A high-throughput and memory-efficient inference and serving engine for LLMs
Mirrored from https://github.com/tduncan-lum/Nvidia-Optimization-Stack---VLLM.git
Synced Nov 11, 22:30 UTC
fix#30092 Kimi-Linear model loading failure with missing indexer_rotary_emb (#30093)
Signed-off-by: baonudesifeizhai <baonudesifeizhai@gmail.com>
Dongjie Zou committed
Dec 5, 2025 at 23:55 UTC
e3fbb6f152fe721506f22b3d8d2ec70c10229569
1 file changed
+1
-1
vllm/model_executor/layers/mla.py
+1
-1
@@ -24,9 +24,9 @@ class MLAModules:
24
q_b_proj: torch.nn.Module | None
25
q_proj: torch.nn.Module | None
26
indexer: torch.nn.Module | None
27
- indexer_rotary_emb: torch.nn.Module | None
27
is_sparse: bool
28
topk_indices_buffer: torch.Tensor | None
29
+ indexer_rotary_emb: torch.nn.Module | None = None
30
31
32
@CustomOp.register("multi_head_latent_attention")