hello all, i tried to run neon on docker environment, when I run cifar10 example, it will crash by "model.fit(train, optimizer=opt_gdm, num_epochs=num_epochs,
cost=cost, callbacks=callbacks)"
The execute command is:
root@6e86070d4014:/tmp# python cifar10_lenet.py -b mkl
Illegal instruction (core dumped)
I have downloaded that core dump, but w/o symbol file, it will show question marks.
any idea how to check that core dump?
one more clue: crash system used Intel(R) Xeon(R) CPU E5-2650 v2 @ 2.60GHz, but for Intel(R) Xeon(R) Gold 6248 CPU @ 2.50GHz, it works fine.
mkl version is: mklml_lnx_2018.0.1.20171227
hello all, i tried to run neon on docker environment, when I run cifar10 example, it will crash by "model.fit(train, optimizer=opt_gdm, num_epochs=num_epochs, cost=cost, callbacks=callbacks)" The execute command is: root@6e86070d4014:/tmp# python cifar10_lenet.py -b mkl Illegal instruction (core dumped)
I have downloaded that core dump, but w/o symbol file, it will show question marks. any idea how to check that core dump?
neon version is: 2.6.0+f9d771b
thanks, Eugene