| 12345678910111213141516171819202122232425262728293031323334353637383940 |
- """Attention ops for b12x.
- - ``paged``: paged-KV self-attention (decode + extend, FP8 KV, MSA
- block-sparse variant) with on-device graph-replay metadata staging.
- - ``dense_mla``: dense compressed-cache MLA with strided physical records and
- optional causal sliding-window masking.
- - ``sparse_mla``: top-k-selected MLA, including strided physical records.
- - ``compressed_mla``: MLA decode directly from compressed KV pages (DSV4).
- - ``nsa_indexer``: the NSA index stage — quantize -> score -> select.
- - ``varlen``: contiguous batched/varlen attention (reduced-assurance tier).
- """
- from __future__ import annotations
- import importlib
- from typing import Any
- _OP_MODULES = (
- "paged",
- "dense_mla",
- "sparse_mla",
- "compressed_mla",
- "nsa_indexer",
- "varlen",
- )
- def __getattr__(name: str) -> Any:
- if name in _OP_MODULES:
- module = importlib.import_module(f".{name}", __name__)
- globals()[name] = module
- return module
- raise AttributeError(f"module {__name__!r} has no attribute {name!r}")
- def __dir__() -> list[str]:
- return sorted(_OP_MODULES)
- __all__ = list(_OP_MODULES)
|
