedConfig,
quant_config: Optional[QuantizationConfig] = None,
quant_config: QuantizationConfig | None = None,
prefix: str = "",
enable_eplb: bool = False,
):
@@ -229,14 +229,14 @@ class Ernie4_5_MoeAttention(nn.Module):
hidden_size: int,
num_heads: int,
num_kv_heads: int,
head_dim: Optional[int] = None,
head_dim: int | None = None,
rope_theta: float = 500000,
rope_scaling: Optional[dict[str, Any]] = None,
rope_scaling: dict[str, Any] | None = None,
max_position_embeddings: int = 131072,
rms_norm_eps: float = 1e-05,
qkv_bias: bool = False,
cache_config: Optional[CacheConfig] = None,
quant_config: Optional[QuantizationConfig] = None,
cache_config: CacheConfig | None = None,
quant_config: QuantizationConfig | None = None,
prefix: str = "",
) -> None:
super().__init__()
@@ -323,8 +323,8 @@ class Ernie4_5_MoeDecoderLayer(nn.Module):
def __init__(
self,
config: PretrainedConfig,
cache_config: Optional[CacheConfig] = None,
quant_config: Optional[QuantizationConfig] = None,
cache_config: CacheConfig | None = None,
quant_config: QuantizationConfig | None = None,
prefix: str = "",
enable_eplb: bool = False,
) -> None:
@@ -391,7 +391,7 @@ class Ernie4_5_MoeDecoderLayer(nn.Module):
self,
positions: torch.Tensor,
hidden_states: torch.Tensor,
residual: Optional[torch.Tensor],
residual: torch.Tensor | None,
) -> torch.Tensor:
# Self Attention
if residual is None:
@@ -467,9 +467,9 @@ class Ernie4_5_MoeModel(nn.Module):
self,
input_ids: torch.Tensor,
positions: torch.Tensor,
intermediate_tensors: Optional[IntermediateTensors] = None,
inputs_embeds: Optional[torch.Tensor] = None,
) -> Union[torch.Tensor, IntermediateTensors]:
intermediate_tensors: IntermediateTensors | None = None,
inputs_embeds: torch.Tensor | None = None,
) -> torch.Tensor | IntermediateTensors:
if get_pp_group().is_first_rank:
if inputs_embeds is not None:
hidden_states = inputs_embeds
@@ -737,9 +737,9 @@ class Ernie4_5_MoeForCausalLM(nn.Module, SupportsPP, SupportsLoRA, MixtureOfExpe
self,
input_ids: torch.Tensor,
positions: torch.Tensor,
intermediate_tensors: Optional[IntermediateTensors] = None,
inputs_embeds: Optional[torch.Tensor] = None,
) -> Union[torch.Tensor, IntermediateTensors]:
intermediate_tensors: IntermediateTensors | None = None,
inputs_embeds: torch.Tensor | None = None,
) -> torch.Tensor | IntermediateTensors:
hidden_states = self.model(
input_ids, positions, intermediate_tensors, inputs_embeds
)
@@ -748,7 +748,7 @@ class Ernie4_5_MoeForCausalLM(nn.Module, SupportsPP, SupportsLoRA, MixtureOfExpe
def compute_logits(
self,
hidden_states: torch.Tensor,
) -> Optional[torch.Tensor]:
) -> torch.Tensor | None:
logits = self.logits_processor(self.lm_head, hidden_states)
return logits