未验证 提交 5fb8c920 编写于 作者: P Pei Yang 提交者: GitHub

fix multihead matmul shared params (#27121)

上级 d6ee0868
...@@ -615,6 +615,16 @@ static int BuildFusionV2(Graph* graph, const std::string& name_scope, ...@@ -615,6 +615,16 @@ static int BuildFusionV2(Graph* graph, const std::string& name_scope,
GET_IR_NODE_FROM_SUBGRAPH(transpose2_qkv_out, transpose2_qkv_out, GET_IR_NODE_FROM_SUBGRAPH(transpose2_qkv_out, transpose2_qkv_out,
multihead_pattern); multihead_pattern);
// If weights or biases in qkv's fc are shared by multiple multihead_matmul
// patterns, we do not support this kind of fusion, this pass will not take
// effect.
bool is_fc_params_shared =
mul0_w->outputs.size() > 1 || mul1_w->outputs.size() > 1 ||
mul2_w->outputs.size() > 1 || eltadd0_b->outputs.size() > 1 ||
eltadd1_b->outputs.size() > 1 || eltadd2_b->outputs.size() > 1;
if (is_fc_params_shared) {
return;
}
fuse_creater(input0, mul0, mul1, mul2, mul0_out, mul1_out, mul2_out, mul0_w, fuse_creater(input0, mul0, mul1, mul2, mul0_out, mul1_out, mul2_out, mul0_w,
mul1_w, mul2_w, eltadd0_b, eltadd1_b, eltadd2_b, eltadd_qk_b, mul1_w, mul2_w, eltadd0_b, eltadd1_b, eltadd2_b, eltadd_qk_b,
reshape2_0, reshape2_qkv_out, scale, scale_out); reshape2_0, reshape2_qkv_out, scale, scale_out);
......
Markdown is supported
0% .
You are about to add 0 people to the discussion. Proceed with caution.
先完成此消息的编辑!
想要评论请 注册