colin
/
TIM-VX
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

Fixed wrong type of hashtable_lookup unit test 

Modified datatype of copied output so that actual float16 output can be
checked  correctly

Type: Bug Fix
Issue: Bug 34696
Signed-off-by: Feiyue Chen <Feiyue.Chen@verisilicon.com>
This commit is contained in:
Feiyue Chen 2022-12-28 15:12:54 +08:00 committed by Sven
parent cc34b5f0ea
commit 06d4747b31
1 changed files with 3 additions and 3 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

6
src/tim/vx/ops/hashtable_lookup_test.cc
View File

@ -61,13 +61,13 @@ TEST(HashtableLookup, int32) {
std::vector<int32_t> value_data = {
7,8,9,10
};
std::vector<float> out_golden = {7,0};
std::vector<uint16_t> out_golden = {18176,0};//check for float16 output
std::vector<int32_t> hit_golden = {1,0};//0 means not hit
EXPECT_TRUE(in_tensor->CopyDataToTensor(in_data.data(), in_data.size() * sizeof(int32_t)));
EXPECT_TRUE(key_tensor->CopyDataToTensor(key_data.data(), key_data.size() * sizeof(int32_t)));
EXPECT_TRUE(value_tensor->CopyDataToTensor(value_data.data(), value_data.size() * sizeof(int32_t)));
EXPECT_TRUE(out_tensor->CopyDataToTensor(out_golden.data(), out_golden.size() * sizeof(float)));
EXPECT_TRUE(out_tensor->CopyDataToTensor(out_golden.data(), out_golden.size()));
EXPECT_TRUE(hit_tensor->CopyDataToTensor(hit_golden.data(), hit_golden.size() * sizeof(int32_t)));
auto op = graph->CreateOperation<tim::vx::ops::HashtableLookup>();
@ -76,7 +76,7 @@ TEST(HashtableLookup, int32) {
EXPECT_TRUE(graph->Compile());
EXPECT_TRUE(graph->Run());
std::vector<float> output(out_golden.size());
std::vector<uint16_t> output(out_golden.size());
std::vector<int32_t> hit(hit_golden.size());
EXPECT_TRUE(out_tensor->CopyDataFromTensor(output.data()));
EXPECT_TRUE(hit_tensor->CopyDataFromTensor(hit.data()));