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

Signed-off-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>
This commit is contained in:
Harry Mellor
2025-10-12 09:51:31 -07:00
committed by GitHub
parent 9bb38130cb
commit 8fcaaf6a16
944 changed files with 9490 additions and 10121 deletions
+17 -17
View File
@@ -19,7 +19,7 @@
import math
from collections.abc import Iterable, Mapping
from itertools import tee
from typing import Annotated, Literal, Optional, Union
from typing import Annotated, Literal
import torch
from torch import nn
@@ -115,7 +115,7 @@ class Llama4VisionMLP(nn.Module):
output_size: int,
bias: bool,
output_activation: bool,
quant_config: Optional[QuantizationConfig] = None,
quant_config: QuantizationConfig | None = None,
prefix: str = "",
use_data_parallel: bool = False,
):
@@ -152,7 +152,7 @@ class Llama4MultiModalProjector(nn.Module):
def __init__(
self,
config,
quant_config: Optional[QuantizationConfig] = None,
quant_config: QuantizationConfig | None = None,
prefix: str = "",
):
super().__init__()
@@ -199,7 +199,7 @@ class Llama4VisionPixelShuffleMLP(nn.Module):
def __init__(
self,
config,
quant_config: Optional[QuantizationConfig] = None,
quant_config: QuantizationConfig | None = None,
prefix: str = "",
use_data_parallel: bool = False,
):
@@ -229,7 +229,7 @@ class Llama4VisionAttention(nn.Module):
def __init__(
self,
config: Llama4VisionConfig,
quant_config: Optional[QuantizationConfig],
quant_config: QuantizationConfig | None,
prefix: str = "",
use_data_parallel: bool = False,
):
@@ -323,7 +323,7 @@ class Llama4VisionEncoderLayer(nn.Module):
def __init__(
self,
config: Llama4VisionConfig,
quant_config: Optional[QuantizationConfig],
quant_config: QuantizationConfig | None,
prefix: str = "",
use_data_parallel: bool = False,
):
@@ -376,7 +376,7 @@ class Llama4VisionEncoder(nn.Module):
def __init__(
self,
config: Llama4VisionConfig,
quant_config: Optional[QuantizationConfig],
quant_config: QuantizationConfig | None,
prefix: str = "",
use_data_parallel: bool = False,
):
@@ -419,7 +419,7 @@ class Llama4UnfoldConvolution(nn.Module):
def __init__(
self,
config: Llama4VisionConfig,
quant_config: Optional[QuantizationConfig] = None,
quant_config: QuantizationConfig | None = None,
prefix: str = "",
use_data_parallel: bool = False,
):
@@ -449,7 +449,7 @@ class Llama4VisionModel(nn.Module):
def __init__(
self,
config: Llama4VisionConfig,
quant_config: Optional[QuantizationConfig] = None,
quant_config: QuantizationConfig | None = None,
prefix: str = "",
use_data_parallel: bool = False,
):
@@ -547,7 +547,7 @@ class Mllama4ProcessingInfo(BaseProcessingInfo):
Llama4Processor, use_fast=kwargs.pop("use_fast", True), **kwargs
)
def get_supported_mm_limits(self) -> Mapping[str, Optional[int]]:
def get_supported_mm_limits(self) -> Mapping[str, int | None]:
# Although vLLM can support more images from an infra capability
# perspective, we do not recommend using >10 images in practice.
return {"image": None}
@@ -699,7 +699,7 @@ class Mllama4DummyInputsBuilder(BaseDummyInputsBuilder[Mllama4ProcessingInfo]):
self,
seq_len: int,
mm_counts: Mapping[str, int],
mm_options: Optional[Mapping[str, BaseDummyOptions]] = None,
mm_options: Mapping[str, BaseDummyOptions] | None = None,
) -> MultiModalDataDict:
num_images = mm_counts.get("image", 0)
@@ -733,7 +733,7 @@ class Llama4ForConditionalGeneration(
supports_encoder_tp_data = True
@classmethod
def get_placeholder_str(cls, modality: str, i: int) -> Optional[str]:
def get_placeholder_str(cls, modality: str, i: int) -> str | None:
if modality.startswith("image"):
return "<|image|>"
@@ -792,7 +792,7 @@ class Llama4ForConditionalGeneration(
def _parse_and_validate_image_input(
self, **kwargs: object
) -> Optional[Llama4ImagePatchInputs]:
) -> Llama4ImagePatchInputs | None:
# num_images, 1, num_chunks, channel, image_size, image_size
pixel_values = kwargs.pop("pixel_values", None)
if pixel_values is None:
@@ -849,10 +849,10 @@ class Llama4ForConditionalGeneration(
self,
input_ids: torch.Tensor,
positions: torch.Tensor,
intermediate_tensors: Optional[IntermediateTensors] = None,
inputs_embeds: Optional[torch.Tensor] = None,
intermediate_tensors: IntermediateTensors | None = None,
inputs_embeds: torch.Tensor | None = None,
**kwargs: object,
) -> Union[torch.Tensor, IntermediateTensors]:
) -> torch.Tensor | IntermediateTensors:
if intermediate_tensors is not None:
inputs_embeds = None
@@ -863,7 +863,7 @@ class Llama4ForConditionalGeneration(
def compute_logits(
self,
hidden_states: torch.Tensor,
) -> Optional[torch.Tensor]:
) -> torch.Tensor | None:
return self.language_model.compute_logits(hidden_states)
def separate_weights(