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

Update deprecated type hinting in models (#18132)

Browse Source 
Signed-off-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>
This commit is contained in:
Harry Mellor
2025-05-15 06:06:50 +01:00
committed by GitHub
parent 83f74c698f
commit 26d0419309
130 changed files with 971 additions and 901 deletions
Download Patch File Download Diff File Expand all files Collapse all files

15
vllm/model_executor/models/granitemoehybrid.py
View File

@@ -1,7 +1,8 @@
# SPDX-License-Identifier: Apache-2.0
"""Inference-only GraniteMoeHybrid model."""
# Added by the IBM Team, 2025
from typing import Iterable, Optional, Set, Tuple
from collections.abc import Iterable
from typing import Optional
import torch
from torch import nn
@@ -381,10 +382,10 @@ class GraniteMoeHybridModel(nn.Module):
hidden_states = self.norm(hidden_states)
return hidden_states
def load_weights(self, weights: Iterable[Tuple[str,
torch.Tensor]]) -> Set[str]:
def load_weights(self, weights: Iterable[tuple[str,
torch.Tensor]]) -> set[str]:
params_dict = dict(self.named_parameters())
loaded_params: Set[str] = set()
loaded_params: set[str] = set()
def _load(n, p):
param = params_dict[n]
@@ -538,7 +539,7 @@ class GraniteMoeHybridForCausalLM(nn.Module, HasInnerState, SupportsLoRA,
return self.mamba_cache.get_seqlen_agnostic_capture_inputs(batch_size)
def _get_mamba_cache_shape(
self) -> Tuple[Tuple[int, int], Tuple[int, int]]:
self) -> tuple[tuple[int, int], tuple[int, int]]:
world_size = get_tensor_model_parallel_world_size()
hidden_size = self.config.hidden_size
@@ -578,7 +579,7 @@ class GraniteMoeHybridForCausalLM(nn.Module, HasInnerState, SupportsLoRA,
sampling_metadata)
return logits
def load_weights(self, weights: Iterable[Tuple[str,
torch.Tensor]]) -> Set[str]:
def load_weights(self, weights: Iterable[tuple[str,
torch.Tensor]]) -> set[str]:
loader = AutoWeightsLoader(self)
return loader.load_weights(weights)