Can’t compile standalone inferencing on RPI

Hello, I was following the instructions in this doc


after i ran sh build.sh on my RPI through ssh, the whole system hangs and shows g++: fatal error: Killed signal terminated program cc1plus. Are there any ways to fix this? i tried increasing the swap size but it still doesn’t work

Hi @weijunawj,

What are the specs of your RPi? Also, can you paste the output where it hangs?

Thanks!

Hi jenny!

I’m using RPi model 3B with a 32bit Raspbian OS and 16Gb sd card. this is my output from the ssh when i tried running it again.

Building standalone classifier
cc -Wall -g -DTF_LITE_DISABLE_X86_NEON=1 -DEI_CLASSIFIER_ENABLE_DETECTION_POSTPROCESS_OP=1 -Wno-strict-aliasing -I. -Os -DNDEBUG -g -c edge-impulse-sdk/CMSIS/DSP/Source/TransformFunctions/arm_cfft_f32.c -o edge-impulse-sdk/CMSIS/DSP/Source/TransformFunctions/arm_cfft_f32.o
cc -Wall -g -DTF_LITE_DISABLE_X86_NEON=1 -DEI_CLASSIFIER_ENABLE_DETECTION_POSTPROCESS_OP=1 -Wno-strict-aliasing -I. -Os -DNDEBUG -g -c edge-impulse-sdk/CMSIS/DSP/Source/TransformFunctions/arm_cfft_radix2_q15.c -o edge-impulse-sdk/CMSIS/DSP/Source/TransformFunctions/arm_cfft_radix2_q15.o
cc -Wall -g -DTF_LITE_DISABLE_X86_NEON=1 -DEI_CLASSIFIER_ENABLE_DETECTION_POSTPROCESS_OP=1 -Wno-strict-aliasing -I. -Os -DNDEBUG -g -c edge-impulse-sdk/CMSIS/DSP/Source/TransformFunctions/arm_cfft_radix4_q15.c -o edge-impulse-sdk/CMSIS/DSP/Source/TransformFunctions/arm_cfft_radix4_q15.o
cc -Wall -g -DTF_LITE_DISABLE_X86_NEON=1 -DEI_CLASSIFIER_ENABLE_DETECTION_POSTPROCESS_OP=1 -Wno-strict-aliasing -I. -Os -DNDEBUG -g -c edge-impulse-sdk/CMSIS/DSP/Source/TransformFunctions/arm_cfft_radix8_f32.c -o edge-impulse-sdk/CMSIS/DSP/Source/TransformFunctions/arm_cfft_radix8_f32.o
cc -Wall -g -DTF_LITE_DISABLE_X86_NEON=1 -DEI_CLASSIFIER_ENABLE_DETECTION_POSTPROCESS_OP=1 -Wno-strict-aliasing -I. -Os -DNDEBUG -g -c edge-impulse-sdk/CMSIS/DSP/Source/TransformFunctions/arm_rfft_q31.c -o edge-impulse-sdk/CMSIS/DSP/Source/TransformFunctions/arm_rfft_q31.o
cc -Wall -g -DTF_LITE_DISABLE_X86_NEON=1 -DEI_CLASSIFIER_ENABLE_DETECTION_POSTPROCESS_OP=1 -Wno-strict-aliasing -I. -Os -DNDEBUG -g -c edge-impulse-sdk/CMSIS/DSP/Source/CommonTables/arm_common_tables.c -o edge-impulse-sdk/CMSIS/DSP/Source/CommonTables/arm_common_tables.o
g++ -Wall -g -DTF_LITE_DISABLE_X86_NEON=1 -DEI_CLASSIFIER_ENABLE_DETECTION_POSTPROCESS_OP=1 -Wno-strict-aliasing -I. -Os -DNDEBUG -g -std=c++14 -c source/main.cpp -o source/main.o
g++ -Wall -g -DTF_LITE_DISABLE_X86_NEON=1 -DEI_CLASSIFIER_ENABLE_DETECTION_POSTPROCESS_OP=1 -Wno-strict-aliasing -I. -Os -DNDEBUG -g -std=c++14 -c tflite-model/trained_model_compiled.cpp -o tflite-model/trained_model_compiled.o
g++ -Wall -g -DTF_LITE_DISABLE_X86_NEON=1 -DEI_CLASSIFIER_ENABLE_DETECTION_POSTPROCESS_OP=1 -Wno-strict-aliasing -I. -Os -DNDEBUG -g -std=c++14 -c edge-impulse-sdk/dsp/kissfft/kiss_fft.cpp -o edge-impulse-sdk/dsp/kissfft/kiss_fft.o
g++ -Wall -g -DTF_LITE_DISABLE_X86_NEON=1 -DEI_CLASSIFIER_ENABLE_DETECTION_POSTPROCESS_OP=1 -Wno-strict-aliasing -I. -Os -DNDEBUG -g -std=c++14 -c edge-impulse-sdk/dsp/kissfft/kiss_fftr.cpp -o edge-impulse-sdk/dsp/kissfft/kiss_fftr.o
g++ -Wall -g -DTF_LITE_DISABLE_X86_NEON=1 -DEI_CLASSIFIER_ENABLE_DETECTION_POSTPROCESS_OP=1 -Wno-strict-aliasing -I. -Os -DNDEBUG -g -std=c++14 -c edge-impulse-sdk/dsp/dct/fast-dct-fft.cpp -o edge-impulse-sdk/dsp/dct/fast-dct-fft.o
g++ -Wall -g -DTF_LITE_DISABLE_X86_NEON=1 -DEI_CLASSIFIER_ENABLE_DETECTION_POSTPROCESS_OP=1 -Wno-strict-aliasing -I. -Os -DNDEBUG -g -std=c++14 -c edge-impulse-sdk/dsp/memory.cpp -o edge-impulse-sdk/dsp/memory.o
g++ -Wall -g -DTF_LITE_DISABLE_X86_NEON=1 -DEI_CLASSIFIER_ENABLE_DETECTION_POSTPROCESS_OP=1 -Wno-strict-aliasing -I. -Os -DNDEBUG -g -std=c++14 -c edge-impulse-sdk/porting/posix/ei_classifier_porting.cpp -o edge-impulse-sdk/porting/posix/ei_classifier_porting.o
g++ -Wall -g -DTF_LITE_DISABLE_X86_NEON=1 -DEI_CLASSIFIER_ENABLE_DETECTION_POSTPROCESS_OP=1 -Wno-strict-aliasing -I. -Os -DNDEBUG -g -std=c++14 -c edge-impulse-sdk/tensorflow/lite/kernels/kernel_util_lite.cc -o edge-impulse-sdk/tensorflow/lite/kernels/kernel_util_lite.o
g++ -Wall -g -DTF_LITE_DISABLE_X86_NEON=1 -DEI_CLASSIFIER_ENABLE_DETECTION_POSTPROCESS_OP=1 -Wno-strict-aliasing -I. -Os -DNDEBUG -g -std=c++14 -c edge-impulse-sdk/tensorflow/lite/kernels/internal/quantization_util.cc -o edge-impulse-sdk/tensorflow/lite/kernels/internal/quantization_util.o
g++ -Wall -g -DTF_LITE_DISABLE_X86_NEON=1 -DEI_CLASSIFIER_ENABLE_DETECTION_POSTPROCESS_OP=1 -Wno-strict-aliasing -I. -Os -DNDEBUG -g -std=c++14 -c edge-impulse-sdk/tensorflow/lite/micro/kernels/activations.cc -o edge-impulse-sdk/tensorflow/lite/micro/kernels/activations.o
g++ -Wall -g -DTF_LITE_DISABLE_X86_NEON=1 -DEI_CLASSIFIER_ENABLE_DETECTION_POSTPROCESS_OP=1 -Wno-strict-aliasing -I. -Os -DNDEBUG -g -std=c++14 -c edge-impulse-sdk/tensorflow/lite/micro/kernels/add.cc -o edge-impulse-sdk/tensorflow/lite/micro/kernels/add.o
g++ -Wall -g -DTF_LITE_DISABLE_X86_NEON=1 -DEI_CLASSIFIER_ENABLE_DETECTION_POSTPROCESS_OP=1 -Wno-strict-aliasing -I. -Os -DNDEBUG -g -std=c++14 -c edge-impulse-sdk/tensorflow/lite/micro/kernels/add_n.cc -o edge-impulse-sdk/tensorflow/lite/micro/kernels/add_n.o
g++ -Wall -g -DTF_LITE_DISABLE_X86_NEON=1 -DEI_CLASSIFIER_ENABLE_DETECTION_POSTPROCESS_OP=1 -Wno-strict-aliasing -I. -Os -DNDEBUG -g -std=c++14 -c edge-impulse-sdk/tensorflow/lite/micro/kernels/arg_min_max.cc -o edge-impulse-sdk/tensorflow/lite/micro/kernels/arg_min_max.o
g++ -Wall -g -DTF_LITE_DISABLE_X86_NEON=1 -DEI_CLASSIFIER_ENABLE_DETECTION_POSTPROCESS_OP=1 -Wno-strict-aliasing -I. -Os -DNDEBUG -g -std=c++14 -c edge-impulse-sdk/tensorflow/lite/micro/kernels/batch_to_space_nd.cc -o edge-impulse-sdk/tensorflow/lite/micro/kernels/batch_to_space_nd.o
g++ -Wall -g -DTF_LITE_DISABLE_X86_NEON=1 -DEI_CLASSIFIER_ENABLE_DETECTION_POSTPROCESS_OP=1 -Wno-strict-aliasing -I. -Os -DNDEBUG -g -std=c++14 -c edge-impulse-sdk/tensorflow/lite/micro/kernels/cast.cc -o edge-impulse-sdk/tensorflow/lite/micro/kernels/cast.o
g++ -Wall -g -DTF_LITE_DISABLE_X86_NEON=1 -DEI_CLASSIFIER_ENABLE_DETECTION_POSTPROCESS_OP=1 -Wno-strict-aliasing -I. -Os -DNDEBUG -g -std=c++14 -c edge-impulse-sdk/tensorflow/lite/micro/kernels/ceil.cc -o edge-impulse-sdk/tensorflow/lite/micro/kernels/ceil.o
g++ -Wall -g -DTF_LITE_DISABLE_X86_NEON=1 -DEI_CLASSIFIER_ENABLE_DETECTION_POSTPROCESS_OP=1 -Wno-strict-aliasing -I. -Os -DNDEBUG -g -std=c++14 -c edge-impulse-sdk/tensorflow/lite/micro/kernels/circular_buffer.cc -o edge-impulse-sdk/tensorflow/lite/micro/kernels/circular_buffer.o
g++ -Wall -g -DTF_LITE_DISABLE_X86_NEON=1 -DEI_CLASSIFIER_ENABLE_DETECTION_POSTPROCESS_OP=1 -Wno-strict-aliasing -I. -Os -DNDEBUG -g -std=c++14 -c edge-impulse-sdk/tensorflow/lite/micro/kernels/comparisons.cc -o edge-impulse-sdk/tensorflow/lite/micro/kernels/comparisons.o
g++ -Wall -g -DTF_LITE_DISABLE_X86_NEON=1 -DEI_CLASSIFIER_ENABLE_DETECTION_POSTPROCESS_OP=1 -Wno-strict-aliasing -I. -Os -DNDEBUG -g -std=c++14 -c edge-impulse-sdk/tensorflow/lite/micro/kernels/concatenation.cc -o edge-impulse-sdk/tensorflow/lite/micro/kernels/concatenation.o
g++ -Wall -g -DTF_LITE_DISABLE_X86_NEON=1 -DEI_CLASSIFIER_ENABLE_DETECTION_POSTPROCESS_OP=1 -Wno-strict-aliasing -I. -Os -DNDEBUG -g -std=c++14 -c edge-impulse-sdk/tensorflow/lite/micro/kernels/conv.cc -o edge-impulse-sdk/tensorflow/lite/micro/kernels/conv.o
g++ -Wall -g -DTF_LITE_DISABLE_X86_NEON=1 -DEI_CLASSIFIER_ENABLE_DETECTION_POSTPROCESS_OP=1 -Wno-strict-aliasing -I. -Os -DNDEBUG -g -std=c++14 -c edge-impulse-sdk/tensorflow/lite/micro/kernels/conv_common.cc -o edge-impulse-sdk/tensorflow/lite/micro/kernels/conv_common.o
g++ -Wall -g -DTF_LITE_DISABLE_X86_NEON=1 -DEI_CLASSIFIER_ENABLE_DETECTION_POSTPROCESS_OP=1 -Wno-strict-aliasing -I. -Os -DNDEBUG -g -std=c++14 -c edge-impulse-sdk/tensorflow/lite/micro/kernels/depthwise_conv.cc -o edge-impulse-sdk/tensorflow/lite/micro/kernels/depthwise_conv.o
g++ -Wall -g -DTF_LITE_DISABLE_X86_NEON=1 -DEI_CLASSIFIER_ENABLE_DETECTION_POSTPROCESS_OP=1 -Wno-strict-aliasing -I. -Os -DNDEBUG -g -std=c++14 -c edge-impulse-sdk/tensorflow/lite/micro/kernels/depthwise_conv_common.cc -o edge-impulse-sdk/tensorflow/lite/micro/kernels/depthwise_conv_common.o
g++ -Wall -g -DTF_LITE_DISABLE_X86_NEON=1 -DEI_CLASSIFIER_ENABLE_DETECTION_POSTPROCESS_OP=1 -Wno-strict-aliasing -I. -Os -DNDEBUG -g -std=c++14 -c edge-impulse-sdk/tensorflow/lite/micro/kernels/dequantize.cc -o edge-impulse-sdk/tensorflow/lite/micro/kernels/dequantize.o
g++ -Wall -g -DTF_LITE_DISABLE_X86_NEON=1 -DEI_CLASSIFIER_ENABLE_DETECTION_POSTPROCESS_OP=1 -Wno-strict-aliasing -I. -Os -DNDEBUG -g -std=c++14 -c edge-impulse-sdk/tensorflow/lite/micro/kernels/div.cc -o edge-impulse-sdk/tensorflow/lite/micro/kernels/div.o
g++ -Wall -g -DTF_LITE_DISABLE_X86_NEON=1 -DEI_CLASSIFIER_ENABLE_DETECTION_POSTPROCESS_OP=1 -Wno-strict-aliasing -I. -Os -DNDEBUG -g -std=c++14 -c edge-impulse-sdk/tensorflow/lite/micro/kernels/elementwise.cc -o edge-impulse-sdk/tensorflow/lite/micro/kernels/elementwise.o
g++ -Wall -g -DTF_LITE_DISABLE_X86_NEON=1 -DEI_CLASSIFIER_ENABLE_DETECTION_POSTPROCESS_OP=1 -Wno-strict-aliasing -I. -Os -DNDEBUG -g -std=c++14 -c edge-impulse-sdk/tensorflow/lite/micro/kernels/elu.cc -o edge-impulse-sdk/tensorflow/lite/micro/kernels/elu.o
g++ -Wall -g -DTF_LITE_DISABLE_X86_NEON=1 -DEI_CLASSIFIER_ENABLE_DETECTION_POSTPROCESS_OP=1 -Wno-strict-aliasing -I. -Os -DNDEBUG -g -std=c++14 -c edge-impulse-sdk/tensorflow/lite/micro/kernels/ethosu.cc -o edge-impulse-sdk/tensorflow/lite/micro/kernels/ethosu.o
g++ -Wall -g -DTF_LITE_DISABLE_X86_NEON=1 -DEI_CLASSIFIER_ENABLE_DETECTION_POSTPROCESS_OP=1 -Wno-strict-aliasing -I. -Os -DNDEBUG -g -std=c++14 -c edge-impulse-sdk/tensorflow/lite/micro/kernels/expand_dims.cc -o edge-impulse-sdk/tensorflow/lite/micro/kernels/expand_dims.o
g++ -Wall -g -DTF_LITE_DISABLE_X86_NEON=1 -DEI_CLASSIFIER_ENABLE_DETECTION_POSTPROCESS_OP=1 -Wno-strict-aliasing -I. -Os -DNDEBUG -g -std=c++14 -c edge-impulse-sdk/tensorflow/lite/micro/kernels/exp.cc -o edge-impulse-sdk/tensorflow/lite/micro/kernels/exp.o
g++ -Wall -g -DTF_LITE_DISABLE_X86_NEON=1 -DEI_CLASSIFIER_ENABLE_DETECTION_POSTPROCESS_OP=1 -Wno-strict-aliasing -I. -Os -DNDEBUG -g -std=c++14 -c edge-impulse-sdk/tensorflow/lite/micro/kernels/fill.cc -o edge-impulse-sdk/tensorflow/lite/micro/kernels/fill.o
g++ -Wall -g -DTF_LITE_DISABLE_X86_NEON=1 -DEI_CLASSIFIER_ENABLE_DETECTION_POSTPROCESS_OP=1 -Wno-strict-aliasing -I. -Os -DNDEBUG -g -std=c++14 -c edge-impulse-sdk/tensorflow/lite/micro/kernels/floor.cc -o edge-impulse-sdk/tensorflow/lite/micro/kernels/floor.o
g++ -Wall -g -DTF_LITE_DISABLE_X86_NEON=1 -DEI_CLASSIFIER_ENABLE_DETECTION_POSTPROCESS_OP=1 -Wno-strict-aliasing -I. -Os -DNDEBUG -g -std=c++14 -c edge-impulse-sdk/tensorflow/lite/micro/kernels/fully_connected.cc -o edge-impulse-sdk/tensorflow/lite/micro/kernels/fully_connected.o
g++ -Wall -g -DTF_LITE_DISABLE_X86_NEON=1 -DEI_CLASSIFIER_ENABLE_DETECTION_POSTPROCESS_OP=1 -Wno-strict-aliasing -I. -Os -DNDEBUG -g -std=c++14 -c edge-impulse-sdk/tensorflow/lite/micro/kernels/fully_connected_common.cc -o edge-impulse-sdk/tensorflow/lite/micro/kernels/fully_connected_common.o
g++ -Wall -g -DTF_LITE_DISABLE_X86_NEON=1 -DEI_CLASSIFIER_ENABLE_DETECTION_POSTPROCESS_OP=1 -Wno-strict-aliasing -I. -Os -DNDEBUG -g -std=c++14 -c edge-impulse-sdk/tensorflow/lite/micro/kernels/hard_swish.cc -o edge-impulse-sdk/tensorflow/lite/micro/kernels/hard_swish.o
g++ -Wall -g -DTF_LITE_DISABLE_X86_NEON=1 -DEI_CLASSIFIER_ENABLE_DETECTION_POSTPROCESS_OP=1 -Wno-strict-aliasing -I. -Os -DNDEBUG -g -std=c++14 -c edge-impulse-sdk/tensorflow/lite/micro/kernels/kernel_util_micro.cc -o edge-impulse-sdk/tensorflow/lite/micro/kernels/kernel_util_micro.o
g++ -Wall -g -DTF_LITE_DISABLE_X86_NEON=1 -DEI_CLASSIFIER_ENABLE_DETECTION_POSTPROCESS_OP=1 -Wno-strict-aliasing -I. -Os -DNDEBUG -g -std=c++14 -c edge-impulse-sdk/tensorflow/lite/micro/kernels/l2norm.cc -o edge-impulse-sdk/tensorflow/lite/micro/kernels/l2norm.o
g++ -Wall -g -DTF_LITE_DISABLE_X86_NEON=1 -DEI_CLASSIFIER_ENABLE_DETECTION_POSTPROCESS_OP=1 -Wno-strict-aliasing -I. -Os -DNDEBUG -g -std=c++14 -c edge-impulse-sdk/tensorflow/lite/micro/kernels/l2_pool_2d.cc -o edge-impulse-sdk/tensorflow/lite/micro/kernels/l2_pool_2d.o
g++ -Wall -g -DTF_LITE_DISABLE_X86_NEON=1 -DEI_CLASSIFIER_ENABLE_DETECTION_POSTPROCESS_OP=1 -Wno-strict-aliasing -I. -Os -DNDEBUG -g -std=c++14 -c edge-impulse-sdk/tensorflow/lite/micro/kernels/leaky_relu.cc -o edge-impulse-sdk/tensorflow/lite/micro/kernels/leaky_relu.o
g++ -Wall -g -DTF_LITE_DISABLE_X86_NEON=1 -DEI_CLASSIFIER_ENABLE_DETECTION_POSTPROCESS_OP=1 -Wno-strict-aliasing -I. -Os -DNDEBUG -g -std=c++14 -c edge-impulse-sdk/tensorflow/lite/micro/kernels/logical.cc -o edge-impulse-sdk/tensorflow/lite/micro/kernels/logical.o
g++ -Wall -g -DTF_LITE_DISABLE_X86_NEON=1 -DEI_CLASSIFIER_ENABLE_DETECTION_POSTPROCESS_OP=1 -Wno-strict-aliasing -I. -Os -DNDEBUG -g -std=c++14 -c edge-impulse-sdk/tensorflow/lite/micro/kernels/logistic.cc -o edge-impulse-sdk/tensorflow/lite/micro/kernels/logistic.o
g++ -Wall -g -DTF_LITE_DISABLE_X86_NEON=1 -DEI_CLASSIFIER_ENABLE_DETECTION_POSTPROCESS_OP=1 -Wno-strict-aliasing -I. -Os -DNDEBUG -g -std=c++14 -c edge-impulse-sdk/tensorflow/lite/micro/kernels/maximum_minimum.cc -o edge-impulse-sdk/tensorflow/lite/micro/kernels/maximum_minimum.o
g++ -Wall -g -DTF_LITE_DISABLE_X86_NEON=1 -DEI_CLASSIFIER_ENABLE_DETECTION_POSTPROCESS_OP=1 -Wno-strict-aliasing -I. -Os -DNDEBUG -g -std=c++14 -c edge-impulse-sdk/tensorflow/lite/micro/kernels/mul.cc -o edge-impulse-sdk/tensorflow/lite/micro/kernels/mul.o
g++ -Wall -g -DTF_LITE_DISABLE_X86_NEON=1 -DEI_CLASSIFIER_ENABLE_DETECTION_POSTPROCESS_OP=1 -Wno-strict-aliasing -I. -Os -DNDEBUG -g -std=c++14 -c edge-impulse-sdk/tensorflow/lite/micro/kernels/neg.cc -o edge-impulse-sdk/tensorflow/lite/micro/kernels/neg.o
g++ -Wall -g -DTF_LITE_DISABLE_X86_NEON=1 -DEI_CLASSIFIER_ENABLE_DETECTION_POSTPROCESS_OP=1 -Wno-strict-aliasing -I. -Os -DNDEBUG -g -std=c++14 -c edge-impulse-sdk/tensorflow/lite/micro/kernels/pack.cc -o edge-impulse-sdk/tensorflow/lite/micro/kernels/pack.o
g++ -Wall -g -DTF_LITE_DISABLE_X86_NEON=1 -DEI_CLASSIFIER_ENABLE_DETECTION_POSTPROCESS_OP=1 -Wno-strict-aliasing -I. -Os -DNDEBUG -g -std=c++14 -c edge-impulse-sdk/tensorflow/lite/micro/kernels/pad.cc -o edge-impulse-sdk/tensorflow/lite/micro/kernels/pad.o
g++ -Wall -g -DTF_LITE_DISABLE_X86_NEON=1 -DEI_CLASSIFIER_ENABLE_DETECTION_POSTPROCESS_OP=1 -Wno-strict-aliasing -I. -Os -DNDEBUG -g -std=c++14 -c edge-impulse-sdk/tensorflow/lite/micro/kernels/pooling.cc -o edge-impulse-sdk/tensorflow/lite/micro/kernels/pooling.o
g++ -Wall -g -DTF_LITE_DISABLE_X86_NEON=1 -DEI_CLASSIFIER_ENABLE_DETECTION_POSTPROCESS_OP=1 -Wno-strict-aliasing -I. -Os -DNDEBUG -g -std=c++14 -c edge-impulse-sdk/tensorflow/lite/micro/kernels/prelu.cc -o edge-impulse-sdk/tensorflow/lite/micro/kernels/prelu.o
g++ -Wall -g -DTF_LITE_DISABLE_X86_NEON=1 -DEI_CLASSIFIER_ENABLE_DETECTION_POSTPROCESS_OP=1 -Wno-strict-aliasing -I. -Os -DNDEBUG -g -std=c++14 -c edge-impulse-sdk/tensorflow/lite/micro/kernels/quantize.cc -o edge-impulse-sdk/tensorflow/lite/micro/kernels/quantize.o
g++ -Wall -g -DTF_LITE_DISABLE_X86_NEON=1 -DEI_CLASSIFIER_ENABLE_DETECTION_POSTPROCESS_OP=1 -Wno-strict-aliasing -I. -Os -DNDEBUG -g -std=c++14 -c edge-impulse-sdk/tensorflow/lite/micro/kernels/quantize_common.cc -o edge-impulse-sdk/tensorflow/lite/micro/kernels/quantize_common.o
g++ -Wall -g -DTF_LITE_DISABLE_X86_NEON=1 -DEI_CLASSIFIER_ENABLE_DETECTION_POSTPROCESS_OP=1 -Wno-strict-aliasing -I. -Os -DNDEBUG -g -std=c++14 -c edge-impulse-sdk/tensorflow/lite/micro/kernels/reduce.cc -o edge-impulse-sdk/tensorflow/lite/micro/kernels/reduce.o
g++ -Wall -g -DTF_LITE_DISABLE_X86_NEON=1 -DEI_CLASSIFIER_ENABLE_DETECTION_POSTPROCESS_OP=1 -Wno-strict-aliasing -I. -Os -DNDEBUG -g -std=c++14 -c edge-impulse-sdk/tensorflow/lite/micro/kernels/reshape.cc -o edge-impulse-sdk/tensorflow/lite/micro/kernels/reshape.o
g++ -Wall -g -DTF_LITE_DISABLE_X86_NEON=1 -DEI_CLASSIFIER_ENABLE_DETECTION_POSTPROCESS_OP=1 -Wno-strict-aliasing -I. -Os -DNDEBUG -g -std=c++14 -c edge-impulse-sdk/tensorflow/lite/micro/kernels/resize_nearest_neighbor.cc -o edge-impulse-sdk/tensorflow/lite/micro/kernels/resize_nearest_neighbor.o
g++ -Wall -g -DTF_LITE_DISABLE_X86_NEON=1 -DEI_CLASSIFIER_ENABLE_DETECTION_POSTPROCESS_OP=1 -Wno-strict-aliasing -I. -Os -DNDEBUG -g -std=c++14 -c edge-impulse-sdk/tensorflow/lite/micro/kernels/round.cc -o edge-impulse-sdk/tensorflow/lite/micro/kernels/round.o
g++ -Wall -g -DTF_LITE_DISABLE_X86_NEON=1 -DEI_CLASSIFIER_ENABLE_DETECTION_POSTPROCESS_OP=1 -Wno-strict-aliasing -I. -Os -DNDEBUG -g -std=c++14 -c edge-impulse-sdk/tensorflow/lite/micro/kernels/shape.cc -o edge-impulse-sdk/tensorflow/lite/micro/kernels/shape.o
g++ -Wall -g -DTF_LITE_DISABLE_X86_NEON=1 -DEI_CLASSIFIER_ENABLE_DETECTION_POSTPROCESS_OP=1 -Wno-strict-aliasing -I. -Os -DNDEBUG -g -std=c++14 -c edge-impulse-sdk/tensorflow/lite/micro/kernels/softmax.cc -o edge-impulse-sdk/tensorflow/lite/micro/kernels/softmax.o
g++ -Wall -g -DTF_LITE_DISABLE_X86_NEON=1 -DEI_CLASSIFIER_ENABLE_DETECTION_POSTPROCESS_OP=1 -Wno-strict-aliasing -I. -Os -DNDEBUG -g -std=c++14 -c edge-impulse-sdk/tensorflow/lite/micro/kernels/softmax_common.cc -o edge-impulse-sdk/tensorflow/lite/micro/kernels/softmax_common.o
g++ -Wall -g -DTF_LITE_DISABLE_X86_NEON=1 -DEI_CLASSIFIER_ENABLE_DETECTION_POSTPROCESS_OP=1 -Wno-strict-aliasing -I. -Os -DNDEBUG -g -std=c++14 -c edge-impulse-sdk/tensorflow/lite/micro/kernels/space_to_batch_nd.cc -o edge-impulse-sdk/tensorflow/lite/micro/kernels/space_to_batch_nd.o
g++ -Wall -g -DTF_LITE_DISABLE_X86_NEON=1 -DEI_CLASSIFIER_ENABLE_DETECTION_POSTPROCESS_OP=1 -Wno-strict-aliasing -I. -Os -DNDEBUG -g -std=c++14 -c edge-impulse-sdk/tensorflow/lite/micro/kernels/split.cc -o edge-impulse-sdk/tensorflow/lite/micro/kernels/split.o
g++ -Wall -g -DTF_LITE_DISABLE_X86_NEON=1 -DEI_CLASSIFIER_ENABLE_DETECTION_POSTPROCESS_OP=1 -Wno-strict-aliasing -I. -Os -DNDEBUG -g -std=c++14 -c edge-impulse-sdk/tensorflow/lite/micro/kernels/split_v.cc -o edge-impulse-sdk/tensorflow/lite/micro/kernels/split_v.o
g++ -Wall -g -DTF_LITE_DISABLE_X86_NEON=1 -DEI_CLASSIFIER_ENABLE_DETECTION_POSTPROCESS_OP=1 -Wno-strict-aliasing -I. -Os -DNDEBUG -g -std=c++14 -c edge-impulse-sdk/tensorflow/lite/micro/kernels/squeeze.cc -o edge-impulse-sdk/tensorflow/lite/micro/kernels/squeeze.o
g++ -Wall -g -DTF_LITE_DISABLE_X86_NEON=1 -DEI_CLASSIFIER_ENABLE_DETECTION_POSTPROCESS_OP=1 -Wno-strict-aliasing -I. -Os -DNDEBUG -g -std=c++14 -c edge-impulse-sdk/tensorflow/lite/micro/kernels/strided_slice.cc -o edge-impulse-sdk/tensorflow/lite/micro/kernels/strided_slice.o
g++ -Wall -g -DTF_LITE_DISABLE_X86_NEON=1 -DEI_CLASSIFIER_ENABLE_DETECTION_POSTPROCESS_OP=1 -Wno-strict-aliasing -I. -Os -DNDEBUG -g -std=c++14 -c edge-impulse-sdk/tensorflow/lite/micro/kernels/sub.cc -o edge-impulse-sdk/tensorflow/lite/micro/kernels/sub.o
g++ -Wall -g -DTF_LITE_DISABLE_X86_NEON=1 -DEI_CLASSIFIER_ENABLE_DETECTION_POSTPROCESS_OP=1 -Wno-strict-aliasing -I. -Os -DNDEBUG -g -std=c++14 -c edge-impulse-sdk/tensorflow/lite/micro/kernels/svdf.cc -o edge-impulse-sdk/tensorflow/lite/micro/kernels/svdf.o
g++ -Wall -g -DTF_LITE_DISABLE_X86_NEON=1 -DEI_CLASSIFIER_ENABLE_DETECTION_POSTPROCESS_OP=1 -Wno-strict-aliasing -I. -Os -DNDEBUG -g -std=c++14 -c edge-impulse-sdk/tensorflow/lite/micro/kernels/svdf_common.cc -o edge-impulse-sdk/tensorflow/lite/micro/kernels/svdf_common.o
g++ -Wall -g -DTF_LITE_DISABLE_X86_NEON=1 -DEI_CLASSIFIER_ENABLE_DETECTION_POSTPROCESS_OP=1 -Wno-strict-aliasing -I. -Os -DNDEBUG -g -std=c++14 -c edge-impulse-sdk/tensorflow/lite/micro/kernels/tanh.cc -o edge-impulse-sdk/tensorflow/lite/micro/kernels/tanh.o
g++ -Wall -g -DTF_LITE_DISABLE_X86_NEON=1 -DEI_CLASSIFIER_ENABLE_DETECTION_POSTPROCESS_OP=1 -Wno-strict-aliasing -I. -Os -DNDEBUG -g -std=c++14 -c edge-impulse-sdk/tensorflow/lite/micro/kernels/tflite_detection_postprocess.cc -o edge-impulse-sdk/tensorflow/lite/micro/kernels/tflite_detection_postprocess.o
g++ -Wall -g -DTF_LITE_DISABLE_X86_NEON=1 -DEI_CLASSIFIER_ENABLE_DETECTION_POSTPROCESS_OP=1 -Wno-strict-aliasing -I. -Os -DNDEBUG -g -std=c++14 -c edge-impulse-sdk/tensorflow/lite/micro/kernels/transpose_conv.cc -o edge-impulse-sdk/tensorflow/lite/micro/kernels/transpose_conv.o
g++ -Wall -g -DTF_LITE_DISABLE_X86_NEON=1 -DEI_CLASSIFIER_ENABLE_DETECTION_POSTPROCESS_OP=1 -Wno-strict-aliasing -I. -Os -DNDEBUG -g -std=c++14 -c edge-impulse-sdk/tensorflow/lite/micro/kernels/unpack.cc -o edge-impulse-sdk/tensorflow/lite/micro/kernels/unpack.o
g++ -Wall -g -DTF_LITE_DISABLE_X86_NEON=1 -DEI_CLASSIFIER_ENABLE_DETECTION_POSTPROCESS_OP=1 -Wno-strict-aliasing -I. -Os -DNDEBUG -g -std=c++14 -c edge-impulse-sdk/tensorflow/lite/micro/kernels/zeros_like.cc -o edge-impulse-sdk/tensorflow/lite/micro/kernels/zeros_like.o
g++ -Wall -g -DTF_LITE_DISABLE_X86_NEON=1 -DEI_CLASSIFIER_ENABLE_DETECTION_POSTPROCESS_OP=1 -Wno-strict-aliasing -I. -Os -DNDEBUG -g -std=c++14 -c edge-impulse-sdk/tensorflow/lite/micro/all_ops_resolver.cc -o edge-impulse-sdk/tensorflow/lite/micro/all_ops_resolver.o
g++ -Wall -g -DTF_LITE_DISABLE_X86_NEON=1 -DEI_CLASSIFIER_ENABLE_DETECTION_POSTPROCESS_OP=1 -Wno-strict-aliasing -I. -Os -DNDEBUG -g -std=c++14 -c edge-impulse-sdk/tensorflow/lite/micro/memory_helpers.cc -o edge-impulse-sdk/tensorflow/lite/micro/memory_helpers.o
g++ -Wall -g -DTF_LITE_DISABLE_X86_NEON=1 -DEI_CLASSIFIER_ENABLE_DETECTION_POSTPROCESS_OP=1 -Wno-strict-aliasing -I. -Os -DNDEBUG -g -std=c++14 -c edge-impulse-sdk/tensorflow/lite/micro/micro_allocator.cc -o edge-impulse-sdk/tensorflow/lite/micro/micro_allocator.o
g++ -Wall -g -DTF_LITE_DISABLE_X86_NEON=1 -DEI_CLASSIFIER_ENABLE_DETECTION_POSTPROCESS_OP=1 -Wno-strict-aliasing -I. -Os -DNDEBUG -g -std=c++14 -c edge-impulse-sdk/tensorflow/lite/micro/micro_error_reporter.cc -o edge-impulse-sdk/tensorflow/lite/micro/micro_error_reporter.o
g++ -Wall -g -DTF_LITE_DISABLE_X86_NEON=1 -DEI_CLASSIFIER_ENABLE_DETECTION_POSTPROCESS_OP=1 -Wno-strict-aliasing -I. -Os -DNDEBUG -g -std=c++14 -c edge-impulse-sdk/tensorflow/lite/micro/micro_interpreter.cc -o edge-impulse-sdk/tensorflow/lite/micro/micro_interpreter.o
g++ -Wall -g -DTF_LITE_DISABLE_X86_NEON=1 -DEI_CLASSIFIER_ENABLE_DETECTION_POSTPROCESS_OP=1 -Wno-strict-aliasing -I. -Os -DNDEBUG -g -std=c++14 -c edge-impulse-sdk/tensorflow/lite/micro/micro_profiler.cc -o edge-impulse-sdk/tensorflow/lite/micro/micro_profiler.o
g++ -Wall -g -DTF_LITE_DISABLE_X86_NEON=1 -DEI_CLASSIFIER_ENABLE_DETECTION_POSTPROCESS_OP=1 -Wno-strict-aliasing -I. -Os -DNDEBUG -g -std=c++14 -c edge-impulse-sdk/tensorflow/lite/micro/micro_string.cc -o edge-impulse-sdk/tensorflow/lite/micro/micro_string.o
g++ -Wall -g -DTF_LITE_DISABLE_X86_NEON=1 -DEI_CLASSIFIER_ENABLE_DETECTION_POSTPROCESS_OP=1 -Wno-strict-aliasing -I. -Os -DNDEBUG -g -std=c++14 -c edge-impulse-sdk/tensorflow/lite/micro/micro_utils.cc -o edge-impulse-sdk/tensorflow/lite/micro/micro_utils.o
g++ -Wall -g -DTF_LITE_DISABLE_X86_NEON=1 -DEI_CLASSIFIER_ENABLE_DETECTION_POSTPROCESS_OP=1 -Wno-strict-aliasing -I. -Os -DNDEBUG -g -std=c++14 -c edge-impulse-sdk/tensorflow/lite/micro/recording_micro_allocator.cc -o edge-impulse-sdk/tensorflow/lite/micro/recording_micro_allocator.o
g++ -Wall -g -DTF_LITE_DISABLE_X86_NEON=1 -DEI_CLASSIFIER_ENABLE_DETECTION_POSTPROCESS_OP=1 -Wno-strict-aliasing -I. -Os -DNDEBUG -g -std=c++14 -c edge-impulse-sdk/tensorflow/lite/micro/recording_simple_memory_allocator.cc -o edge-impulse-sdk/tensorflow/lite/micro/recording_simple_memory_allocator.o
g++ -Wall -g -DTF_LITE_DISABLE_X86_NEON=1 -DEI_CLASSIFIER_ENABLE_DETECTION_POSTPROCESS_OP=1 -Wno-strict-aliasing -I. -Os -DNDEBUG -g -std=c++14 -c edge-impulse-sdk/tensorflow/lite/micro/schema_utils.cc -o edge-impulse-sdk/tensorflow/lite/micro/schema_utils.o
g++ -Wall -g -DTF_LITE_DISABLE_X86_NEON=1 -DEI_CLASSIFIER_ENABLE_DETECTION_POSTPROCESS_OP=1 -Wno-strict-aliasing -I. -Os -DNDEBUG -g -std=c++14 -c edge-impulse-sdk/tensorflow/lite/micro/simple_memory_allocator.cc -o edge-impulse-sdk/tensorflow/lite/micro/simple_memory_allocator.o
g++ -Wall -g -DTF_LITE_DISABLE_X86_NEON=1 -DEI_CLASSIFIER_ENABLE_DETECTION_POSTPROCESS_OP=1 -Wno-strict-aliasing -I. -Os -DNDEBUG -g -std=c++14 -c edge-impulse-sdk/tensorflow/lite/micro/test_helpers.cc -o edge-impulse-sdk/tensorflow/lite/micro/test_helpers.o
g++ -Wall -g -DTF_LITE_DISABLE_X86_NEON=1 -DEI_CLASSIFIER_ENABLE_DETECTION_POSTPROCESS_OP=1 -Wno-strict-aliasing -I. -Os -DNDEBUG -g -std=c++14 -c edge-impulse-sdk/tensorflow/lite/micro/memory_planner/greedy_memory_planner.cc -o edge-impulse-sdk/tensorflow/lite/micro/memory_planner/greedy_memory_planner.o
g++ -Wall -g -DTF_LITE_DISABLE_X86_NEON=1 -DEI_CLASSIFIER_ENABLE_DETECTION_POSTPROCESS_OP=1 -Wno-strict-aliasing -I. -Os -DNDEBUG -g -std=c++14 -c edge-impulse-sdk/tensorflow/lite/micro/memory_planner/linear_memory_planner.cc -o edge-impulse-sdk/tensorflow/lite/micro/memory_planner/linear_memory_planner.o
g++ -Wall -g -DTF_LITE_DISABLE_X86_NEON=1 -DEI_CLASSIFIER_ENABLE_DETECTION_POSTPROCESS_OP=1 -Wno-strict-aliasing -I. -Os -DNDEBUG -g -std=c++14 -c edge-impulse-sdk/tensorflow/lite/core/api/error_reporter.cc -o edge-impulse-sdk/tensorflow/lite/core/api/error_reporter.o
g++ -Wall -g -DTF_LITE_DISABLE_X86_NEON=1 -DEI_CLASSIFIER_ENABLE_DETECTION_POSTPROCESS_OP=1 -Wno-strict-aliasing -I. -Os -DNDEBUG -g -std=c++14 -c edge-impulse-sdk/tensorflow/lite/core/api/flatbuffer_conversions.cc -o edge-impulse-sdk/tensorflow/lite/core/api/flatbuffer_conversions.o
g++ -Wall -g -DTF_LITE_DISABLE_X86_NEON=1 -DEI_CLASSIFIER_ENABLE_DETECTION_POSTPROCESS_OP=1 -Wno-strict-aliasing -I. -Os -DNDEBUG -g -std=c++14 -c edge-impulse-sdk/tensorflow/lite/core/api/op_resolver.cc -o edge-impulse-sdk/tensorflow/lite/core/api/op_resolver.o
g++ -Wall -g -DTF_LITE_DISABLE_X86_NEON=1 -DEI_CLASSIFIER_ENABLE_DETECTION_POSTPROCESS_OP=1 -Wno-strict-aliasing -I. -Os -DNDEBUG -g -std=c++14 -c edge-impulse-sdk/tensorflow/lite/core/api/tensor_utils.cc -o edge-impulse-sdk/tensorflow/lite/core/api/tensor_utils.o
client_loop: send disconnect: Connection reset

it hangs once it reaches the 2nd last line above^ until it got disconnected

Hello @weijunawj,

Can you try to edit the build.sh to use make -j2 instead of make -j.
I am suspecting an out of memory issue when you build with too many processes in parallel…

Regards,

Louis

1 Like

Thank u soooo much Louis!! U SAVED ME ONCE AGAIN :sob::sob::sob:

2 Likes

Hi Loise, i was facing similar issue as “weijunawj” getting on RPi. but after using make -j2, i am getting below error:
++: fatal error: Killed signal terminated program cc1plus
compilation terminated.

Please suggest.

Hi @sgp2002,

This sounds like it could be a memory or CPU issue. Could you try running make without -j2 and see if you still have the same issue? What type of computer are you trying to compile on? Is it an RPi? If so, what version?

Thanks Shwan, I tried “make” also, but RPi is getting crash.
I am trying on RPi 3b. Today i tried on Ubuntu laptop where 32GB RAM available, there also i am facing similar problem.
Actually, I am having existing keras model, that i am converting in ONNX and tried. Today I tried converting on TFlite, but facing similar issue.
Would you please help me out?

Hello @sgp2002,

Can you share your project ID so I can try to replicate with your model please?

Best,

Louis

Hi Louis, Its 202172.

Hello @sgp2002,

So I tried to compile both GitHub - edgeimpulse/example-standalone-inferencing-linux: Builds and runs an exported impulse locally (Linux) and GitHub - edgeimpulse/example-standalone-inferencing: Builds and runs an exported impulse locally (C++).

I downloaded your impulse, copied the content of the downloaded archive in the repository and ran:

  • For the standalone example:
make -j

For the standalone-linux:

APP_CUSTOM=1 make -j

Both worked but took a couple of minutes on my macbook pro.

Maybe your raspberry is not powerful enough to compile your model?
Can you make sure you followed this procedure:

I downloaded your impulse, copied the content of the downloaded archive in the repository and ran the commands:

Best,

Louis

Thanks louis, I tried compiling on my Linux laptop. It compiled properly. But getting hang on Raspberry Pi 3b. waited around 1hr, then it crashed.

  1. May I know, how much minimum RAM is required to compile? Raspberry pi 3b is having 1GB RAM only.
  2. I tried cross compiling on Linux laptop, created the output app, copied it to Raspberry Pi 3b and tried running it. But getting below error:
    pi@raspberrypi:~ $ ./app
    -bash: ./app: cannot execute binary file: Exec format error
    Que: Can it work like this? Or need to compile on Raspberry Pi 3b itself? Please suggest.

Hello @louis
I have tried all methods build with make, make -j, make -j2 , But facing error -
g++ -Wall -g -Wno-strict-aliasing -I. -Isource -Imodel-parameters -Itflite-model -Ithird_party -Iutils -Os -DNDEBUG -DEI_CLASSIFIER_ENABLE_DETECTION_POSTPROCESS_OP=1 -g -DTF_LITE_DISABLE_X86_NEON=1 -Ialsa-lib-1.2.7.2/include -std=c++14 -c edge-impulse-sdk/tensorflow/lite/micro/micro_allocator.cc -o edge-impulse-sdk/tensorflow/lite/micro/micro_allocator.o
g++: fatal error: Killed signal terminated program cc1plus
compilation terminated.
My machine is imx6ull and below it’s sepecifications
total used free shared buff/cache available
Mem: 232Mi 37Mi 128Mi 48Mi 66Mi 139Mi
Swap: 0B 0B 0B