Skip to content

vllm.models.kimi_k3.nvidia.ops.cute_dsl

Modules:

  • kda_skinny_gemm –

    Kimi-K3 TP8 skinny GEMMs for the KDA F_A/beta and F_B projections.

  • latent_moe_tail –

    CuTe DSL kernels for KimiK3LatentMoETailOp.