No response to vision detection requests on new server

I have DeepStack running on a LUBUNTU vm (under hypervisor) and it works fine. I wanted to use the GPU version with a spare 1080 I have so I built out a new LUBUNTU install (tried plain UBUNTU First with same results) on a bare metal HP Z800 workstation. Every time I can connect to the landing page and see the DeepStack is activated message (curious as to how its active since I have not registered on this machine at all) , but I can not use any of the endpoints. Any calls to an end point just timeout, and I do not see anything in the Docker logs at all. I have disabled the firewall, and I have tried with both the CPU and GPU versions of the image. I have rebuilt the OS a couple of times with the same results. I have tried with just the VISION-DETECTION end point, as well as the VISION-DETECTION and VISION-FACE endpoints (as I was using both of these on the vm installation)
The machine has 2 nics, I don’t know if that could be causing any issues. I am pretty green when it comes to Docker and Linux in general so I am at the extent of my troubleshooting ability without some direction. Any help is greatly appreciated.

John

EDIT:
I was able to use the -noavx version, so I tried the gpu-noavx version which apparently requires a paid key. Ok - I PAID for a 3 month key and it also doesnt work. I get nothing -just times out when trying to use it.
So the standard deepquestai/deepstack container doesnt work. The deepquestai/deepstack:gpu doesnt work , the deepquestai/deepstack:noavx DOES work but the deepquestai/deepstack:gpu-noavx (paid) does not work. This is running on an older xeon machine.

Your CPU needs to support the AVX instructions… most pre gen2 Intels and Xeons do not have these. The Image detection hangs as the calculation cannot be performed.

The NoAVX version is probably an old image with an old version of tensorflow (pre 2.0) binaries were build without AVX instructions. The subscription model seems to be abandoned but the old images remain

1 Like

Hello @JohnH

Kindly use the latest update deepquestai/deepstack:latest or deepquestai/deepstack:cpu-x4-beta as it adds full support for both avx and noavx systems and it is entirely free. We are open sourcing deepstack soon.