Skip to content

vllm.models.muse_spark.nvidia.specialized_decode.cutedsl.attention

Modules:

  • attn2 –

    Muse Spark paged decode attention of the 8-token verify step, fused with the

  • kv_format –

    The KV cache formats the QKV front writes and attn2 reads, head_dim 64 and

  • qkv_front –

    The Muse Spark attention front for decode batches (1 <= M <= 16) in one