Skip to content

N
ngraph
Switch branch/tag
History Find file
Name
Last commit
 Last update
..
files Yet another serialization option (#619)
models RNN Fusion using Pattern Matcher (#741)
ref_generators add enable gpu convolution tests flag in py (#898)
util Move runtime::Manager functionality into runtime::Backend (#875)
CMakeLists.txt nnpi and nnpi tester in unit test (#929)
algebraic_simplification.cpp Replace with Broadcast if exists in Algebraic Simplifier (#908)
autodiff.in.cpp Update to enable pass backend unit tests (#904)
backend_api.cpp Move runtime::Manager functionality into runtime::Backend (#875)
backend_debug_api.cpp Replace interpreter with Adam's simplified implementation (#915)
backend_performance.cpp Move runtime::Manager functionality into runtime::Backend (#875)
backend_test.in.cpp gpu select (#919)
build_graph.cpp get_input_op -> get_argument (#852)
builder.cpp Remove legacy Backend API (#848)
builder_autobroadcast.cpp Get value types out of public API, multi-values from Function (#340)
codegen.cpp set the EIGEN_MPL2_ONLY flag and add tests to make sure it is set for compile and codegen (#655)
convolution_test.in.cpp add enable gpu convolution tests flag in py (#898)
copy.cpp get_input_op -> get_argument (#852)
core_fusion.cpp Rename Any to Skip since that is exactly what it should have been called from the beginning (#891)
cpio.cpp Yet another serialization option (#619)
cpu_fusion.cpp Rename Any to Skip since that is exactly what it should have been called from the beginning (#891)
cpu_test.cpp CUDNN BatchNorm (inference/forward/backward) (#893)
cudnn.cpp GPU kernels for reshape, GEMM, EW ADD/Mult, Maximum (#440)
distributed.cpp Remove legacy Backend API (#848)
element_type.cpp Directory rename (#701)
file_util.cpp Add LICENSE and switch to Intel Copyright (#466)
graph_partition.cpp Fix cmake files and tests to allow building ngraph without the CPU backend (#910)
includes.cpp Rename Any to Skip since that is exactly what it should have been called from the beginning (#891)
inliner.cpp Replace using aliases with actual classes (#428)
input_output_assign.cpp get_input_op -> get_argument (#852)
main.cpp add AllReduce op and MPI support (#425)
mkldnn.cpp fix warnings (#571)
ngraph.cpp Add LICENSE and switch to Intel Copyright (#466)
op.cpp Element Type simplification (#313)
pass_liveness.cpp op::Result ver3 (#594)
pass_manager.cpp Replace using aliases with actual classes (#428)
pass_memory_layout.cpp op::Result ver3 (#594)
pattern.cpp label on skip (#909)
reshape_elimination.cpp Rename Any to Skip since that is exactly what it should have been called from the beginning (#891)
serialize.cpp Remove legacy Backend API (#848)
shape.cpp Add LICENSE and switch to Intel Copyright (#466)
tensor.cpp Remove legacy Backend API (#848)
type_prop.cpp Add logical-and, logical-or ops (#892)
update_reference.sh Change convolution reference to work with f32 (#409)
util.cpp Remove legacy Backend API (#848)
uuid.cpp Add LICENSE and switch to Intel Copyright (#466)
zero_dim_tensor_elimination.cpp Zero Dimension Tensor Elimination (#617)