update llga to 85405a9 to fix INT8 RetinaNet accuracy issue (#1032) · intel/intel-extension-for-pytorch@e0c719b

Commit

update llga to 85405a9 to fix INT8 RetinaNet accuracy issue (#1032)

Browse files

chunyuan-w committed Aug 5, 2022

1 parent e2fbae0 commit e0c719b

+0 −4		src/backend/dnnl/dnnl_backend.cpp
+0 −31		src/backend/dnnl/dnnl_op_def.hpp
+0 −2		src/backend/dnnl/dnnl_opset.hpp
+0 −2		src/backend/dnnl/internal_ops.hpp
+0 −41		src/backend/dnnl/patterns/eltwise_fusion.cpp
+0 −12		tests/benchdnn/inputs_graph/eltwise/test_graph_eltwise_int8_ci
+0 −255		tests/cpp/unit/backend/dnnl/test_compile.cpp
+0 −120		tests/cpp/unit/backend/dnnl/test_pass.cpp