Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

Fix attr(qkv int8) in multihead_pass #35504

Merged
merged 1 commit into from
Sep 7, 2021
Merged

Fix attr(qkv int8) in multihead_pass #35504

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
2 changes: 0 additions & 2 deletions paddle/fluid/framework/ir/multihead_matmul_fuse_pass.cc
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -903,8 +903,6 @@ int MultiHeadMatmulV2FusePass::BuildFusionV2(Graph* graph,
float, softmax_qk_op_desc->GetAttr("out_threshold"));
multihead_op_desc.SetAttr("dp_probs", qkv_plugin_scale);
}
} else {
multihead_op_desc.SetAttr("qkv2context_plugin_int8", false);
}

auto* multihead = graph->CreateOpNode(&multihead_op_desc);
Expand Down
3 changes: 1 addition & 2 deletions paddle/fluid/inference/tensorrt/convert/multihead_matmul_op.cc
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -42,8 +42,7 @@ class MultiheadMatMulOpConverter : public OpConverter {

float* weight_data = nullptr;
bool enable_int8 = op_desc.HasAttr("enable_int8");
bool qkv2context_plugin_int8 =
BOOST_GET_CONST(bool, op_desc.GetAttr("qkv2context_plugin_int8"));
bool qkv2context_plugin_int8 = op_desc.HasAttr("qkv2context_plugin_int8");
float in_scale = 0.;

if (enable_int8) {
Expand Down