[CPU][BugFix] Multiple fixes to w4a8_int8 CPU MoE path (#46739)

Signed-off-by: Fadi Arafeh <fadi.arafeh@arm.com>
This commit is contained in:
Fadi Arafeh
2026-07-06 05:39:20 +00:00
committed by GitHub
parent 394edc8108
commit f1073c050c
10 changed files with 189 additions and 187 deletions
+2 -1
View File
@@ -278,7 +278,8 @@ TORCH_LIBRARY_EXPAND(TORCH_EXTENSION_NAME, ops) {
ops.def(
"dynamic_4bit_int_moe("
"Tensor x, Tensor topk_ids, Tensor topk_weights,"
"Tensor w13_packed, Tensor w2_packed, int H, int I, int I2,"
"Tensor w13_packed, Tensor w2_packed,"
"int hidden_size, int intermediate_size,"
"int group_size, bool apply_router_weight_on_input, int activation_kind"
") -> Tensor");