Closed danlakey closed 5 years ago
It looks to me that in the VM some CPU instructions are not supported, facebox uses AVX2 instructions, can you verify some way that VM has that instructions available?
You can maybe verify that is the instruction set, using the docker image machinebox/facebox_noavx
(not officially supported, but available for the home assistant project)
Excellent bug report. :)
@dahernan thank you for the suggestion of machinebox/facebox_noavx
, that appears to have resolved my issue! I executed docker pull machinebox/facebox_noavx
and the launched the container via docker run -p 8080:8080 -e "MB_KEY=$MB_KEY" machinebox/facebox_noavx
and was able to continue on with the demo successfully.
I actually came across MachineBox from the Home Assistant Community and am looking to implement Facebox as a potential solution for in-house presence detection (who is in certain rooms of the house).
Will this version machinebox/facebox_noavx
ever become officially supported?
Thank you again for the quick reply.
Great, glad to hear that. We don't have plans right to support noavx, we currently focus on cloud build, and still are small team
Any solution for objectbox? Naive pull of machinebox/objectbox_noavx:latest did not work :( .
I'm trying to run facebox on bare metal (no VM) on a Intel Core 2 Quad Q6600 and ran into this issue. noavx got me a step further, but now even though facebox says it's ready, any request to train responds "box is not ready, try again later".
Encountering an error with the built in Facebox demo. Below are my system specs: Dell PowerEdge T310 running VMWare ESXi 5.5 One VM with Ubuntu 18.04 LTS with 2 CPU and 4 GB RAM Docker version 18.09.3, build 774a1f4
Attached a full dump of the console when the demo page button is clicked. First few lines of the print out are:
facebox crash log.txt
Thank you for your help, looking forward to being able to use this product.