feat(EX): Algorithm Factor Replacement Engine — dlopen-based CUDA kernel dispatch
Factors: 0 (moe_topk_softmax), 2 (moe_fused_gemm), 5 (gdn_chunk_fwd) Fixes: topk_softmax fallback (2304x/token), GDN NaN (frac=0.9998-1.0)
This commit is contained in:
3
ex_engine/python/__init__.py
Normal file
3
ex_engine/python/__init__.py
Normal file
@@ -0,0 +1,3 @@
|
||||
from .ex_loader import EXEngine, get_engine
|
||||
|
||||
__all__ = ["EXEngine", "get_engine"]
|
||||
Reference in New Issue
Block a user