Skip to content

P
Paddle

PaddlePaddle / Paddle
大约 1 年 前同步成功

通知 2299
Star 20931
Fork 5422
P
Paddle
未验证 提交 9dac1e71 编写于 作者: G gem5 提交者： GitHub
浏览文件
操作

Enable identity_scale_op_clean_pass by default (#49227)

上级 85e92421
显示空白变更内容
内联 并排
Showing with 3 addition and 2 deletion
paddle/fluid/inference/api/paddle_pass_builder.cc
浏览文件 @ 9dac1e71
...@@ -162,6 +162,7 @@ const std::vector<std::string> kLiteSubgraphPasses({ ...@@ -162,6 +162,7 @@ const std::vector<std::string> kLiteSubgraphPasses({
// support fp16/bf16 precision, temporarily use low precision pass to prevent // support fp16/bf16 precision, temporarily use low precision pass to prevent
// running errors. After fusion operator supports low precision, delete this. // running errors. After fusion operator supports low precision, delete this.
const std::vector<std::string> kGpuLowerPrecisionPasses{ const std::vector<std::string> kGpuLowerPrecisionPasses{
"identity_scale_op_clean_pass",
"simplify_with_basic_ops_pass", "simplify_with_basic_ops_pass",
"delete_quant_dequant_linear_op_pass", "delete_quant_dequant_linear_op_pass",
"delete_weight_dequant_linear_op_pass", "delete_weight_dequant_linear_op_pass",
...@@ -208,7 +209,7 @@ const std::vector<std::string> kCINNCompilerPasses{ ...@@ -208,7 +209,7 @@ const std::vector<std::string> kCINNCompilerPasses{
GpuPassStrategy::GpuPassStrategy() : PassStrategy({}) { GpuPassStrategy::GpuPassStrategy() : PassStrategy({}) {
passes_.assign({ passes_.assign({
// "identity_scale_op_clean_pass", // "identity_scale_op_clean_pass", //
"is_test_pass", // "is_test_pass", //
"simplify_with_basic_ops_pass", // "simplify_with_basic_ops_pass", //
"delete_quant_dequant_linear_op_pass", // "delete_quant_dequant_linear_op_pass", //
......
Markdown is supported
0% 或 .
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 或