[docs] Fix the docs build (#48008)
Signed-off-by: Martin Hickey <martin.hickey@ie.ibm.com> Co-authored-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>
Martin Hickey committed
Jul 8, 2026 at 23:47 UTC
6cf7b26bd4bff60bf378e1af14044280ac0d214c
2 files changed
+6
-5
vllm/model_executor/layers/fused_moe/layer.py
+4
-3
@@ -186,9 +186,10 @@ def FusedMoE(
186
has_bias: Whether expert layers have bias terms
187
is_sequence_parallel: Whether sequence parallelism is enabled
188
reduce_results: Whether to all-reduce the final output. Setting this
189
- to False (to fuse the all-reduce downstream) is only honored on the
190
- late-AR path.
191
- expert_mapping: Expert parameter mapping for weight loading
189
+ to False (to fuse the all-reduce downstream) is only honored on
190
+ the late-AR path.
191
+ ckpt_names: Checkpoint parameter name tuple (gate_proj, down_proj,
192
+ up_proj) used for weight loading
193
n_shared_experts: Number of shared experts to fuse into the routed
194
grouped GEMM (ROCm; requires aiter FSE or the router-append path)
195
router_logits_dtype: Data type for router logits buffers
vllm/v1/kv_offload/tiering/p2p/manager.py
+2
-2
@@ -11,7 +11,7 @@ from __future__ import annotations
11
import time
12
from collections.abc import Iterable, Sequence
13
from dataclasses import dataclass, field
14
-from typing import TYPE_CHECKING
14
+from typing import TYPE_CHECKING, Any
15
16
from typing_extensions import override
17
@@ -115,7 +115,7 @@ class P2PSecondaryTierManager(SecondaryTierManager):
115
port: int = 7777,
116
backends: list[str] | None = None,
117
num_threads: int = 4,
118
- **kwargs,
118
+ **kwargs: Any,
119
) -> None:
120
"""Initialize the P2P secondary tier manager.
121