Add new moe wna16 marlin gemm (#14122)

This commit is contained in:
Xiaoyu Zhang
2025-12-01 23:07:53 +08:00
committed by GitHub
parent 2e8f54e61e
commit 3de09aadbc
15 changed files with 823 additions and 876 deletions
+2 -1
View File
@@ -299,7 +299,8 @@ TORCH_LIBRARY_FRAGMENT(sgl_kernel, m) {
*/
m.def(
"moe_wna16_marlin_gemm(Tensor! a, Tensor? c_or_none,"
"Tensor! b_q_weight, Tensor! b_scales, Tensor? b_zeros_or_none,"
"Tensor! b_q_weight, Tensor? b_bias_or_none, Tensor! b_scales,"
"Tensor? global_scale_or_none, Tensor? b_zeros_or_none,"
"Tensor? g_idx_or_none, Tensor? perm_or_none, Tensor! workspace,"
"Tensor sorted_token_ids,"
"Tensor! expert_ids, Tensor! num_tokens_past_padded,"