[Model] Introduce Kimi Linear to vLLM (#27809)

Signed-off-by: lizhiyuan <lizhiyuan@moonshot.cn> Signed-off-by: Zhiyuan Li <uniartisan2017@gmail.com>
2025-10-30 21:02:27 +08:00
parent 1994de99ea
commit 4e68cc9b6a
15 changed files with 1325 additions and 48 deletions
--- a/vllm/config/compilation.py
+++ b/vllm/config/compilation.py
@@ -453,6 +453,7 @@ class CompilationConfig:
        "vllm::linear_attention",
        "vllm::plamo2_mamba_mixer",
        "vllm::gdn_attention",
+        "vllm::kda_attention",
        "vllm::sparse_attn_indexer",
    ]

--- a/vllm/config/model.py
+++ b/vllm/config/model.py
@@ -1236,6 +1236,7 @@ class ModelConfig:
            "deepseek_v32",
            "deepseek_mtp",
            "kimi_k2",
+            "kimi_linear",
            "longcat_flash",
        ):
            return self.hf_text_config.kv_lora_rank is not None