No MxNetCpp.h

Good afternoon,
Could someone give a hand on compiling mxnet w/ C++ support?

I’ve tried to compile mxnet cpp w/o success, it compiles w/o error, but no MxNetCpp.h.
Tried linux_gpu.cmake and also making a config.mk with the following flags, anyway w/o cpp files:

USE_OPENCV = 1
USE_OPENMP = 1
USE_MKL_IF_AVAILABLE = 1
USE_MKLDNN = 1
USE_LAPACK = 1
USE_SSE = 1
USE_F16C = 1
USE_OPERATOR_TUNING = 1
USE_GPERFTOOLS = 1
USE_JEMALLOC = 1
USE_BLAS = mkl
USE_CUDA = 1
USE_CUDA_PATH = /usr/local/cuda
USE_CUDNN = 1
ENABLE_CUDA_RTC = 1
USE_NVTX = 1
USE_CPP_PACKAGE = 1
install_manifest.txt:
/usr/local/lib/libmxnet.so.2.0.0
/usr/local/lib/libmxnet.so.2
/usr/local/lib/libmxnet.so
/usr/local/include/dlpack/dlpack.h
/usr/local/include/dmlc/any.h
/usr/local/include/dmlc/array_view.h
/usr/local/include/dmlc/base.h
/usr/local/include/dmlc/blockingconcurrentqueue.h
/usr/local/include/dmlc/build_config.h
/usr/local/include/dmlc/build_config_default.h
/usr/local/include/dmlc/common.h
/usr/local/include/dmlc/concurrency.h
/usr/local/include/dmlc/concurrentqueue.h
/usr/local/include/dmlc/config.h
/usr/local/include/dmlc/data.h
/usr/local/include/dmlc/endian.h
/usr/local/include/dmlc/filesystem.h
/usr/local/include/dmlc/input_split_shuffle.h
/usr/local/include/dmlc/io.h
/usr/local/include/dmlc/json.h
/usr/local/include/dmlc/logging.h
/usr/local/include/dmlc/lua.h
/usr/local/include/dmlc/memory.h
/usr/local/include/dmlc/memory_io.h
/usr/local/include/dmlc/omp.h
/usr/local/include/dmlc/optional.h
/usr/local/include/dmlc/parameter.h
/usr/local/include/dmlc/recordio.h
/usr/local/include/dmlc/registry.h
/usr/local/include/dmlc/serializer.h
/usr/local/include/dmlc/strtonum.h
/usr/local/include/dmlc/threadediter.h
/usr/local/include/dmlc/thread_group.h
/usr/local/include/dmlc/thread_local.h
/usr/local/include/dmlc/timer.h
/usr/local/include/dmlc/type_traits.h
/usr/local/include/mshadow/base.h
/usr/local/include/mshadow/bfloat.h
/usr/local/include/mshadow/cuda/reduce.cuh
/usr/local/include/mshadow/cuda/tensor_gpu-inl.cuh
/usr/local/include/mshadow/dot_engine-inl.h
/usr/local/include/mshadow/expression.h
/usr/local/include/mshadow/expr_engine-inl.h
/usr/local/include/mshadow/expr_scalar-inl.h
/usr/local/include/mshadow/extension/broadcast.h
/usr/local/include/mshadow/extension/broadcast_with_axis.h
/usr/local/include/mshadow/extension/channel_pool.h
/usr/local/include/mshadow/extension/channel_unpool.h
/usr/local/include/mshadow/extension/choose.h
/usr/local/include/mshadow/extension/complex.h
/usr/local/include/mshadow/extension/concat.h
/usr/local/include/mshadow/extension/crop.h
/usr/local/include/mshadow/extension/fill.h
/usr/local/include/mshadow/extension/flip.h
/usr/local/include/mshadow/extension/implicit_gemm.h
/usr/local/include/mshadow/extension/mask.h
/usr/local/include/mshadow/extension/mirror.h
/usr/local/include/mshadow/extension/one_hot.h
/usr/local/include/mshadow/extension/pack_col2patch.h
/usr/local/include/mshadow/extension/pad.h
/usr/local/include/mshadow/extension/range.h
/usr/local/include/mshadow/extension/reduceto1d.h
/usr/local/include/mshadow/extension/reduce_with_axis.h
/usr/local/include/mshadow/extension/reshape.h
/usr/local/include/mshadow/extension/slice.h
/usr/local/include/mshadow/extension/slice_ex.h
/usr/local/include/mshadow/extension/spatial_pool.h
/usr/local/include/mshadow/extension/spatial_unpool.h
/usr/local/include/mshadow/extension/spatial_upsampling_nearest.h
/usr/local/include/mshadow/extension/swapaxis.h
/usr/local/include/mshadow/extension/take.h
/usr/local/include/mshadow/extension/take_grad.h
/usr/local/include/mshadow/extension/transpose.h
/usr/local/include/mshadow/extension/unpack_patch2col.h
/usr/local/include/mshadow/extension.h
/usr/local/include/mshadow/half.h
/usr/local/include/mshadow/half2.h
/usr/local/include/mshadow/io.h
/usr/local/include/mshadow/logging.h
/usr/local/include/mshadow/packet/plain-inl.h
/usr/local/include/mshadow/packet/sse-inl.h
/usr/local/include/mshadow/packet-inl.h
/usr/local/include/mshadow/random.h
/usr/local/include/mshadow/README.md
/usr/local/include/mshadow/stream_gpu-inl.h
/usr/local/include/mshadow/tensor.h
/usr/local/include/mshadow/tensor_container.h
/usr/local/include/mshadow/tensor_cpu-inl.h
/usr/local/include/mshadow/tensor_gpu-inl.h
/usr/local/include/mxnet/api_registry.h
/usr/local/include/mxnet/base.h
/usr/local/include/mxnet/c_api.h
/usr/local/include/mxnet/c_api_error.h
/usr/local/include/mxnet/c_api_test.h
/usr/local/include/mxnet/engine.h
/usr/local/include/mxnet/executor.h
/usr/local/include/mxnet/expr_operator.h
/usr/local/include/mxnet/graph_attr_types.h
/usr/local/include/mxnet/imperative.h
/usr/local/include/mxnet/io.h
/usr/local/include/mxnet/ir/expr.h
/usr/local/include/mxnet/kvstore.h
/usr/local/include/mxnet/libinfo.h
/usr/local/include/mxnet/lib_api.h
/usr/local/include/mxnet/ndarray.h
/usr/local/include/mxnet/node/container.h
/usr/local/include/mxnet/node/node.h
/usr/local/include/mxnet/operator.h
/usr/local/include/mxnet/operator_util.h
/usr/local/include/mxnet/op_attr_types.h
/usr/local/include/mxnet/random_generator.h
/usr/local/include/mxnet/resource.h
/usr/local/include/mxnet/rtc.h
/usr/local/include/mxnet/runtime/container.h
/usr/local/include/mxnet/runtime/c_runtime_api.h
/usr/local/include/mxnet/runtime/data_type.h
/usr/local/include/mxnet/runtime/ffi_helper.h
/usr/local/include/mxnet/runtime/memory.h
/usr/local/include/mxnet/runtime/ndarray.h
/usr/local/include/mxnet/runtime/ndarray_handle.h
/usr/local/include/mxnet/runtime/object.h
/usr/local/include/mxnet/runtime/packed_func.h
/usr/local/include/mxnet/runtime/py_arg.h
/usr/local/include/mxnet/runtime/registry.h
/usr/local/include/mxnet/storage.h
/usr/local/include/mxnet/tensor_blob.h
/usr/local/include/mxnet/tuple.h
/usr/local/include/nnvm/base.h
/usr/local/include/nnvm/c_api.h
/usr/local/include/nnvm/graph.h
/usr/local/include/nnvm/graph_attr_types.h
/usr/local/include/nnvm/layout.h
/usr/local/include/nnvm/node.h
/usr/local/include/nnvm/op.h
/usr/local/include/nnvm/op_attr_types.h
/usr/local/include/nnvm/pass.h
/usr/local/include/nnvm/pass_functions.h
/usr/local/include/nnvm/symbolic.h
/usr/local/include/nnvm/tuple.h
/usr/local/share/doc/dnnl/reference/html/assets/mathjax/config/dnnl.js
/usr/local/lib/libdnnl.a
/usr/local/include/mkldnn/dnnl.h
/usr/local/include/mkldnn/dnnl_config.h
/usr/local/include/mkldnn/dnnl_debug.h
/usr/local/include/mkldnn/dnnl_types.h
/usr/local/include/mkldnn/dnnl_version.h
/usr/local/include/mkldnn/mkldnn.h
/usr/local/include/mkldnn/mkldnn_config.h
/usr/local/include/mkldnn/mkldnn_debug.h
/usr/local/include/mkldnn/mkldnn_dnnl_mangling.h
/usr/local/include/mkldnn/mkldnn_types.h
/usr/local/include/mkldnn/mkldnn_version.h
/usr/local/include/mkldnn/dnnl.h
/usr/local/include/mkldnn/dnnl.hpp
/usr/local/include/mkldnn/dnnl_config.h
/usr/local/include/mkldnn/dnnl_debug.h
/usr/local/include/mkldnn/dnnl_threadpool_iface.hpp
/usr/local/include/mkldnn/dnnl_types.h
/usr/local/include/mkldnn/dnnl_version.h
/usr/local/include/mkldnn/mkldnn.h
/usr/local/include/mkldnn/mkldnn.hpp
/usr/local/include/mkldnn/mkldnn_config.h
/usr/local/include/mkldnn/mkldnn_debug.h
/usr/local/include/mkldnn/mkldnn_dnnl_mangling.h
/usr/local/include/mkldnn/mkldnn_types.h
/usr/local/include/mkldnn/mkldnn_version.h
/usr/local/lib/cmake/dnnl/dnnl-config.cmake
/usr/local/lib/cmake/dnnl/dnnl-config-version.cmake
/usr/local/lib/cmake/dnnl/dnnl-targets.cmake
/usr/local/lib/cmake/dnnl/dnnl-targets-noconfig.cmake
/usr/local/lib/libmkldnn.a
/usr/local/share/doc/dnnl/LICENSE
/usr/local/share/doc/dnnl/THIRD-PARTY-PROGRAMS
/usr/local/share/doc/dnnl/README
/usr/local/lib/libomp.so
/usr/local/include/omp.h
/usr/local/include/omp-tools.h
/usr/local/include/ompt.h
/usr/local/include/dmlc/any.h
/usr/local/include/dmlc/array_view.h
/usr/local/include/dmlc/base.h
/usr/local/include/dmlc/blockingconcurrentqueue.h
/usr/local/include/dmlc/build_config.h
/usr/local/include/dmlc/build_config_default.h
/usr/local/include/dmlc/common.h
/usr/local/include/dmlc/concurrency.h
/usr/local/include/dmlc/concurrentqueue.h
/usr/local/include/dmlc/config.h
/usr/local/include/dmlc/data.h
/usr/local/include/dmlc/endian.h
/usr/local/include/dmlc/filesystem.h
/usr/local/include/dmlc/input_split_shuffle.h
/usr/local/include/dmlc/io.h
/usr/local/include/dmlc/json.h
/usr/local/include/dmlc/logging.h
/usr/local/include/dmlc/lua.h
/usr/local/include/dmlc/memory.h
/usr/local/include/dmlc/memory_io.h
/usr/local/include/dmlc/omp.h
/usr/local/include/dmlc/optional.h
/usr/local/include/dmlc/parameter.h
/usr/local/include/dmlc/recordio.h
/usr/local/include/dmlc/registry.h
/usr/local/include/dmlc/serializer.h
/usr/local/include/dmlc/strtonum.h
/usr/local/include/dmlc/threadediter.h
/usr/local/include/dmlc/thread_group.h
/usr/local/include/dmlc/thread_local.h
/usr/local/include/dmlc/timer.h
/usr/local/include/dmlc/type_traits.h
/usr/local/include/dmlc/build_config.h
/usr/local/lib/libdmlc.a
/usr/local/lib/cmake/dmlc/DMLCTargets.cmake
/usr/local/lib/cmake/dmlc/DMLCTargets-noconfig.cmake
/usr/local/lib/cmake/dmlc/dmlc-config.cmake
/usr/local/lib/cmake/dmlc/dmlc-config-version.cmake

// Ubuntu 20 w/ latest Cuda (11.0.3-1) & CUDnn (8.02.39), Tensorrt 7.1.3.4-1
±----------------------------------------------------------------------------+
| NVIDIA-SMI 450.51.06 Driver Version: 450.51.06 CUDA Version: 11.0 |
|-------------------------------±---------------------±---------------------+
| GPU Name Persistence-M| Bus-Id Disp.A | Volatile Uncorr. ECC |
| Fan Temp Perf Pwr:Usage/Cap| Memory-Usage | GPU-Util Compute M. |
| | | MIG M. |
|===============================+======================+======================|
| 0 GeForce RTX 2060 On | 00000000:01:00.0 Off | N/A |
| N/A 48C P8 3W / N/A | 3MiB / 5934MiB | 0% Default |
| | | N/A |
±------------------------------±---------------------±---------------------+

±----------------------------------------------------------------------------+
| Processes: |
| GPU GI CI PID Type Process name GPU Memory |
| ID ID Usage |
|=============================================================================|
| No running processes found |
±----------------------------------------------------------------------------+

//