Please help - Trouble compiling TF 1.14 CP37 GPU Cuda and TensorRT

Hi, I have lost count of how many times I have reflashed my jetson Xavier and reinstalled everything to be able to get TF 1.14 using Cuda, TensorRT and the compilation still fails after hours :((( I really need this to work with python3.7 because the industrial camera we bought only works with python 3.7 and above and the manufacturer doesnt have a driver that would work for python 3.6 or below. I am running Jetpack 4.3 didnt touch anything with the built in libraries. I added an SSD drive and started compiling TF 1.14 using an 8GB swap file. This time my boot drive is not overrun.

If anyone knows a wheel for TF 1.14 with GPU support that will work on the Xavier please let me know.

In the meantime, this is the error im getting:

ERROR: /media/xavier/storage/tensorflow/tensorflow/lite/kernels/BUILD:286:1: C++ compilation of rule ‘//tensorflow/lite/kernels:builtin_op_kernels’ failed (Exit 1): crosstool_wrapper_driver_is_not_gcc failed: error executing command
(cd /home/xavier/.cache/bazel/bazel_xavier/49531cc0ad8c2ca1aee55603d0e1d4ef/execroot/org_tensorflow &&
exec env -
CUDA_TOOLKIT_PATH=/usr/local/cuda
GCC_HOST_COMPILER_PATH=/usr/bin/gcc-5
LD_LIBRARY_PATH=/usr/local/cuda-10.0/lib64:
PATH=/home/xavier/bazel-0.24.1-dist/output:/usr/local/cuda-10.0/bin:/home/xavier/.local/bin:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin
PWD=/proc/self/cwd
PYTHON_BIN_PATH=/usr/bin/python3.7
PYTHON_LIB_PATH=/usr/local/lib/python3.7/dist-packages
TF_CONFIGURE_IOS=0
TF_CUDA_CLANG=0
TF_CUDA_COMPUTE_CAPABILITIES=7.2
TF_NEED_CUDA=1
TF_NEED_OPENCL_SYCL=0
TF_NEED_ROCM=0
TF_NEED_TENSORRT=1
external/local_config_cuda/crosstool/clang/bin/crosstool_wrapper_driver_is_not_gcc -MD -MF bazel-out/aarch64-opt/bin/tensorflow/lite/kernels/objs/builtin_op_kernels/depthwise_conv.pic.d '-frandom-seed=bazel-out/aarch64-opt/bin/tensorflow/lite/kernels/objs/builtin_op_kernels/depthwise_conv.pic.o’ -DEIGEN_MPL2_ONLY ‘-DEIGEN_MAX_ALIGN_BYTES=64’ ‘-DEIGEN_HAS_TYPE_TRAITS=0’ -iquote . -iquote bazel-out/aarch64-opt/genfiles -iquote bazel-out/aarch64-opt/bin -iquote external/gemmlowp -iquote bazel-out/aarch64-opt/genfiles/external/gemmlowp -iquote bazel-out/aarch64-opt/bin/external/gemmlowp -iquote external/com_google_absl -iquote bazel-out/aarch64-opt/genfiles/external/com_google_absl -iquote bazel-out/aarch64-opt/bin/external/com_google_absl -iquote external/arm_neon_2_x86_sse -iquote bazel-out/aarch64-opt/genfiles/external/arm_neon_2_x86_sse -iquote bazel-out/aarch64-opt/bin/external/arm_neon_2_x86_sse -iquote external/eigen_archive -iquote bazel-out/aarch64-opt/genfiles/external/eigen_archive -iquote bazel-out/aarch64-opt/bin/external/eigen_archive -iquote external/local_config_sycl -iquote bazel-out/aarch64-opt/genfiles/external/local_config_sycl -iquote bazel-out/aarch64-opt/bin/external/local_config_sycl -iquote external/flatbuffers -iquote bazel-out/aarch64-opt/genfiles/external/flatbuffers -iquote bazel-out/aarch64-opt/bin/external/flatbuffers -iquote external/fft2d -iquote bazel-out/aarch64-opt/genfiles/external/fft2d -iquote bazel-out/aarch64-opt/bin/external/fft2d -iquote external/farmhash_archive -iquote bazel-out/aarch64-opt/genfiles/external/farmhash_archive -iquote bazel-out/aarch64-opt/bin/external/farmhash_archive -isystem external/eigen_archive -isystem bazel-out/aarch64-opt/genfiles/external/eigen_archive -isystem bazel-out/aarch64-opt/bin/external/eigen_archive -isystem tensorflow/lite/schema -isystem bazel-out/aarch64-opt/genfiles/tensorflow/lite/schema -isystem bazel-out/aarch64-opt/bin/tensorflow/lite/schema -isystem external/flatbuffers/include -isystem bazel-out/aarch64-opt/genfiles/external/flatbuffers/include -isystem bazel-out/aarch64-opt/bin/external/flatbuffers/include -isystem external/farmhash_archive/src -isystem bazel-out/aarch64-opt/genfiles/external/farmhash_archive/src -isystem bazel-out/aarch64-opt/bin/external/farmhash_archive/src ‘-std=c++11’ -Wno-builtin-macro-redefined '-D__DATE=“redacted”’ '-D__TIMESTAMP
_=“redacted”’ ‘-D__TIME__=“redacted”’ -fPIC -U_FORTIFY_SOURCE ‘-D_FORTIFY_SOURCE=1’ -fstack-protector -Wall -fno-omit-frame-pointer -no-canonical-prefixes -fno-canonical-system-headers -DNDEBUG -g0 -O2 -ffunction-sections -fdata-sections ‘-march=native’ -Wno-sign-compare ‘-D_GLIBCXX_USE_CXX11_ABI=0’ -DFARMHASH_NO_CXX_STRING -Wno-sign-compare ‘-Wno-error=reorder’ -c tensorflow/lite/kernels/depthwise_conv.cc -o bazel-out/aarch64-opt/bin/tensorflow/lite/kernels/_objs/builtin_op_kernels/depthwise_conv.pic.o)
Execution platform: @bazel_tools//platforms:host_platform
In file included from ./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8.h:23:0,
from ./tensorflow/lite/kernels/internal/optimized/depthwiseconv_multithread.h:21,
from tensorflow/lite/kernels/depthwise_conv.cc:28:
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h: In static member function ‘static void tflite::optimized_ops::depthwise_conv::PackMacroBlock<(tflite::DepthwiseConvImplementation)3, (tflite::DepthwiseConvDepthMultiplication)0, 0>::PackMacroBlockNeon(const uint8*, int8*, const tflite::optimized_ops::depthwise_conv::DepthwiseConvDotProdParams*)’:
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:5981:24: note: use -flax-vector-conversions to permit conversions between vectors with differing element types or numbers of subparts
input_data_a = vld1q_u8(input_data_0);
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:5981:24: error: cannot convert ‘uint8x16_t {aka __vector(16) unsigned char}’ to ‘int8x16_t {aka __vector(16) signed char}’ in assignment
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:5982:24: error: cannot convert ‘uint8x16_t {aka __vector(16) unsigned char}’ to ‘int8x16_t {aka __vector(16) signed char}’ in assignment
input_data_b = vld1q_u8(input_data_0 + 1 * input_depth);
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:5983:24: error: cannot convert ‘uint8x16_t {aka __vector(16) unsigned char}’ to ‘int8x16_t {aka __vector(16) signed char}’ in assignment
input_data_c = vld1q_u8(input_data_0 + 2 * input_depth);
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:5984:24: error: cannot convert ‘uint8x16_t {aka __vector(16) unsigned char}’ to ‘int8x16_t {aka __vector(16) signed char}’ in assignment
input_data_d = vld1q_u8(input_data_0 + 3 * input_depth);
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:5993:55: error: cannot convert ‘const uint8x16_t {aka const __vector(16) unsigned char}’ to ‘int8x16_t {aka __vector(16) signed char}’ for argument ‘2’ to ‘int8x16_t veorq_s8(int8x16_t, int8x16_t)’
work_reg_a = veorq_s8(work_reg_a, sign_bit);
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:5994:55: error: cannot convert ‘const uint8x16_t {aka const __vector(16) unsigned char}’ to ‘int8x16_t {aka __vector(16) signed char}’ for argument ‘2’ to ‘int8x16_t veorq_s8(int8x16_t, int8x16_t)’
work_reg_b = veorq_s8(work_reg_b, sign_bit);
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:6000:26: error: cannot convert ‘uint8x16_t {aka __vector(16) unsigned char}’ to ‘int8x16_t {aka __vector(16) signed char}’ in assignment
input_data_a = vld1q_u8(input_data_0);
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:6001:26: error: cannot convert ‘uint8x16_t {aka __vector(16) unsigned char}’ to ‘int8x16_t {aka __vector(16) signed char}’ in assignment
input_data_b = vld1q_u8(input_data_0 + 1 * input_depth);
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:6009:61: error: cannot convert ‘const uint8x16_t {aka const __vector(16) unsigned char}’ to ‘int8x16_t {aka __vector(16) signed char}’ for argument ‘2’ to ‘int8x16_t veorq_s8(int8x16_t, int8x16_t)’
work_reg_a_sp = veorq_s8(work_reg_a_sp, sign_bit);
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:6010:61: error: cannot convert ‘const uint8x16_t {aka const __vector(16) unsigned char}’ to ‘int8x16_t {aka __vector(16) signed char}’ for argument ‘2’ to ‘int8x16_t veorq_s8(int8x16_t, int8x16_t)’
work_reg_b_sp = veorq_s8(work_reg_b_sp, sign_bit);
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:6012:26: error: cannot convert ‘uint8x16_t {aka __vector(16) unsigned char}’ to ‘int8x16_t {aka __vector(16) signed char}’ in assignment
input_data_c = vld1q_u8(input_data_0 + 2 * input_depth);
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:6013:26: error: cannot convert ‘uint8x16_t {aka __vector(16) unsigned char}’ to ‘int8x16_t {aka __vector(16) signed char}’ in assignment
input_data_d = vld1q_u8(input_data_0 + 3 * input_depth);
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:6029:53: error: cannot convert ‘const uint8x16_t {aka const __vector(16) unsigned char}’ to ‘int8x16_t {aka __vector(16) signed char}’ for argument ‘2’ to ‘int8x16_t veorq_s8(int8x16_t, int8x16_t)’
work_reg_a = veorq_s8(work_reg_a, sign_bit);
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:6030:53: error: cannot convert ‘const uint8x16_t {aka const __vector(16) unsigned char}’ to ‘int8x16_t {aka __vector(16) signed char}’ for argument ‘2’ to ‘int8x16_t veorq_s8(int8x16_t, int8x16_t)’
work_reg_b = veorq_s8(work_reg_b, sign_bit);
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:6042:59: error: cannot convert ‘const uint8x16_t {aka const __vector(16) unsigned char}’ to ‘int8x16_t {aka __vector(16) signed char}’ for argument ‘2’ to ‘int8x16_t veorq_s8(int8x16_t, int8x16_t)’
work_reg_a_sp = veorq_s8(work_reg_a_sp, sign_bit);
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:6043:59: error: cannot convert ‘const uint8x16_t {aka const __vector(16) unsigned char}’ to ‘int8x16_t {aka __vector(16) signed char}’ for argument ‘2’ to ‘int8x16_t veorq_s8(int8x16_t, int8x16_t)’
work_reg_b_sp = veorq_s8(work_reg_b_sp, sign_bit);
^
In file included from ./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8.h:23:0,
from ./tensorflow/lite/kernels/internal/optimized/depthwiseconv_multithread.h:21,
from tensorflow/lite/kernels/depthwise_conv.cc:28:
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:46:71: error: cannot convert ‘int8x16_t {aka __vector(16) signed char}’ to ‘uint64x2_t {aka __vector(2) long unsigned int}’ for argument ‘2’ to ‘uint64x2_t vld1q_lane_u64(const uint64_t*, uint64x2_t, int)’
vld1q_lane_u64(reinterpret_cast<const uint64_t*>(src), reg, lane_num)
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:6053:26: note: in expansion of macro ‘vld1q_lane_s8x8’
input_data_a = vld1q_lane_s8x8(input_data_0, input_data_a, 0);
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:46:71: error: cannot convert ‘int8x16_t {aka __vector(16) signed char}’ to ‘uint64x2_t {aka __vector(2) long unsigned int}’ for argument ‘2’ to ‘uint64x2_t vld1q_lane_u64(const uint64_t*, uint64x2_t, int)’
vld1q_lane_u64(reinterpret_cast<const uint64_t*>(src), reg, lane_num)
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:6055:15: note: in expansion of macro ‘vld1q_lane_s8x8’
vld1q_lane_s8x8(input_data_0 + 1 * input_depth, input_data_b, 0);
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:46:71: error: cannot convert ‘int8x16_t {aka __vector(16) signed char}’ to ‘uint64x2_t {aka __vector(2) long unsigned int}’ for argument ‘2’ to ‘uint64x2_t vld1q_lane_u64(const uint64_t*, uint64x2_t, int)’
vld1q_lane_u64(reinterpret_cast<const uint64_t*>(src), reg, lane_num)
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:6057:15: note: in expansion of macro ‘vld1q_lane_s8x8’
vld1q_lane_s8x8(input_data_0 + 2 * input_depth, input_data_c, 0);
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:46:71: error: cannot convert ‘int8x16_t {aka __vector(16) signed char}’ to ‘uint64x2_t {aka __vector(2) long unsigned int}’ for argument ‘2’ to ‘uint64x2_t vld1q_lane_u64(const uint64_t*, uint64x2_t, int)’
vld1q_lane_u64(reinterpret_cast<const uint64_t*>(src), reg, lane_num)
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:6059:15: note: in expansion of macro ‘vld1q_lane_s8x8’
vld1q_lane_s8x8(input_data_0 + 3 * input_depth, input_data_d, 0);
^
In file included from ./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8.h:23:0,
from ./tensorflow/lite/kernels/internal/optimized/depthwiseconv_multithread.h:21,
from tensorflow/lite/kernels/depthwise_conv.cc:28:
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:6066:53: error: cannot convert ‘const uint8x16_t {aka const __vector(16) unsigned char}’ to ‘int8x16_t {aka __vector(16) signed char}’ for argument ‘2’ to ‘int8x16_t veorq_s8(int8x16_t, int8x16_t)’
work_reg_a = veorq_s8(work_reg_a, sign_bit);
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:6067:53: error: cannot convert ‘const uint8x16_t {aka const __vector(16) unsigned char}’ to ‘int8x16_t {aka __vector(16) signed char}’ for argument ‘2’ to ‘int8x16_t veorq_s8(int8x16_t, int8x16_t)’
work_reg_b = veorq_s8(work_reg_b, sign_bit);
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:6085:24: error: cannot convert ‘uint8x16_t {aka __vector(16) unsigned char}’ to ‘int8x16_t {aka __vector(16) signed char}’ in assignment
input_data_c = vdupq_n_u8(kSignBit);
^
In file included from ./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8.h:23:0,
from ./tensorflow/lite/kernels/internal/optimized/depthwiseconv_multithread.h:21,
from tensorflow/lite/kernels/depthwise_conv.cc:28:
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:46:71: error: cannot convert ‘int8x16_t {aka __vector(16) signed char}’ to ‘uint64x2_t {aka __vector(2) long unsigned int}’ for argument ‘2’ to ‘uint64x2_t vld1q_lane_u64(const uint64_t*, uint64x2_t, int)’
vld1q_lane_u64(reinterpret_cast<const uint64_t*>(src), reg, lane_num)
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:6086:26: note: in expansion of macro ‘vld1q_lane_s8x8’
input_data_a = vld1q_lane_s8x8(input_data_0, input_data_a, 0);
^
In file included from ./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8.h:23:0,
from ./tensorflow/lite/kernels/internal/optimized/depthwiseconv_multithread.h:21,
from tensorflow/lite/kernels/depthwise_conv.cc:28:
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:6087:24: error: cannot convert ‘uint8x16_t {aka __vector(16) unsigned char}’ to ‘int8x16_t {aka __vector(16) signed char}’ in assignment
input_data_d = vdupq_n_u8(kSignBit);
^
In file included from ./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8.h:23:0,
from ./tensorflow/lite/kernels/internal/optimized/depthwiseconv_multithread.h:21,
from tensorflow/lite/kernels/depthwise_conv.cc:28:
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:46:71: error: cannot convert ‘int8x16_t {aka __vector(16) signed char}’ to ‘uint64x2_t {aka __vector(2) long unsigned int}’ for argument ‘2’ to ‘uint64x2_t vld1q_lane_u64(const uint64_t*, uint64x2_t, int)’
vld1q_lane_u64(reinterpret_cast<const uint64_t*>(src), reg, lane_num)
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:6090:17: note: in expansion of macro ‘vld1q_lane_s8x8’
vld1q_lane_s8x8(input_data_0 + input_depth, input_data_b, 0);
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:46:71: error: cannot convert ‘int8x16_t {aka __vector(16) signed char}’ to ‘uint64x2_t {aka __vector(2) long unsigned int}’ for argument ‘2’ to ‘uint64x2_t vld1q_lane_u64(const uint64_t*, uint64x2_t, int)’
vld1q_lane_u64(reinterpret_cast<const uint64_t*>(src), reg, lane_num)
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:6092:30: note: in expansion of macro ‘vld1q_lane_s8x8’
input_data_c = vld1q_lane_s8x8(input_data_0 + 2 * input_depth,
^
In file included from ./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8.h:23:0,
from ./tensorflow/lite/kernels/internal/optimized/depthwiseconv_multithread.h:21,
from tensorflow/lite/kernels/depthwise_conv.cc:28:
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:6099:53: error: cannot convert ‘const uint8x16_t {aka const __vector(16) unsigned char}’ to ‘int8x16_t {aka __vector(16) signed char}’ for argument ‘2’ to ‘int8x16_t veorq_s8(int8x16_t, int8x16_t)’
work_reg_a = veorq_s8(work_reg_a, sign_bit);
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:6100:53: error: cannot convert ‘const uint8x16_t {aka const __vector(16) unsigned char}’ to ‘int8x16_t {aka __vector(16) signed char}’ for argument ‘2’ to ‘int8x16_t veorq_s8(int8x16_t, int8x16_t)’
work_reg_b = veorq_s8(work_reg_b, sign_bit);
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h: In static member function ‘static void tflite::optimized_ops::depthwise_conv::PackMacroBlock<(tflite::DepthwiseConvImplementation)3, (tflite::DepthwiseConvDepthMultiplication)0, 1>::PackMacroBlockNeon(int32, int32, const uint8*, int8*, const tflite::optimized_ops::depthwise_conv::DepthwiseConvDotProdParams*)’:
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:6244:28: error: cannot convert ‘uint8x16_t {aka __vector(16) unsigned char}’ to ‘int8x16_t {aka __vector(16) signed char}’ in assignment
input_data_a = vld1q_u8(input_data_0);
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:6245:28: error: cannot convert ‘uint8x16_t {aka __vector(16) unsigned char}’ to ‘int8x16_t {aka __vector(16) signed char}’ in assignment
input_data_b = vld1q_u8(input_data_0 + 1 * input_depth);
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:6246:28: error: cannot convert ‘uint8x16_t {aka __vector(16) unsigned char}’ to ‘int8x16_t {aka __vector(16) signed char}’ in assignment
input_data_c = vld1q_u8(input_data_0 + 2 * input_depth);
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:6247:28: error: cannot convert ‘uint8x16_t {aka __vector(16) unsigned char}’ to ‘int8x16_t {aka __vector(16) signed char}’ in assignment
input_data_d = vld1q_u8(input_data_0 + 3 * input_depth);
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:6256:59: error: cannot convert ‘const uint8x16_t {aka const __vector(16) unsigned char}’ to ‘int8x16_t {aka __vector(16) signed char}’ for argument ‘2’ to ‘int8x16_t veorq_s8(int8x16_t, int8x16_t)’
work_reg_a = veorq_s8(work_reg_a, sign_bit);
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:6257:59: error: cannot convert ‘const uint8x16_t {aka const __vector(16) unsigned char}’ to ‘int8x16_t {aka __vector(16) signed char}’ for argument ‘2’ to ‘int8x16_t veorq_s8(int8x16_t, int8x16_t)’
work_reg_b = veorq_s8(work_reg_b, sign_bit);
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:6263:30: error: cannot convert ‘uint8x16_t {aka __vector(16) unsigned char}’ to ‘int8x16_t {aka __vector(16) signed char}’ in assignment
input_data_a = vld1q_u8(input_data_0);
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:6264:30: error: cannot convert ‘uint8x16_t {aka __vector(16) unsigned char}’ to ‘int8x16_t {aka __vector(16) signed char}’ in assignment
input_data_b = vld1q_u8(input_data_0 + 1 * input_depth);
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:6272:65: error: cannot convert ‘const uint8x16_t {aka const __vector(16) unsigned char}’ to ‘int8x16_t {aka __vector(16) signed char}’ for argument ‘2’ to ‘int8x16_t veorq_s8(int8x16_t, int8x16_t)’
work_reg_a_sp = veorq_s8(work_reg_a_sp, sign_bit);
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:6273:65: error: cannot convert ‘const uint8x16_t {aka const __vector(16) unsigned char}’ to ‘int8x16_t {aka __vector(16) signed char}’ for argument ‘2’ to ‘int8x16_t veorq_s8(int8x16_t, int8x16_t)’
work_reg_b_sp = veorq_s8(work_reg_b_sp, sign_bit);
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:6275:30: error: cannot convert ‘uint8x16_t {aka __vector(16) unsigned char}’ to ‘int8x16_t {aka __vector(16) signed char}’ in assignment
input_data_c = vld1q_u8(input_data_0 + 2 * input_depth);
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:6276:30: error: cannot convert ‘uint8x16_t {aka __vector(16) unsigned char}’ to ‘int8x16_t {aka __vector(16) signed char}’ in assignment
input_data_d = vld1q_u8(input_data_0 + 3 * input_depth);
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:6292:57: error: cannot convert ‘const uint8x16_t {aka const __vector(16) unsigned char}’ to ‘int8x16_t {aka __vector(16) signed char}’ for argument ‘2’ to ‘int8x16_t veorq_s8(int8x16_t, int8x16_t)’
work_reg_a = veorq_s8(work_reg_a, sign_bit);
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:6293:57: error: cannot convert ‘const uint8x16_t {aka const __vector(16) unsigned char}’ to ‘int8x16_t {aka __vector(16) signed char}’ for argument ‘2’ to ‘int8x16_t veorq_s8(int8x16_t, int8x16_t)’
work_reg_b = veorq_s8(work_reg_b, sign_bit);
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:6305:63: error: cannot convert ‘const uint8x16_t {aka const __vector(16) unsigned char}’ to ‘int8x16_t {aka __vector(16) signed char}’ for argument ‘2’ to ‘int8x16_t veorq_s8(int8x16_t, int8x16_t)’
work_reg_a_sp = veorq_s8(work_reg_a_sp, sign_bit);
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:6306:63: error: cannot convert ‘const uint8x16_t {aka const __vector(16) unsigned char}’ to ‘int8x16_t {aka __vector(16) signed char}’ for argument ‘2’ to ‘int8x16_t veorq_s8(int8x16_t, int8x16_t)’
work_reg_b_sp = veorq_s8(work_reg_b_sp, sign_bit);
^
In file included from ./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8.h:23:0,
from ./tensorflow/lite/kernels/internal/optimized/depthwiseconv_multithread.h:21,
from tensorflow/lite/kernels/depthwise_conv.cc:28:
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:46:71: error: cannot convert ‘int8x16_t {aka __vector(16) signed char}’ to ‘uint64x2_t {aka __vector(2) long unsigned int}’ for argument ‘2’ to ‘uint64x2_t vld1q_lane_u64(const uint64_t*, uint64x2_t, int)’
vld1q_lane_u64(reinterpret_cast<const uint64_t*>(src), reg, lane_num)
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:6316:30: note: in expansion of macro ‘vld1q_lane_s8x8’
input_data_a = vld1q_lane_s8x8(input_data_0, input_data_a, 0);
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:46:71: error: cannot convert ‘int8x16_t {aka __vector(16) signed char}’ to ‘uint64x2_t {aka __vector(2) long unsigned int}’ for argument ‘2’ to ‘uint64x2_t vld1q_lane_u64(const uint64_t*, uint64x2_t, int)’
vld1q_lane_u64(reinterpret_cast<const uint64_t*>(src), reg, lane_num)
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:6317:30: note: in expansion of macro ‘vld1q_lane_s8x8’
input_data_b = vld1q_lane_s8x8(input_data_0 + 1 * input_depth,
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:46:71: error: cannot convert ‘int8x16_t {aka __vector(16) signed char}’ to ‘uint64x2_t {aka __vector(2) long unsigned int}’ for argument ‘2’ to ‘uint64x2_t vld1q_lane_u64(const uint64_t*, uint64x2_t, int)’
vld1q_lane_u64(reinterpret_cast<const uint64_t*>(src), reg, lane_num)
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:6319:30: note: in expansion of macro ‘vld1q_lane_s8x8’
input_data_c = vld1q_lane_s8x8(input_data_0 + 2 * input_depth,
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:46:71: error: cannot convert ‘int8x16_t {aka __vector(16) signed char}’ to ‘uint64x2_t {aka __vector(2) long unsigned int}’ for argument ‘2’ to ‘uint64x2_t vld1q_lane_u64(const uint64_t*, uint64x2_t, int)’
vld1q_lane_u64(reinterpret_cast<const uint64_t*>(src), reg, lane_num)
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:6321:30: note: in expansion of macro ‘vld1q_lane_s8x8’
input_data_d = vld1q_lane_s8x8(input_data_0 + 3 * input_depth,
^
In file included from ./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8.h:23:0,
from ./tensorflow/lite/kernels/internal/optimized/depthwiseconv_multithread.h:21,
from tensorflow/lite/kernels/depthwise_conv.cc:28:
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:6329:57: error: cannot convert ‘const uint8x16_t {aka const __vector(16) unsigned char}’ to ‘int8x16_t {aka __vector(16) signed char}’ for argument ‘2’ to ‘int8x16_t veorq_s8(int8x16_t, int8x16_t)’
work_reg_a = veorq_s8(work_reg_a, sign_bit);
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:6330:57: error: cannot convert ‘const uint8x16_t {aka const __vector(16) unsigned char}’ to ‘int8x16_t {aka __vector(16) signed char}’ for argument ‘2’ to ‘int8x16_t veorq_s8(int8x16_t, int8x16_t)’
work_reg_b = veorq_s8(work_reg_b, sign_bit);
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:6344:28: error: cannot convert ‘uint8x16_t {aka __vector(16) unsigned char}’ to ‘int8x16_t {aka __vector(16) signed char}’ in assignment
input_data_a = vdupq_n_u8(-input_offset);
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:6345:28: error: cannot convert ‘uint8x16_t {aka __vector(16) unsigned char}’ to ‘int8x16_t {aka __vector(16) signed char}’ in assignment
input_data_b = vdupq_n_u8(-input_offset);
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:6346:28: error: cannot convert ‘uint8x16_t {aka __vector(16) unsigned char}’ to ‘int8x16_t {aka __vector(16) signed char}’ in assignment
input_data_c = vdupq_n_u8(-input_offset);
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:6347:28: error: cannot convert ‘uint8x16_t {aka __vector(16) unsigned char}’ to ‘int8x16_t {aka __vector(16) signed char}’ in assignment
input_data_d = vdupq_n_u8(-input_offset);
^
In file included from ./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8.h:23:0,
from ./tensorflow/lite/kernels/internal/optimized/depthwiseconv_multithread.h:21,
from tensorflow/lite/kernels/depthwise_conv.cc:28:
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:46:71: error: cannot convert ‘int8x16_t {aka __vector(16) signed char}’ to ‘uint64x2_t {aka __vector(2) long unsigned int}’ for argument ‘2’ to ‘uint64x2_t vld1q_lane_u64(const uint64_t*, uint64x2_t, int)’
vld1q_lane_u64(reinterpret_cast<const uint64_t*>(src), reg, lane_num)
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:6349:32: note: in expansion of macro ‘vld1q_lane_s8x8’
input_data_a = vld1q_lane_s8x8(input_data_0, input_data_a, 0);
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:46:71: error: cannot convert ‘int8x16_t {aka __vector(16) signed char}’ to ‘uint64x2_t {aka __vector(2) long unsigned int}’ for argument ‘2’ to ‘uint64x2_t vld1q_lane_u64(const uint64_t*, uint64x2_t, int)’
vld1q_lane_u64(reinterpret_cast<const uint64_t*>(src), reg, lane_num)
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:6351:34: note: in expansion of macro ‘vld1q_lane_s8x8’
input_data_b = vld1q_lane_s8x8(input_data_0 + input_depth,
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:46:71: error: cannot convert ‘int8x16_t {aka __vector(16) signed char}’ to ‘uint64x2_t {aka __vector(2) long unsigned int}’ for argument ‘2’ to ‘uint64x2_t vld1q_lane_u64(const uint64_t*, uint64x2_t, int)’
vld1q_lane_u64(reinterpret_cast<const uint64_t*>(src), reg, lane_num)
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:6354:36: note: in expansion of macro ‘vld1q_lane_s8x8’
input_data_c = vld1q_lane_s8x8(
^
In file included from ./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8.h:23:0,
from ./tensorflow/lite/kernels/internal/optimized/depthwiseconv_multithread.h:21,
from tensorflow/lite/kernels/depthwise_conv.cc:28:
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:6362:57: error: cannot convert ‘const uint8x16_t {aka const __vector(16) unsigned char}’ to ‘int8x16_t {aka __vector(16) signed char}’ for argument ‘2’ to ‘int8x16_t veorq_s8(int8x16_t, int8x16_t)’
work_reg_a = veorq_s8(work_reg_a, sign_bit);
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:6363:57: error: cannot convert ‘const uint8x16_t {aka const __vector(16) unsigned char}’ to ‘int8x16_t {aka __vector(16) signed char}’ for argument ‘2’ to ‘int8x16_t veorq_s8(int8x16_t, int8x16_t)’
work_reg_b = veorq_s8(work_reg_b, sign_bit);
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:6389:28: error: cannot convert ‘uint8x16_t {aka __vector(16) unsigned char}’ to ‘int8x16_t {aka __vector(16) signed char}’ in assignment
input_data_a = vdupq_n_u8(-input_offset);
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:6390:28: error: cannot convert ‘uint8x16_t {aka __vector(16) unsigned char}’ to ‘int8x16_t {aka __vector(16) signed char}’ in assignment
input_data_b = vld1q_u8(input_data_0 + 1 * input_depth);
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:6391:28: error: cannot convert ‘uint8x16_t {aka __vector(16) unsigned char}’ to ‘int8x16_t {aka __vector(16) signed char}’ in assignment
input_data_c = vld1q_u8(input_data_0 + 2 * input_depth);
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:6392:28: error: cannot convert ‘uint8x16_t {aka __vector(16) unsigned char}’ to ‘int8x16_t {aka __vector(16) signed char}’ in assignment
input_data_d = vld1q_u8(input_data_0 + 3 * input_depth);
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:6401:59: error: cannot convert ‘const uint8x16_t {aka const __vector(16) unsigned char}’ to ‘int8x16_t {aka __vector(16) signed char}’ for argument ‘2’ to ‘int8x16_t veorq_s8(int8x16_t, int8x16_t)’
work_reg_a = veorq_s8(work_reg_a, sign_bit);
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:6402:59: error: cannot convert ‘const uint8x16_t {aka const __vector(16) unsigned char}’ to ‘int8x16_t {aka __vector(16) signed char}’ for argument ‘2’ to ‘int8x16_t veorq_s8(int8x16_t, int8x16_t)’
work_reg_b = veorq_s8(work_reg_b, sign_bit);
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:6408:30: error: cannot convert ‘uint8x16_t {aka __vector(16) unsigned char}’ to ‘int8x16_t {aka __vector(16) signed char}’ in assignment
input_data_a = vdupq_n_u8(-input_offset);
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:6409:30: error: cannot convert ‘uint8x16_t {aka __vector(16) unsigned char}’ to ‘int8x16_t {aka __vector(16) signed char}’ in assignment
input_data_b = vld1q_u8(input_data_0 + 1 * input_depth);
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:6417:65: error: cannot convert ‘const uint8x16_t {aka const __vector(16) unsigned char}’ to ‘int8x16_t {aka __vector(16) signed char}’ for argument ‘2’ to ‘int8x16_t veorq_s8(int8x16_t, int8x16_t)’
work_reg_a_sp = veorq_s8(work_reg_a_sp, sign_bit);
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:6418:65: error: cannot convert ‘const uint8x16_t {aka const __vector(16) unsigned char}’ to ‘int8x16_t {aka __vector(16) signed char}’ for argument ‘2’ to ‘int8x16_t veorq_s8(int8x16_t, int8x16_t)’
work_reg_b_sp = veorq_s8(work_reg_b_sp, sign_bit);
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:6420:30: error: cannot convert ‘uint8x16_t {aka __vector(16) unsigned char}’ to ‘int8x16_t {aka __vector(16) signed char}’ in assignment
input_data_c = vld1q_u8(input_data_0 + 2 * input_depth);
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:6421:30: error: cannot convert ‘uint8x16_t {aka __vector(16) unsigned char}’ to ‘int8x16_t {aka __vector(16) signed char}’ in assignment
input_data_d = vld1q_u8(input_data_0 + 3 * input_depth);
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:6437:57: error: cannot convert ‘const uint8x16_t {aka const __vector(16) unsigned char}’ to ‘int8x16_t {aka __vector(16) signed char}’ for argument ‘2’ to ‘int8x16_t veorq_s8(int8x16_t, int8x16_t)’
work_reg_a = veorq_s8(work_reg_a, sign_bit);
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:6438:57: error: cannot convert ‘const uint8x16_t {aka const __vector(16) unsigned char}’ to ‘int8x16_t {aka __vector(16) signed char}’ for argument ‘2’ to ‘int8x16_t veorq_s8(int8x16_t, int8x16_t)’
work_reg_b = veorq_s8(work_reg_b, sign_bit);
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:6450:63: error: cannot convert ‘const uint8x16_t {aka const __vector(16) unsigned char}’ to ‘int8x16_t {aka __vector(16) signed char}’ for argument ‘2’ to ‘int8x16_t veorq_s8(int8x16_t, int8x16_t)’
work_reg_a_sp = veorq_s8(work_reg_a_sp, sign_bit);
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:6451:63: error: cannot convert ‘const uint8x16_t {aka const __vector(16) unsigned char}’ to ‘int8x16_t {aka __vector(16) signed char}’ for argument ‘2’ to ‘int8x16_t veorq_s8(int8x16_t, int8x16_t)’
work_reg_b_sp = veorq_s8(work_reg_b_sp, sign_bit);
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:6461:28: error: cannot convert ‘uint8x16_t {aka __vector(16) unsigned char}’ to ‘int8x16_t {aka __vector(16) signed char}’ in assignment
input_data_a = vdupq_n_u8(-input_offset);
^
In file included from ./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8.h:23:0,
from ./tensorflow/lite/kernels/internal/optimized/depthwiseconv_multithread.h:21,
from tensorflow/lite/kernels/depthwise_conv.cc:28:
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:46:71: error: cannot convert ‘int8x16_t {aka __vector(16) signed char}’ to ‘uint64x2_t {aka __vector(2) long unsigned int}’ for argument ‘2’ to ‘uint64x2_t vld1q_lane_u64(const uint64_t*, uint64x2_t, int)’
vld1q_lane_u64(reinterpret_cast<const uint64_t*>(src), reg, lane_num)
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:6462:30: note: in expansion of macro ‘vld1q_lane_s8x8’
input_data_b = vld1q_lane_s8x8(input_data_0 + 1 * input_depth,
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:46:71: error: cannot convert ‘int8x16_t {aka __vector(16) signed char}’ to ‘uint64x2_t {aka __vector(2) long unsigned int}’ for argument ‘2’ to ‘uint64x2_t vld1q_lane_u64(const uint64_t*, uint64x2_t, int)’
vld1q_lane_u64(reinterpret_cast<const uint64_t*>(src), reg, lane_num)
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:6464:30: note: in expansion of macro ‘vld1q_lane_s8x8’
input_data_c = vld1q_lane_s8x8(input_data_0 + 2 * input_depth,
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:46:71: error: cannot convert ‘int8x16_t {aka __vector(16) signed char}’ to ‘uint64x2_t {aka __vector(2) long unsigned int}’ for argument ‘2’ to ‘uint64x2_t vld1q_lane_u64(const uint64_t*, uint64x2_t, int)’
vld1q_lane_u64(reinterpret_cast<const uint64_t*>(src), reg, lane_num)
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:6466:30: note: in expansion of macro ‘vld1q_lane_s8x8’
input_data_d = vld1q_lane_s8x8(input_data_0 + 3 * input_depth,
^
In file included from ./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8.h:23:0,
from ./tensorflow/lite/kernels/internal/optimized/depthwiseconv_multithread.h:21,
from tensorflow/lite/kernels/depthwise_conv.cc:28:
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:6474:57: error: cannot convert ‘const uint8x16_t {aka const __vector(16) unsigned char}’ to ‘int8x16_t {aka __vector(16) signed char}’ for argument ‘2’ to ‘int8x16_t veorq_s8(int8x16_t, int8x16_t)’
work_reg_a = veorq_s8(work_reg_a, sign_bit);
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:6475:57: error: cannot convert ‘const uint8x16_t {aka const __vector(16) unsigned char}’ to ‘int8x16_t {aka __vector(16) signed char}’ for argument ‘2’ to ‘int8x16_t veorq_s8(int8x16_t, int8x16_t)’
work_reg_b = veorq_s8(work_reg_b, sign_bit);
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:6490:28: error: cannot convert ‘uint8x16_t {aka __vector(16) unsigned char}’ to ‘int8x16_t {aka __vector(16) signed char}’ in assignment
input_data_a = vdupq_n_u8(-input_offset);
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:6491:28: error: cannot convert ‘uint8x16_t {aka __vector(16) unsigned char}’ to ‘int8x16_t {aka __vector(16) signed char}’ in assignment
input_data_b = vdupq_n_u8(-input_offset);
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:6492:28: error: cannot convert ‘uint8x16_t {aka __vector(16) unsigned char}’ to ‘int8x16_t {aka __vector(16) signed char}’ in assignment
input_data_c = vdupq_n_u8(-input_offset);
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:6493:28: error: cannot convert ‘uint8x16_t {aka __vector(16) unsigned char}’ to ‘int8x16_t {aka __vector(16) signed char}’ in assignment
input_data_d = vdupq_n_u8(-input_offset);
^
In file included from ./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8.h:23:0,
from ./tensorflow/lite/kernels/internal/optimized/depthwiseconv_multithread.h:21,
from tensorflow/lite/kernels/depthwise_conv.cc:28:
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:46:71: error: cannot convert ‘int8x16_t {aka __vector(16) signed char}’ to ‘uint64x2_t {aka __vector(2) long unsigned int}’ for argument ‘2’ to ‘uint64x2_t vld1q_lane_u64(const uint64_t*, uint64x2_t, int)’
vld1q_lane_u64(reinterpret_cast<const uint64_t*>(src), reg, lane_num)
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:6496:32: note: in expansion of macro ‘vld1q_lane_s8x8’
input_data_b = vld1q_lane_s8x8(input_data_0 + input_depth,
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:46:71: error: cannot convert ‘int8x16_t {aka __vector(16) signed char}’ to ‘uint64x2_t {aka __vector(2) long unsigned int}’ for argument ‘2’ to ‘uint64x2_t vld1q_lane_u64(const uint64_t*, uint64x2_t, int)’
vld1q_lane_u64(reinterpret_cast<const uint64_t*>(src), reg, lane_num)
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:6499:34: note: in expansion of macro ‘vld1q_lane_s8x8’
input_data_c = vld1q_lane_s8x8(input_data_0 + 2 * input_depth,
^
In file included from ./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8.h:23:0,
from ./tensorflow/lite/kernels/internal/optimized/depthwiseconv_multithread.h:21,
from tensorflow/lite/kernels/depthwise_conv.cc:28:
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:6506:57: error: cannot convert ‘const uint8x16_t {aka const __vector(16) unsigned char}’ to ‘int8x16_t {aka __vector(16) signed char}’ for argument ‘2’ to ‘int8x16_t veorq_s8(int8x16_t, int8x16_t)’
work_reg_a = veorq_s8(work_reg_a, sign_bit);
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:6507:57: error: cannot convert ‘const uint8x16_t {aka const __vector(16) unsigned char}’ to ‘int8x16_t {aka __vector(16) signed char}’ for argument ‘2’ to ‘int8x16_t veorq_s8(int8x16_t, int8x16_t)’
work_reg_b = veorq_s8(work_reg_b, sign_bit);
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h: In static member function ‘static void tflite::optimized_ops::depthwise_conv::PackMacroBlock<(tflite::DepthwiseConvImplementation)3, (tflite::DepthwiseConvDepthMultiplication)1, 1>::PackMacroBlockNeon(int32, int32, const uint8*, int8*, const tflite::optimized_ops::depthwise_conv::DepthwiseConvDotProdParams*)’:
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:6647:75: error: cannot convert ‘int8x8_t {aka __vector(8) signed char}’ to ‘uint64x1_t {aka __vector(1) long unsigned int}’ for argument ‘1’ to ‘uint64x1_t vshl_u64(uint64x1_t, int64x1_t)’
padding_mask = vshl_u64(padding_mask, vdup_n_s64(8 * copy_remaining));
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:6659:20: error: cannot convert ‘uint8x16_t {aka __vector(16) unsigned char}’ to ‘int8x16_t {aka __vector(16) signed char}’ in assignment
work_reg = vld1q_u8(input_block_data + input_block_offset);
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:6660:56: error: cannot convert ‘const uint8x16_t {aka const __vector(16) unsigned char}’ to ‘int8x16_t {aka __vector(16) signed char}’ for argument ‘1’ to ‘int8x16_t vextq_s8(int8x16_t, int8x16_t, int)’
work_reg = vextq_s8(padding_reg, work_reg, 15);
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:6661:49: error: cannot convert ‘const uint8x16_t {aka const __vector(16) unsigned char}’ to ‘int8x16_t {aka __vector(16) signed char}’ for argument ‘2’ to ‘int8x16_t veorq_s8(int8x16_t, int8x16_t)’
work_reg = veorq_s8(work_reg, sign_bit);
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:6669:20: error: cannot convert ‘uint8x16_t {aka __vector(16) unsigned char}’ to ‘int8x16_t {aka __vector(16) signed char}’ in assignment
work_reg =
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:6671:49: error: cannot convert ‘const uint8x16_t {aka const __vector(16) unsigned char}’ to ‘int8x16_t {aka __vector(16) signed char}’ for argument ‘2’ to ‘int8x16_t veorq_s8(int8x16_t, int8x16_t)’
work_reg = veorq_s8(work_reg, sign_bit);
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:6679:25: error: cannot convert ‘uint8x8_t {aka __vector(8) unsigned char}’ to ‘int8x8_t {aka __vector(8) signed char}’ in assignment
half_work_reg =
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:6681:70: error: cannot convert ‘const uint8x16_t {aka const __vector(16) unsigned char}’ to ‘int8x16_t {aka __vector(16) signed char}’ for argument ‘1’ to ‘int8x8_t vget_low_s8(int8x16_t)’
half_work_reg = veor_s8(half_work_reg, vget_low_s8(sign_bit));
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:6701:25: error: cannot convert ‘uint8x8_t {aka __vector(8) unsigned char}’ to ‘int8x8_t {aka __vector(8) signed char}’ in assignment
half_work_reg =
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:6705:76: error: cannot convert ‘int8x8_t {aka __vector(8) signed char}’ to ‘uint64x1_t {aka __vector(1) long unsigned int}’ for argument ‘1’ to ‘uint64x1_t vshl_u64(uint64x1_t, int64x1_t)’
vshl_u64(half_work_reg, vdup_n_s64(-8 * (8 - copy_remaining)));
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:6707:60: error: cannot convert ‘const uint8x16_t {aka const __vector(16) unsigned char}’ to ‘int8x16_t {aka __vector(16) signed char}’ for argument ‘1’ to ‘int8x8_t vget_low_s8(int8x16_t)’
vbsl_s8(padding_mask, vget_low_s8(padding_reg), half_work_reg);
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:6709:70: error: cannot convert ‘const uint8x16_t {aka const __vector(16) unsigned char}’ to ‘int8x16_t {aka __vector(16) signed char}’ for argument ‘1’ to ‘int8x8_t vget_low_s8(int8x16_t)’
half_work_reg = veor_s8(half_work_reg, vget_low_s8(sign_bit));
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:6729:75: error: cannot convert ‘int8x8_t {aka __vector(8) signed char}’ to ‘uint64x1_t {aka __vector(1) long unsigned int}’ for argument ‘1’ to ‘uint64x1_t vshl_u64(uint64x1_t, int64x1_t)’
padding_mask = vshl_u64(padding_mask, vdup_n_s64(8 * copy_remaining));
^
In file included from ./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8.h:23:0,
from ./tensorflow/lite/kernels/internal/optimized/depthwiseconv_multithread.h:21,
from tensorflow/lite/kernels/depthwise_conv.cc:28:
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:48:67: error: cannot convert ‘int8x8_t {aka __vector(8) signed char}’ to ‘int32x2_t {aka __vector(2) int}’ for argument ‘2’ to ‘int32x2_t vld1_lane_s32(const int32_t*, int32x2_t, int)’
vld1_lane_s32(reinterpret_cast<const int32*>(src), reg, lane_num)
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:6741:27: note: in expansion of macro ‘vld1_lane_8x4’
half_work_reg = vld1_lane_8x4(input_block_data + input_block_offset,
^
In file included from ./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8.h:23:0,
from ./tensorflow/lite/kernels/internal/optimized/depthwiseconv_multithread.h:21,
from tensorflow/lite/kernels/depthwise_conv.cc:28:
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:6743:58: error: cannot convert ‘const uint8x16_t {aka const __vector(16) unsigned char}’ to ‘int8x16_t {aka __vector(16) signed char}’ for argument ‘1’ to ‘int8x8_t vget_low_s8(int8x16_t)’
half_work_reg = vext_s8(vget_low_s8(padding_reg), half_work_reg, 7);
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:6744:70: error: cannot convert ‘const uint8x16_t {aka const __vector(16) unsigned char}’ to ‘int8x16_t {aka __vector(16) signed char}’ for argument ‘1’ to ‘int8x8_t vget_low_s8(int8x16_t)’
half_work_reg = veor_s8(half_work_reg, vget_low_s8(sign_bit));
^
In file included from ./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8.h:23:0,
from ./tensorflow/lite/kernels/internal/optimized/depthwiseconv_multithread.h:21,
from tensorflow/lite/kernels/depthwise_conv.cc:28:
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:37:64: error: cannot convert ‘int8x8_t {aka __vector(8) signed char}’ to ‘uint32x2_t {aka __vector(2) unsigned int}’ for argument ‘2’ to ‘void vst1_lane_u32(uint32_t*, uint32x2_t, int)’
vst1_lane_u32(reinterpret_cast<uint32_t*>(dst), reg, lane_num)
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:6746:11: note: in expansion of macro ‘vst1_lane_8x4’
vst1_lane_8x4(scratch_data, half_work_reg, 0);
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:48:67: error: cannot convert ‘int8x8_t {aka __vector(8) signed char}’ to ‘int32x2_t {aka __vector(2) int}’ for argument ‘2’ to ‘int32x2_t vld1_lane_s32(const int32_t*, int32x2_t, int)’
vld1_lane_s32(reinterpret_cast<const int32*>(src), reg, lane_num)
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:6753:15: note: in expansion of macro ‘vld1_lane_8x4’
vld1_lane_8x4(input_block_data + input_block_offset + copy_done,
^
In file included from ./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8.h:23:0,
from ./tensorflow/lite/kernels/internal/optimized/depthwiseconv_multithread.h:21,
from tensorflow/lite/kernels/depthwise_conv.cc:28:
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:6755:70: error: cannot convert ‘const uint8x16_t {aka const __vector(16) unsigned char}’ to ‘int8x16_t {aka __vector(16) signed char}’ for argument ‘1’ to ‘int8x8_t vget_low_s8(int8x16_t)’
half_work_reg = veor_s8(half_work_reg, vget_low_s8(sign_bit));
^
In file included from ./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8.h:23:0,
from ./tensorflow/lite/kernels/internal/optimized/depthwiseconv_multithread.h:21,
from tensorflow/lite/kernels/depthwise_conv.cc:28:
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:37:64: error: cannot convert ‘int8x8_t {aka __vector(8) signed char}’ to ‘uint32x2_t {aka __vector(2) unsigned int}’ for argument ‘2’ to ‘void vst1_lane_u32(uint32_t*, uint32x2_t, int)’
vst1_lane_u32(reinterpret_cast<uint32_t*>(dst), reg, lane_num)
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:6759:11: note: in expansion of macro ‘vst1_lane_8x4’
vst1_lane_8x4(scratch_data + start_width + copy_done, half_work_reg,
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:48:67: error: cannot convert ‘int8x8_t {aka __vector(8) signed char}’ to ‘int32x2_t {aka __vector(2) int}’ for argument ‘2’ to ‘int32x2_t vld1_lane_s32(const int32_t*, int32x2_t, int)’
vld1_lane_s32(reinterpret_cast<const int32*>(src), reg, lane_num)
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:6776:27: note: in expansion of macro ‘vld1_lane_8x4’
half_work_reg = vld1_lane_8x4(
^
In file included from ./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8.h:23:0,
from ./tensorflow/lite/kernels/internal/optimized/depthwiseconv_multithread.h:21,
from tensorflow/lite/kernels/depthwise_conv.cc:28:
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:6781:76: error: cannot convert ‘int8x8_t {aka __vector(8) signed char}’ to ‘uint64x1_t {aka __vector(1) long unsigned int}’ for argument ‘1’ to ‘uint64x1_t vshl_u64(uint64x1_t, int64x1_t)’
vshl_u64(half_work_reg, vdup_n_s64(-8 * (4 - copy_remaining)));
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:6783:60: error: cannot convert ‘const uint8x16_t {aka const __vector(16) unsigned char}’ to ‘int8x16_t {aka __vector(16) signed char}’ for argument ‘1’ to ‘int8x8_t vget_low_s8(int8x16_t)’
vbsl_s8(padding_mask, vget_low_s8(padding_reg), half_work_reg);
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:6785:70: error: cannot convert ‘const uint8x16_t {aka const __vector(16) unsigned char}’ to ‘int8x16_t {aka __vector(16) signed char}’ for argument ‘1’ to ‘int8x8_t vget_low_s8(int8x16_t)’
half_work_reg = veor_s8(half_work_reg, vget_low_s8(sign_bit));
^
In file included from ./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8.h:23:0,
from ./tensorflow/lite/kernels/internal/optimized/depthwiseconv_multithread.h:21,
from tensorflow/lite/kernels/depthwise_conv.cc:28:
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:37:64: error: cannot convert ‘int8x8_t {aka __vector(8) signed char}’ to ‘uint32x2_t {aka __vector(2) unsigned int}’ for argument ‘2’ to ‘void vst1_lane_u32(uint32_t*, uint32x2_t, int)’
vst1_lane_u32(reinterpret_cast<uint32_t*>(dst), reg, lane_num)
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:6789:11: note: in expansion of macro ‘vst1_lane_8x4’
vst1_lane_8x4(scratch_data + start_width + copy_done, half_work_reg,
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:37:64: error: cannot convert ‘int8x8_t {aka __vector(8) signed char}’ to ‘uint32x2_t {aka __vector(2) unsigned int}’ for argument ‘2’ to ‘void vst1_lane_u32(uint32_t*, uint32x2_t, int)’
vst1_lane_u32(reinterpret_cast<uint32_t*>(dst), reg, lane_num)
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:6798:9: note: in expansion of macro ‘vst1_lane_8x4’
vst1_lane_8x4(scratch_data + start_width + copy_done, half_work_reg, 0);
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:37:64: error: cannot convert ‘int8x8_t {aka __vector(8) signed char}’ to ‘uint32x2_t {aka __vector(2) unsigned int}’ for argument ‘2’ to ‘void vst1_lane_u32(uint32_t*, uint32x2_t, int)’
vst1_lane_u32(reinterpret_cast<uint32_t*>(dst), reg, lane_num)
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:6799:9: note: in expansion of macro ‘vst1_lane_8x4’
vst1_lane_8x4(scratch_data + start_width + copy_done + 4, half_work_reg,
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:37:64: error: cannot convert ‘int8x8_t {aka __vector(8) signed char}’ to ‘uint32x2_t {aka __vector(2) unsigned int}’ for argument ‘2’ to ‘void vst1_lane_u32(uint32_t*, uint32x2_t, int)’
vst1_lane_u32(reinterpret_cast<uint32_t*>(dst), reg, lane_num)
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:6801:9: note: in expansion of macro ‘vst1_lane_8x4’
vst1_lane_8x4(scratch_data + start_width + copy_done + 8, half_work_reg,
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:37:64: error: cannot convert ‘int8x8_t {aka __vector(8) signed char}’ to ‘uint32x2_t {aka __vector(2) unsigned int}’ for argument ‘2’ to ‘void vst1_lane_u32(uint32_t*, uint32x2_t, int)’
vst1_lane_u32(reinterpret_cast<uint32_t*>(dst), reg, lane_num)
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:6803:9: note: in expansion of macro ‘vst1_lane_8x4’
vst1_lane_8x4(scratch_data + start_width + copy_done + 12,
^
In file included from ./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8.h:23:0,
from ./tensorflow/lite/kernels/internal/optimized/depthwiseconv_multithread.h:21,
from tensorflow/lite/kernels/depthwise_conv.cc:28:
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:6818:23: error: cannot convert ‘uint8x8_t {aka __vector(8) unsigned char}’ to ‘int8x8_t {aka __vector(8) signed char}’ in assignment
half_work_reg = vdup_n_u8(-input_offset);
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:6831:68: error: cannot convert ‘const uint8x16_t {aka const __vector(16) unsigned char}’ to ‘int8x16_t {aka __vector(16) signed char}’ for argument ‘1’ to ‘int8x8_t vget_low_s8(int8x16_t)’
half_work_reg = veor_s8(half_work_reg, vget_low_s8(sign_bit));
^
In file included from ./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8.h:23:0,
from ./tensorflow/lite/kernels/internal/optimized/depthwiseconv_multithread.h:21,
from tensorflow/lite/kernels/depthwise_conv.cc:28:
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:37:64: error: cannot convert ‘int8x8_t {aka __vector(8) signed char}’ to ‘uint32x2_t {aka __vector(2) unsigned int}’ for argument ‘2’ to ‘void vst1_lane_u32(uint32_t*, uint32x2_t, int)’
vst1_lane_u32(reinterpret_cast<uint32_t*>(dst), reg, lane_num)
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:6842:9: note: in expansion of macro ‘vst1_lane_8x4’
vst1_lane_8x4(scratch_data_base + scratch_data_offset + 4,
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:37:64: error: cannot convert ‘int8x8_t {aka __vector(8) signed char}’ to ‘uint32x2_t {aka __vector(2) unsigned int}’ for argument ‘2’ to ‘void vst1_lane_u32(uint32_t*, uint32x2_t, int)’
vst1_lane_u32(reinterpret_cast<uint32_t*>(dst), reg, lane_num)
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:6844:9: note: in expansion of macro ‘vst1_lane_8x4’
vst1_lane_8x4(scratch_data_base + scratch_data_offset + 8,
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:37:64: error: cannot convert ‘int8x8_t {aka __vector(8) signed char}’ to ‘uint32x2_t {aka __vector(2) unsigned int}’ for argument ‘2’ to ‘void vst1_lane_u32(uint32_t*, uint32x2_t, int)’
vst1_lane_u32(reinterpret_cast<uint32_t*>(dst), reg, lane_num)
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:6846:9: note: in expansion of macro ‘vst1_lane_8x4’
vst1_lane_8x4(scratch_data_base + scratch_data_offset + 12,
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:37:64: error: cannot convert ‘int8x8_t {aka __vector(8) signed char}’ to ‘uint32x2_t {aka __vector(2) unsigned int}’ for argument ‘2’ to ‘void vst1_lane_u32(uint32_t*, uint32x2_t, int)’
vst1_lane_u32(reinterpret_cast<uint32_t*>(dst), reg, lane_num)
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:6848:9: note: in expansion of macro ‘vst1_lane_8x4’
vst1_lane_8x4(scratch_data_base + scratch_data_offset + 16,
^
In file included from ./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8.h:23:0,
from ./tensorflow/lite/kernels/internal/optimized/depthwiseconv_multithread.h:21,
from tensorflow/lite/kernels/depthwise_conv.cc:28:
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:6857:75: error: cannot convert ‘int8x8_t {aka __vector(8) signed char}’ to ‘uint64x1_t {aka __vector(1) long unsigned int}’ for argument ‘1’ to ‘uint64x1_t vshl_u64(uint64x1_t, int64x1_t)’
padding_mask = vshl_u64(padding_mask, vdup_n_s64(8 * copy_remaining));
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:6859:57: error: cannot convert ‘int8x8_t {aka __vector(8) signed char}’ to ‘uint8x8_t {aka __vector(8) unsigned char}’ for argument ‘2’ to ‘uint8x8_t vset_lane_u8(uint8_t, uint8x8_t, int)’
padding_mask = vset_lane_u8(255, padding_mask, 0);
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:6864:54: error: cannot convert ‘int8x8_t {aka __vector(8) signed char}’ to ‘uint64x1_t {aka __vector(1) long unsigned int}’ for argument ‘1’ to ‘uint64x1_t vshl_n_u64(uint64x1_t, int)’
half_work_reg = vshl_n_u64(half_work_reg, 8);
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:6871:54: error: cannot convert ‘int8x8_t {aka __vector(8) signed char}’ to ‘int64x1_t {aka __vector(1) long int}’ for argument ‘1’ to ‘int64x1_t vshl_n_s64(int64x1_t, int)’
half_work_reg = vshl_n_s64(half_work_reg, 8);
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:6874:58: error: cannot convert ‘const uint8x16_t {aka const __vector(16) unsigned char}’ to ‘int8x16_t {aka __vector(16) signed char}’ for argument ‘1’ to ‘int8x8_t vget_low_s8(int8x16_t)’
vbsl_s8(padding_mask, vget_low_s8(padding_reg), half_work_reg);
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:6876:68: error: cannot convert ‘const uint8x16_t {aka const __vector(16) unsigned char}’ to ‘int8x16_t {aka __vector(16) signed char}’ for argument ‘1’ to ‘int8x8_t vget_low_s8(int8x16_t)’
half_work_reg = veor_s8(half_work_reg, vget_low_s8(sign_bit));
^
In file included from ./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8.h:23:0,
from ./tensorflow/lite/kernels/internal/optimized/depthwiseconv_multithread.h:21,
from tensorflow/lite/kernels/depthwise_conv.cc:28:
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:37:64: error: cannot convert ‘int8x8_t {aka __vector(8) signed char}’ to ‘uint32x2_t {aka __vector(2) unsigned int}’ for argument ‘2’ to ‘void vst1_lane_u32(uint32_t*, uint32x2_t, int)’
vst1_lane_u32(reinterpret_cast<uint32_t*>(dst), reg, lane_num)
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:6880:9: note: in expansion of macro ‘vst1_lane_8x4’
vst1_lane_8x4(scratch_data_base + scratch_data_offset, half_work_reg,
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:37:64: error: cannot convert ‘int8x8_t {aka __vector(8) signed char}’ to ‘uint32x2_t {aka __vector(2) unsigned int}’ for argument ‘2’ to ‘void vst1_lane_u32(uint32_t*, uint32x2_t, int)’
vst1_lane_u32(reinterpret_cast<uint32_t*>(dst), reg, lane_num)
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:6888:9: note: in expansion of macro ‘vst1_lane_8x4’
vst1_lane_8x4(scratch_data_base + scratch_data_offset + 4,
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:37:64: error: cannot convert ‘int8x8_t {aka __vector(8) signed char}’ to ‘uint32x2_t {aka __vector(2) unsigned int}’ for argument ‘2’ to ‘void vst1_lane_u32(uint32_t*, uint32x2_t, int)’
vst1_lane_u32(reinterpret_cast<uint32_t*>(dst), reg, lane_num)
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:6890:9: note: in expansion of macro ‘vst1_lane_8x4’
vst1_lane_8x4(scratch_data_base + scratch_data_offset + 8,
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:37:64: error: cannot convert ‘int8x8_t {aka __vector(8) signed char}’ to ‘uint32x2_t {aka __vector(2) unsigned int}’ for argument ‘2’ to ‘void vst1_lane_u32(uint32_t*, uint32x2_t, int)’
vst1_lane_u32(reinterpret_cast<uint32_t*>(dst), reg, lane_num)
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:6892:9: note: in expansion of macro ‘vst1_lane_8x4’
vst1_lane_8x4(scratch_data_base + scratch_data_offset + 12,
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:37:64: error: cannot convert ‘int8x8_t {aka __vector(8) signed char}’ to ‘uint32x2_t {aka __vector(2) unsigned int}’ for argument ‘2’ to ‘void vst1_lane_u32(uint32_t*, uint32x2_t, int)’
vst1_lane_u32(reinterpret_cast<uint32_t*>(dst), reg, lane_num)
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:6894:9: note: in expansion of macro ‘vst1_lane_8x4’
vst1_lane_8x4(scratch_data_base + scratch_data_offset + 16,
^
In file included from ./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8.h:23:0,
from ./tensorflow/lite/kernels/internal/optimized/depthwiseconv_multithread.h:21,
from tensorflow/lite/kernels/depthwise_conv.cc:28:
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h: In static member function ‘static void tflite::optimized_ops::depthwise_conv::PackMacroBlock<(tflite::DepthwiseConvImplementation)3, (tflite::DepthwiseConvDepthMultiplication)1, 0>::PackMacroBlockNeon(int32, int32, const uint8*, int8*, const tflite::optimized_ops::depthwise_conv::DepthwiseConvDotProdParams*)’:
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:6993:20: error: cannot convert ‘uint8x16_t {aka __vector(16) unsigned char}’ to ‘int8x16_t {aka __vector(16) signed char}’ in assignment
work_reg =
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:6995:49: error: cannot convert ‘const uint8x16_t {aka const __vector(16) unsigned char}’ to ‘int8x16_t {aka __vector(16) signed char}’ for argument ‘2’ to ‘int8x16_t veorq_s8(int8x16_t, int8x16_t)’
work_reg = veorq_s8(work_reg, sign_bit);
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:7002:25: error: cannot convert ‘uint8x8_t {aka __vector(8) unsigned char}’ to ‘int8x8_t {aka __vector(8) signed char}’ in assignment
half_work_reg =
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:7004:70: error: cannot convert ‘const uint8x16_t {aka const __vector(16) unsigned char}’ to ‘int8x16_t {aka __vector(16) signed char}’ for argument ‘1’ to ‘int8x8_t vget_low_s8(int8x16_t)’
half_work_reg = veor_s8(half_work_reg, vget_low_s8(sign_bit));
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:7023:25: error: cannot convert ‘uint8x8_t {aka __vector(8) unsigned char}’ to ‘int8x8_t {aka __vector(8) signed char}’ in assignment
half_work_reg =
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:7027:76: error: cannot convert ‘int8x8_t {aka __vector(8) signed char}’ to ‘uint64x1_t {aka __vector(1) long unsigned int}’ for argument ‘1’ to ‘uint64x1_t vshl_u64(uint64x1_t, int64x1_t)’
vshl_u64(half_work_reg, vdup_n_s64(-8 * (8 - copy_remaining)));
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:7029:70: error: cannot convert ‘const uint8x16_t {aka const __vector(16) unsigned char}’ to ‘int8x16_t {aka __vector(16) signed char}’ for argument ‘1’ to ‘int8x8_t vget_low_s8(int8x16_t)’
half_work_reg = veor_s8(half_work_reg, vget_low_s8(sign_bit));
^
In file included from ./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8.h:23:0,
from ./tensorflow/lite/kernels/internal/optimized/depthwiseconv_multithread.h:21,
from tensorflow/lite/kernels/depthwise_conv.cc:28:
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:48:67: error: cannot convert ‘int8x8_t {aka __vector(8) signed char}’ to ‘int32x2_t {aka __vector(2) int}’ for argument ‘2’ to ‘int32x2_t vld1_lane_s32(const int32_t*, int32x2_t, int)’
vld1_lane_s32(reinterpret_cast<const int32*>(src), reg, lane_num)
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:7057:15: note: in expansion of macro ‘vld1_lane_8x4’
vld1_lane_8x4(input_block_data + input_block_offset + copy_done,
^
In file included from ./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8.h:23:0,
from ./tensorflow/lite/kernels/internal/optimized/depthwiseconv_multithread.h:21,
from tensorflow/lite/kernels/depthwise_conv.cc:28:
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:7059:70: error: cannot convert ‘const uint8x16_t {aka const __vector(16) unsigned char}’ to ‘int8x16_t {aka __vector(16) signed char}’ for argument ‘1’ to ‘int8x8_t vget_low_s8(int8x16_t)’
half_work_reg = veor_s8(half_work_reg, vget_low_s8(sign_bit));
^
In file included from ./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8.h:23:0,
from ./tensorflow/lite/kernels/internal/optimized/depthwiseconv_multithread.h:21,
from tensorflow/lite/kernels/depthwise_conv.cc:28:
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:37:64: error: cannot convert ‘int8x8_t {aka __vector(8) signed char}’ to ‘uint32x2_t {aka __vector(2) unsigned int}’ for argument ‘2’ to ‘void vst1_lane_u32(uint32_t*, uint32x2_t, int)’
vst1_lane_u32(reinterpret_cast<uint32_t*>(dst), reg, lane_num)
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:7062:11: note: in expansion of macro ‘vst1_lane_8x4’
vst1_lane_8x4(scratch_data + copy_done, half_work_reg, 0);
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:48:67: error: cannot convert ‘int8x8_t {aka __vector(8) signed char}’ to ‘int32x2_t {aka __vector(2) int}’ for argument ‘2’ to ‘int32x2_t vld1_lane_s32(const int32_t*, int32x2_t, int)’
vld1_lane_s32(reinterpret_cast<const int32*>(src), reg, lane_num)
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:7078:27: note: in expansion of macro ‘vld1_lane_8x4’
half_work_reg = vld1_lane_8x4(
^
In file included from ./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8.h:23:0,
from ./tensorflow/lite/kernels/internal/optimized/depthwiseconv_multithread.h:21,
from tensorflow/lite/kernels/depthwise_conv.cc:28:
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:7083:76: error: cannot convert ‘int8x8_t {aka __vector(8) signed char}’ to ‘uint64x1_t {aka __vector(1) long unsigned int}’ for argument ‘1’ to ‘uint64x1_t vshl_u64(uint64x1_t, int64x1_t)’
vshl_u64(half_work_reg, vdup_n_s64(-8 * (4 - copy_remaining)));
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:7085:70: error: cannot convert ‘const uint8x16_t {aka const __vector(16) unsigned char}’ to ‘int8x16_t {aka __vector(16) signed char}’ for argument ‘1’ to ‘int8x8_t vget_low_s8(int8x16_t)’
half_work_reg = veor_s8(half_work_reg, vget_low_s8(sign_bit));
^
In file included from ./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8.h:23:0,
from ./tensorflow/lite/kernels/internal/optimized/depthwiseconv_multithread.h:21,
from tensorflow/lite/kernels/depthwise_conv.cc:28:
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:37:64: error: cannot convert ‘int8x8_t {aka __vector(8) signed char}’ to ‘uint32x2_t {aka __vector(2) unsigned int}’ for argument ‘2’ to ‘void vst1_lane_u32(uint32_t*, uint32x2_t, int)’
vst1_lane_u32(reinterpret_cast<uint32_t*>(dst), reg, lane_num)
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:7088:11: note: in expansion of macro ‘vst1_lane_8x4’
vst1_lane_8x4(scratch_data + copy_done, half_work_reg, 0);
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:37:64: error: cannot convert ‘int8x8_t {aka __vector(8) signed char}’ to ‘uint32x2_t {aka __vector(2) unsigned int}’ for argument ‘2’ to ‘void vst1_lane_u32(uint32_t*, uint32x2_t, int)’
vst1_lane_u32(reinterpret_cast<uint32_t*>(dst), reg, lane_num)
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:7094:9: note: in expansion of macro ‘vst1_lane_8x4’
vst1_lane_8x4(scratch_data + copy_done, half_work_reg, 0);
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:37:64: error: cannot convert ‘int8x8_t {aka __vector(8) signed char}’ to ‘uint32x2_t {aka __vector(2) unsigned int}’ for argument ‘2’ to ‘void vst1_lane_u32(uint32_t*, uint32x2_t, int)’
vst1_lane_u32(reinterpret_cast<uint32_t*>(dst), reg, lane_num)
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:7095:9: note: in expansion of macro ‘vst1_lane_8x4’
vst1_lane_8x4(scratch_data + copy_done + 4, half_work_reg, 0);
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:37:64: error: cannot convert ‘int8x8_t {aka __vector(8) signed char}’ to ‘uint32x2_t {aka __vector(2) unsigned int}’ for argument ‘2’ to ‘void vst1_lane_u32(uint32_t*, uint32x2_t, int)’
vst1_lane_u32(reinterpret_cast<uint32_t*>(dst), reg, lane_num)
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:7096:9: note: in expansion of macro ‘vst1_lane_8x4’
vst1_lane_8x4(scratch_data + copy_done + 8, half_work_reg, 0);
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:37:64: error: cannot convert ‘int8x8_t {aka __vector(8) signed char}’ to ‘uint32x2_t {aka __vector(2) unsigned int}’ for argument ‘2’ to ‘void vst1_lane_u32(uint32_t*, uint32x2_t, int)’
vst1_lane_u32(reinterpret_cast<uint32_t*>(dst), reg, lane_num)
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:7097:9: note: in expansion of macro ‘vst1_lane_8x4’
vst1_lane_8x4(scratch_data + copy_done + 12, half_work_reg, 0);
^
In file included from ./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8.h:23:0,
from ./tensorflow/lite/kernels/internal/optimized/depthwiseconv_multithread.h:21,
from tensorflow/lite/kernels/depthwise_conv.cc:28:
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:7107:54: error: cannot convert ‘int8x8_t {aka __vector(8) signed char}’ to ‘uint64x1_t {aka __vector(1) long unsigned int}’ for argument ‘1’ to ‘uint64x1_t vshl_n_u64(uint64x1_t, int)’
half_work_reg = vshl_n_u64(half_work_reg, 8);
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:7114:68: error: cannot convert ‘const uint8x16_t {aka const __vector(16) unsigned char}’ to ‘int8x16_t {aka __vector(16) signed char}’ for argument ‘1’ to ‘int8x8_t vget_low_s8(int8x16_t)’
half_work_reg = veor_s8(half_work_reg, vget_low_s8(sign_bit));
^
In file included from ./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8.h:23:0,
from ./tensorflow/lite/kernels/internal/optimized/depthwiseconv_multithread.h:21,
from tensorflow/lite/kernels/depthwise_conv.cc:28:
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:37:64: error: cannot convert ‘int8x8_t {aka __vector(8) signed char}’ to ‘uint32x2_t {aka __vector(2) unsigned int}’ for argument ‘2’ to ‘void vst1_lane_u32(uint32_t*, uint32x2_t, int)’
vst1_lane_u32(reinterpret_cast<uint32_t*>(dst), reg, lane_num)
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:7118:9: note: in expansion of macro ‘vst1_lane_8x4’
vst1_lane_8x4(scratch_data_base + scratch_data_offset, half_work_reg,
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:37:64: error: cannot convert ‘int8x8_t {aka __vector(8) signed char}’ to ‘uint32x2_t {aka __vector(2) unsigned int}’ for argument ‘2’ to ‘void vst1_lane_u32(uint32_t*, uint32x2_t, int)’
vst1_lane_u32(reinterpret_cast<uint32_t*>(dst), reg, lane_num)
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:7124:9: note: in expansion of macro ‘vst1_lane_8x4’
vst1_lane_8x4(scratch_data_base + scratch_data_offset + 4,
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:37:64: error: cannot convert ‘int8x8_t {aka __vector(8) signed char}’ to ‘uint32x2_t {aka __vector(2) unsigned int}’ for argument ‘2’ to ‘void vst1_lane_u32(uint32_t*, uint32x2_t, int)’
vst1_lane_u32(reinterpret_cast<uint32_t*>(dst), reg, lane_num)
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:7126:9: note: in expansion of macro ‘vst1_lane_8x4’
vst1_lane_8x4(scratch_data_base + scratch_data_offset + 8,
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:37:64: error: cannot convert ‘int8x8_t {aka __vector(8) signed char}’ to ‘uint32x2_t {aka __vector(2) unsigned int}’ for argument ‘2’ to ‘void vst1_lane_u32(uint32_t*, uint32x2_t, int)’
vst1_lane_u32(reinterpret_cast<uint32_t*>(dst), reg, lane_num)
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:7128:9: note: in expansion of macro ‘vst1_lane_8x4’
vst1_lane_8x4(scratch_data_base + scratch_data_offset + 12,
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:37:64: error: cannot convert ‘int8x8_t {aka __vector(8) signed char}’ to ‘uint32x2_t {aka __vector(2) unsigned int}’ for argument ‘2’ to ‘void vst1_lane_u32(uint32_t*, uint32x2_t, int)’
vst1_lane_u32(reinterpret_cast<uint32_t*>(dst), reg, lane_num)
^
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8_3x3_filter.h:7130:9: note: in expansion of macro ‘vst1_lane_8x4’
vst1_lane_8x4(scratch_data_base + scratch_data_offset + 16,
^
Target //tensorflow/tools/pip_package:build_pip_package failed to build
INFO: Elapsed time: 3320.032s, Critical Path: 297.50s
INFO: 2544 processes: 2544 local.
FAILED: Build did NOT complete successfully
xavier@xavier-desktop:/media/xavier/storage/tensorflow$

Did you already try the official TF builds available here:

I see you are already using gcc-5. Did you check all other prerequisites for building TF including Bazel version etc?

You will have to apply some patches to TF like here:

I remember that for some versions TensorRT support is broken and you will have to disable it

Yes i already saw those they are for python 3.6 but as i mentioned i need it for python 3.7 because we found the right camera for our purpose but the manufacturers SDK only supports python 3.7 as minimum for some reason something to do with the API and C++

I did the patch mentioned on here, I was able to build TF 1.13 wwithout TensorRT: Building Tensorflow 1.13 on Jetson Xavier

I’ll read the link you sent me and try again

Thank you!

Hi again,

I followed the steps and i am building 1.15 natively on the JEtson Xavier but now i have a new error, hoping you could help m

Thanks!

ERROR: /media/xavier/storage/tensorflow/tensorflow/lite/kernels/BUILD:375:1: C++ compilation of rule ‘//tensorflow/lite/kernels:builtin_op_kernels’ failed (Exit 1): crosstool_wrapper_driver_is_not_gcc failed: error executing command
(cd /home/xavier/.cache/bazel/bazel_xavier/49531cc0ad8c2ca1aee55603d0e1d4ef/execroot/org_tensorflow &&
exec env -
CUDA_TOOLKIT_PATH=/usr/local/cuda
GCC_HOST_COMPILER_PATH=/usr/bin/gcc-5
LD_LIBRARY_PATH=/usr/local/cuda-10.0/lib64:
PATH=/home/xavier/bazel-0.24.1-dist/output:/usr/local/cuda-10.0/bin:/home/xavier/.local/bin:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin
PWD=/proc/self/cwd
PYTHON_BIN_PATH=/usr/bin/python3.7
PYTHON_LIB_PATH=/usr/local/lib/python3.7/dist-packages
TF2_BEHAVIOR=0
TF_CONFIGURE_IOS=0
TF_CUDA_COMPUTE_CAPABILITIES=7.2
TF_ENABLE_XLA=1
TF_NEED_CUDA=1
TF_NEED_TENSORRT=1
external/local_config_cuda/crosstool/clang/bin/crosstool_wrapper_driver_is_not_gcc -MD -MF bazel-out/aarch64-opt/bin/tensorflow/lite/kernels/objs/builtin_op_kernels/depthwise_conv.pic.d '-frandom-seed=bazel-out/aarch64-opt/bin/tensorflow/lite/kernels/objs/builtin_op_kernels/depthwise_conv.pic.o’ -DTFLITE_WITH_RUY -DEIGEN_MPL2_ONLY ‘-DEIGEN_MAX_ALIGN_BYTES=64’ ‘-DEIGEN_HAS_TYPE_TRAITS=0’ -iquote . -iquote bazel-out/aarch64-opt/genfiles -iquote bazel-out/aarch64-opt/bin -iquote external/gemmlowp -iquote bazel-out/aarch64-opt/genfiles/external/gemmlowp -iquote bazel-out/aarch64-opt/bin/external/gemmlowp -iquote external/arm_neon_2_x86_sse -iquote bazel-out/aarch64-opt/genfiles/external/arm_neon_2_x86_sse -iquote bazel-out/aarch64-opt/bin/external/arm_neon_2_x86_sse -iquote external/eigen_archive -iquote bazel-out/aarch64-opt/genfiles/external/eigen_archive -iquote bazel-out/aarch64-opt/bin/external/eigen_archive -iquote external/local_config_sycl -iquote bazel-out/aarch64-opt/genfiles/external/local_config_sycl -iquote bazel-out/aarch64-opt/bin/external/local_config_sycl -iquote external/flatbuffers -iquote bazel-out/aarch64-opt/genfiles/external/flatbuffers -iquote bazel-out/aarch64-opt/bin/external/flatbuffers -iquote external/fft2d -iquote bazel-out/aarch64-opt/genfiles/external/fft2d -iquote bazel-out/aarch64-opt/bin/external/fft2d -iquote external/farmhash_archive -iquote bazel-out/aarch64-opt/genfiles/external/farmhash_archive -iquote bazel-out/aarch64-opt/bin/external/farmhash_archive -isystem external/eigen_archive -isystem bazel-out/aarch64-opt/genfiles/external/eigen_archive -isystem bazel-out/aarch64-opt/bin/external/eigen_archive -isystem external/flatbuffers/include -isystem bazel-out/aarch64-opt/genfiles/external/flatbuffers/include -isystem bazel-out/aarch64-opt/bin/external/flatbuffers/include -isystem tensorflow/lite/schema -isystem bazel-out/aarch64-opt/genfiles/tensorflow/lite/schema -isystem bazel-out/aarch64-opt/bin/tensorflow/lite/schema -isystem external/farmhash_archive/src -isystem bazel-out/aarch64-opt/genfiles/external/farmhash_archive/src -isystem bazel-out/aarch64-opt/bin/external/farmhash_archive/src ‘-std=c++11’ -Wno-builtin-macro-redefined '-D__DATE=“redacted”’ '-D__TIMESTAMP
_=“redacted”’ ‘-D__TIME__=“redacted”’ -fPIC -U_FORTIFY_SOURCE ‘-D_FORTIFY_SOURCE=1’ -fstack-protector -Wall -fno-omit-frame-pointer -no-canonical-prefixes -fno-canonical-system-headers -DNDEBUG -g0 -O2 -ffunction-sections -fdata-sections -w ‘-march=native’ -Wno-sign-compare ‘-D_GLIBCXX_USE_CXX11_ABI=0’ -DFARMHASH_NO_CXX_STRING -Wno-sign-compare ‘-Wno-error=reorder’ -c tensorflow/lite/kernels/depthwise_conv.cc -o bazel-out/aarch64-opt/bin/tensorflow/lite/kernels/objs/builtin_op_kernels/depthwise_conv.pic.o)
Execution platform: @bazel_tools//platforms:host_platform
In file included from ./tensorflow/lite/kernels/internal/optimized/depthwiseconv_multithread.h:22:0,
from tensorflow/lite/kernels/depthwise_conv.cc:29:
./tensorflow/lite/kernels/internal/optimized/depthwiseconv_uint8.h:2097:2: error: #endif without #if
#endif // TENSORFLOW_LITE_KERNELS_INTERNAL_OPTIMIZED_DEPTHWISECONV_UINT8_H

^
Target //tensorflow/tools/pip_package:build_pip_package failed to build
INFO: Elapsed time: 3528.294s, Critical Path: 303.11s
INFO: 2482 processes: 2482 local.
FAILED: Build did NOT complete successfully
xavier@xavier-desktop:/media/xavier/storage/tensorflow$

You may try different gcc version from the 5.x branch, e.g. 5.2 or 5.3.

As said before: TensorRT support seems to be broken for a number of TF 1.x versions. See the huge number of issues in the TF Github when searching for “tensorrt”
I would first try building with TensorRT disabled (it is on by default when running ./configure) and when that was succesful proceed from there.

1 Like

Thanks I think i figured it out. During the edits on that github link, in the conv_unit8.h section when you have to add the #if

In TF 1.15 there are two #if statements so i replaced both by one liner as shown in the github.

And it complained that theres an endif without an if. I fixed that and currently it is compiling. I have TF1.15 compiling (with tensor RT) on the xavier and i setup a Ubuntu desktop on an old workstation and I am compiling TF 1.14 for the Xavier on there. Currently at 19,432/19457 so its looking good.

If all succeeds, i’ll be sharing the wheels on this forum

1 Like

Ok i have good news and bad news. On my ubuntu desktop (x86 machine) i was able to compile TF 1.14 for Python3.7 AND including Tensor RT but the wheel name turns out to be for linux_x86 so when i tried to install this on my jetson it throws this error:

tensorflow-1.14.1-cp37-cp37m-linux_x86_64.whl is not a supported wheel on this platform.

What flag or additional step should I do to be able to compile wheels for an aarch64 from an x_86 desktop?

Thanks!
Moth

I did it!

2 Likes

Really good to hear this.
Thanks for sharing this with us.

1 Like