train_dl_classifier_batch freeeze problem.

Evgeniy · 发表于 2021-5-21 23:22:58

Hi, colleagues!

I have next problem: Halcon 18.11 can't start operator train_dl_classifier_batch. Just freeze. And nothing happening. Just running and something do continuosly. Also this problem have in set_dl_classifier_param(DlHandle, 'runtime_init', 'immediately').

Please help

Evgeniy · 发表于 2021-5-21 23:25:00

This problem for all examples by deep_learning:

classify_pill_defects_deep_learning.hdev
and
classify_fruit_deep_learning.hdev

绝地武士 · 发表于 2021-5-22 10:35:56

How is your computer configuration?which is your graphics card used?

Evgeniy · 发表于 2021-5-24 19:31:42

Intel(R) Xeon(R) Silver 4114 CPU 2.2GHz / 64 Gb / NVidia GeForce RTX3070

Evgeniy · 发表于 2021-5-24 19:41:10

Halcon 18.11

Evgeniy · 发表于 2021-5-28 14:02:52

As it turned out, the problem was related to the video card and the old version of Halcon.

Hangs disappeared after reinstalling Halcon from version 18.11 to version 20.11.

Apparently the old version of Halcon is not able to work correctly with very new video cards.

I was checked this rigth now.

Criss · 发表于 2021-5-28 14:23:55

The solution is usually to upgrade the latest halcon version and update the latest nvidia graphics driver