[MoE] Move various experts classes to fused_moe/experts/ (#41979)

Signed-off-by: Jackmin801 <ongjackm@gmail.com>
Signed-off-by: Robert Shaw <robertgshaw2@gmail.com>
Signed-off-by: Jackmin801 <56836461+Jackmin801@users.noreply.github.com>
Signed-off-by: Bill Nell <bnell@redhat.com>
Co-authored-by: Jackmin801 <ongjackm@gmail.com>
Co-authored-by: Claude <noreply@anthropic.com>
Co-authored-by: Robert Shaw <robertgshaw2@gmail.com>
Co-authored-by: Robert Shaw <114415538+robertgshaw2-redhat@users.noreply.github.com>
Co-authored-by: Jackmin801 <56836461+Jackmin801@users.noreply.github.com>
This commit is contained in:
bnellnm
2026-05-11 07:54:33 +08:00
committed by GitHub
co-authored by Jackmin801 Claude Robert Shaw Robert Shaw Jackmin801
parent 21943d4c25
commit 1b57eb41f2
34 changed files with 586 additions and 553 deletions
+1 -1
View File
@@ -1,7 +1,7 @@
#include <cstdint>
/* Adapted from ./csrc/quantization/gguf/mmq.cuh
based on ./vllm/model_executor/layers/fused_moe/fused_moe.py */
based on ./vllm/model_executor/layers/fused_moe/experts/triton_moe.py */
template <typename scalar_t, int qk, int qr, int qi, bool need_sum,
typename block_q_t, int mmq_x, int mmq_y, int nwarps,
allocate_tiles_cuda_t allocate_tiles, load_tiles_cuda_t load_tiles,