@johndoe Reload Galcore without adding any settings.
On the terminal where you execute this command, print your environment variables
@johndoe Reload Galcore without adding any settings.
On the terminal where you execute this command, print your environment variables
export VIVANTE_SDK_DIR=/home/khadas/Just_for_get_op_time/data/vcmdtools
export LD_LIBRARY_PATH=/home/khadas/Just_for_get_op_time/data/drivers_64_exportdata
export VIV_VX_DEBUG_LEVEL=1
export VIV_VX_PROFILE=1
export CNN_PERF=0
export NN_LAYER_DUMP=0
@johndoe Can you restart the board and try again without reloading Galcore? I think the environment variable is polluted and it can’t get out.
I restarted the board. Didn’t reload galcore. Added the above environment variables but still not getting the execution times
( 0, 0, 0) ( 0, 0, 0.000000%, 0.000000%, NONE)
PreLoadWeightBiases = 66304 6.323242%
---------------------------End VerifyTiling -------------------------
Verify Graph: 4509ms or 4509410us
Start run graph [1] times...
[ 1] TOTAL_READ_BANDWIDTH (MByte): 6.408952
[ 2] TOTAL_WRITE_BANDWIDTH (MByte): 3.544906
[ 3] AXI_READ_BANDWIDTH (MByte): 4.967385
[ 4] AXI_WRITE_BANDWIDTH (MByte): 3.543929
[ 5] DDR_READ_BANDWIDTH (MByte): 1.441568
[ 6] DDR_WRITE_BANDWIDTH (MByte): 0.000977
[ 7] GPUTOTALCYCLES: 2045773
[ 8] GPUIDLECYCLES: 580846
VPC_ELAPSETIME: 2830
*********
Run the 1 time: 3.00ms or 3237.00us
@johndoe Package your original file and conversion script to me, I will find time to test next week
This are my folder contents link
Conversion script:
../bin/convertonnx \
--onnx-model squeezenet1_1.onnx \
--net-output squeezenet1_1.json \
--data-output squeezenet1_1.data
../../bin/tensorzonex \
--action quantization \
--source text \
--source-file data/validation_tf.txt \
--channel-mean-value '103.94 116.78 123.68 58.82' \
--reorder-channel '2 1 0' \
--model-input squeezenet1_1.json \
--model-data squeezenet1_1.data \
--quantized-dtype asymmetric_affine-u8 \
--quantized-rebuild
../bin/ovxgenerator \
--model-input squeezenet1_1.json \
--data-input squeezenet1_1.data \
--model-quantize squeezenet1_1.quantize \
--reorder-channel '2 1 0' \
--channel-mean-value '103.94 116.78 123.68 58.82' \
--export-dtype quantized
mkdir -p squeezenet_case_code
mv *.h *.c .project .cproject *.vcxproj BUILD *.linux *.export.data squeezenet_case_code
rm *.data *.quantize *.json
according to step 3. i can not understanding what should be the exact contect of 2_export_case_code.sh after the editing.
should it be follows?
#!/bin/bash
NAME=mobilenet_tf
ACUITY_PATH=../bin/
export_ovxlib=${ACUITY_PATH}ovxgenerator
$export_ovxlib \
--model-input ${NAME}.json \
--data-input ${NAME}.data \
--model-quantize ${NAME}.quantize \
--reorder-channel '0 1 2' \
--channel-mean-value '128 128 128 128' \
--export-dtype quantized \
# --optimize VIPNANOQI_PID0XE8 \
# --viv-sdk ${ACUITY_PATH}vcmdtools \
# --pack-nbg-unify \
#Note:
# --optimize VIPNANOQI_PID0XB9
# when exporting nbg case for different platforms, the paramsters are different.
# you can set VIPNANOQI_PID0X7D VIPNANOQI_PID0X88 VIPNANOQI_PID0X99
# VIPNANOQI_PID0XA1 VIPNANOQI_PID0XB9 VIPNANOQI_PID0XBE VIPNANOQI_PID0XE8
# Refer to sectoin 3.4(Step 3) of the <Model_Transcoding and Running User Guide_V0.8> documdent
rm -rf nbg_unify_${NAME}
mv ../*_nbg_unify nbg_unify_${NAME}
cd nbg_unify_${NAME}
mv network_binary.nb ${NAME}.nb
cd ..
#save normal case demo export.data
mkdir -p ${NAME}_normal_case_demo
mv *.h *.c .project .cproject *.vcxproj BUILD *.linux *.export.data ${NAME}_normal_case_demo
# delete normal_case demo source
#rm *.h *.c .project .cproject *.vcxproj BUILD *.linux *.export.data
rm *.data *.quantize *.json
or should it be as below?
#!/bin/bash
NAME=mobilenet_tf
ACUITY_PATH=../bin/
export_ovxlib=${ACUITY_PATH}ovxgenerator
$export_ovxlib \
# --model-input ${NAME}.json \
# --data-input ${NAME}.data \
# --model-quantize ${NAME}.quantize \
--reorder-channel '0 1 2' \
--channel-mean-value '128 128 128 128' \
--export-dtype quantized \
# --optimize VIPNANOQI_PID0XE8 \
# --viv-sdk ${ACUITY_PATH}vcmdtools \
# --pack-nbg-unify \
#Note:
# --optimize VIPNANOQI_PID0XB9
# when exporting nbg case for different platforms, the paramsters are different.
# you can set VIPNANOQI_PID0X7D VIPNANOQI_PID0X88 VIPNANOQI_PID0X99
# VIPNANOQI_PID0XA1 VIPNANOQI_PID0XB9 VIPNANOQI_PID0XBE VIPNANOQI_PID0XE8
# Refer to sectoin 3.4(Step 3) of the <Model_Transcoding and Running User Guide_V0.8> documdent
rm -rf nbg_unify_${NAME}
mv ../*_nbg_unify nbg_unify_${NAME}
cd nbg_unify_${NAME}
mv network_binary.nb ${NAME}.nb
cd ..
#save normal case demo export.data
mkdir -p ${NAME}_normal_case_demo
mv *.h *.c .project .cproject *.vcxproj BUILD *.linux *.export.data ${NAME}_normal_case_demo
# delete normal_case demo source
#rm *.h *.c .project .cproject *.vcxproj BUILD *.linux *.export.data
rm *.data *.quantize *.json
Can you please share the exact contents after the required editing of 2_export_case_code.sh
@Frank @johndoe