the correct data as follow:
543.429 76.3822 594.336 210.146 0.867586 0
378.291 79.1234 414.697 174.536 0.802472 0
53.5174 37.2288 163.915 353.05 0.794809 0
134.581 119.821 266.834 264.356 0.749265 0
271.192 84.3156 299.226 182.198 0.731514 0
346.405 79.6145 380.143 180.17 0.711018 0
206.175 73.101 264.53 200.231 0.665582 0
309.502 87.152 341.577 160.295 0.637177 0
325.73 144.886 399.379 254.945 0.615658 0
290.677 84.1234 323.912 177.064 0.577647 0
100.565 40.8382 160.28 230.896 0.518276 0
363.7 136.358 424.807 224.517 0.508566 0
602.741 122.713 627.097 142.253 0.448372 0
123.509 222.057 267.914 317.438 0.493788 56
369.073 197.882 550.159 311.72 0.575214 60
488.226 86.3548 563.139 132.268 0.820506 62
after quantization on borad :
543.923 75.7296 593.641 212.502 0.847142 0
44.6422 42.1153 163.474 346.575 0.805613 0
377.14 79.4789 415.256 175.19 0.777702 0
203.153 64.9407 264.533 201.713 0.720819 0
346.062 78.2363 382.334 185.349 0.623575 0
307.857 87.8984 341.986 156.734 0.567722 0
272.842 143.625 400.895 261.567 0.516111 0
368.079 135.083 427.135 220.16 0.43871 0
271.554 92.449 298.01 176.125 0.427736 0
94.9267 45.8206 159.414 215.314 0.343526 0
601.751 122.242 627.2 142.945 0.338131 0
271.449 89.378 298.452 179.476 0.362646 24
93.1184 45.8421 160.219 215.292 0.326112 26
126.366 218.453 266.783 319.255 0.380229 56
250.269 145.704 396.101 273.539 0.229514 57
364.608 201.969 552.584 323.696 0.305701 60
We can find that after quantification, the score becomes smaller
But when i run the command as follow and debug the outpus on PC,the data seems right!
tensorzone=${ACUITY_PATH}tensorzonex
$tensorzone
–action inference
–source text
–source-file ./verify.list
–channel-mean-value ‘0 0 0 255’
–model-input …/${NAME}.json
–model-data …/${NAME}.data
–dtype quantized"
debug outpus:
img_boxes.size:16
543.647 77.8307 596.121 206.863 0.858643 0
52.5562 38.7353 162.118 351.888 0.808135 0
377.724 79.2998 415.408 172.841 0.784823 0
133.909 122.218 269.756 266.9 0.746892 0
347.793 80.1796 379.332 177.878 0.703232 0
204.673 71.8114 266.002 200.844 0.683806 0
271.285 84.5801 299.351 181.252 0.683376 0
309.281 87.0134 341.304 160.112 0.616298 0
325.253 142.976 399.364 255.371 0.581223 0
103.835 42.1638 158.988 225.174 0.534385 0
223.586 87.7342 263.99 183.365 0.528073 0
290.586 84.3736 324.237 176.945 0.516509 0
359.012 136.198 424.088 227.892 0.50728 0
602.133 121.871 627.561 142.556 0.4664 0
366.987 201.873 551.688 310.23 0.560215 60
488.153 84.9329 563.66 132.247 0.836944 62