- Rename ops.py to attention_ops.py - Remove layout/scale params: fixed blhd, auto scale - Replace causal_offset with is_causal bool - Move SDPA fallback to backend, ops only calls CUDA kernels - Update __init__.py exports
43 lines
1.2 KiB
Python
43 lines
1.2 KiB
Python
"""CUDA attention kernel wrappers with torch fallback.
|
|
|
|
Public API:
|
|
- ``attn_decode`` — single-query decode attention
|
|
- ``attn_prefill`` — multi-query prefill attention
|
|
- ``attn_paged_decode`` — paged decode attention (direct page-table access)
|
|
- ``AttentionBackend`` — ABC for attention computation strategies
|
|
- ``TorchNativeBackend`` — default SDPA backend with KV cache I/O
|
|
|
|
Layout convention: all q/k/v are ``[batch, seq_len, n_heads, head_dim]``
|
|
(blhd). Scale is always ``1/sqrt(head_dim)``.
|
|
|
|
Each wrapper dispatches to its compiled CUDA kernel (``astrai.extension.attn_*``)
|
|
when available, otherwise falls back to ``torch.nn.functional.scaled_dot_product_attention``.
|
|
"""
|
|
|
|
from astrai.extension.attention_backend import (
|
|
ATTN_BACKEND,
|
|
AttentionBackend,
|
|
TorchNativeBackend,
|
|
attn_backend,
|
|
get_backend,
|
|
)
|
|
from astrai.extension.attention_ops import (
|
|
attn_decode,
|
|
attn_paged_decode,
|
|
attn_prefill,
|
|
)
|
|
from astrai.extension.loader import KERNEL_NAMES, is_available
|
|
|
|
__all__ = [
|
|
"ATTN_BACKEND",
|
|
"AttentionBackend",
|
|
"TorchNativeBackend",
|
|
"attn_backend",
|
|
"get_backend",
|
|
"attn_decode",
|
|
"attn_paged_decode",
|
|
"attn_prefill",
|
|
"is_available",
|
|
"KERNEL_NAMES",
|
|
]
|