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

42
vllm/model_executor/models/zamba2.py
View File

@@ -10,7 +10,7 @@ model alternates between state space model layers and attention-based layers.
from collections.abc import Iterable
from itertools import cycle
from typing import Any, Optional, Union
from typing import Any
import torch
from torch import nn
@@ -60,8 +60,8 @@ class Zamba2LoRA(nn.Module):
self,
input_dim: int,
rank: int,
output_dim: Union[int, list[int]],
quant_config: Optional[QuantizationConfig] = None,
output_dim: int | list[int],
quant_config: QuantizationConfig | None = None,
prefix: str = "",
):
"""Initialize the attention layer.
@@ -106,8 +106,8 @@ class Zamba2Attention(nn.Module):
config: Zamba2Config,
bare_block_idx: int,
num_hybrid_layers: int,
cache_config: Optional[CacheConfig] = None,
quant_config: Optional[QuantizationConfig] = None,
cache_config: CacheConfig | None = None,
quant_config: QuantizationConfig | None = None,
prefix: str = "",
) -> None:
"""Initialize the attention layer.
@@ -288,7 +288,7 @@ class Zamba2MLP(nn.Module):
config: Zamba2Config,
bare_block_idx: int,
num_hybrid_layers: dict[int, int],
quant_config: Optional[QuantizationConfig] = None,
quant_config: QuantizationConfig | None = None,
prefix: str = "",
) -> None:
"""Initialize the MLP layer.
@@ -386,8 +386,8 @@ class Zamba2AttentionDecoderLayer(nn.Module):
config: Zamba2Config,
bare_block_idx: int,
num_hybrid_layers: int,
cache_config: Optional[CacheConfig] = None,
quant_config: Optional[QuantizationConfig] = None,
cache_config: CacheConfig | None = None,
quant_config: QuantizationConfig | None = None,
prefix: str = "",
) -> None:
"""Initialize the decoder layer.
@@ -484,9 +484,9 @@ class Zamba2MambaDecoderLayer(nn.Module):
def __init__(
self,
config: Zamba2Config,
model_config: Optional[ModelConfig] = None,
cache_config: Optional[CacheConfig] = None,
quant_config: Optional[QuantizationConfig] = None,
model_config: ModelConfig | None = None,
cache_config: CacheConfig | None = None,
quant_config: QuantizationConfig | None = None,
prefix: str = "",
) -> None:
"""Initialize the Mamba decoder layer.
@@ -523,9 +523,9 @@ class Zamba2MambaDecoderLayer(nn.Module):
def forward(
self,
hidden_states: torch.Tensor,
transformer_hidden_states: Optional[torch.Tensor] = None,
positions: Optional[torch.Tensor] = None,
original_hidden_states: Optional[torch.Tensor] = None,
transformer_hidden_states: torch.Tensor | None = None,
positions: torch.Tensor | None = None,
original_hidden_states: torch.Tensor | None = None,
) -> torch.Tensor:
"""Forward pass through the Mamba decoder layer.
@@ -581,9 +581,9 @@ class Zamba2HybridLayer(nn.Module):
shared_transformer: Zamba2AttentionDecoderLayer,
config: Zamba2Config,
block_idx: int,
model_config: Optional[ModelConfig] = None,
cache_config: Optional[CacheConfig] = None,
quant_config: Optional[QuantizationConfig] = None,
model_config: ModelConfig | None = None,
cache_config: CacheConfig | None = None,
quant_config: QuantizationConfig | None = None,
prefix: str = "",
) -> None:
"""Initialize the hybrid layer.
@@ -764,8 +764,8 @@ class Zamba2Model(nn.Module):
self,
input_ids: torch.Tensor,
positions: torch.Tensor,
inputs_embeds: Optional[torch.Tensor] = None,
) -> Union[torch.Tensor, IntermediateTensors]:
inputs_embeds: torch.Tensor | None = None,
) -> torch.Tensor | IntermediateTensors:
"""Forward pass through the model.
Args:
@@ -947,7 +947,7 @@ class Zamba2ForCausalLM(nn.Module, HasInnerState, IsHybrid):
self,
input_ids: torch.Tensor,
positions: torch.Tensor,
inputs_embeds: Optional[torch.Tensor] = None,
inputs_embeds: torch.Tensor | None = None,
**kwargs: Any,
) -> torch.Tensor:
"""Forward pass through the model.
@@ -973,7 +973,7 @@ class Zamba2ForCausalLM(nn.Module, HasInnerState, IsHybrid):
def compute_logits(
self,
hidden_states: torch.Tensor,
) -> Optional[torch.Tensor]:
) -> torch.Tensor | None:
"""Compute logits for next token prediction.
Args: