Recent content by Molbo

You need to configure different subnets to allow Wireguard to route network traffic to every location. The subnets going to be isolated anyway. Your IP allocation example is correct.

That should work, just make sure you are using different subnets in every location and cameras are using static ips. WireGuard is really nice software and easy to use, especially if you are not afraid to use CLI and edit config files manually :)

Right, but it's not clear to how to add your own images to the model, it looks like the only option is to use existent Open Images Dataset.

Is it possible to fine tune a custom model by adding some number of my own images? Somewhat similar to LLM fine-tuning? And slightly related to that, imagine if BI could help us to produce new set of training images, defining a bounding box and allowing us to label it :)

Oh, huh, how did I miss that one, thank you wittaj.

Well, I'm tracking deer intrusions into 20 acres partially forested property with pretty complex topography. I cannot cover it densely enough to use low resolution motion detection. So, I want to have relatively few hires wide angle cameras able to spot a deer from far away. And, tbh, I’m not...

After watching my son assembling his new gaming rig, I'm thinking about building something custom for BI. Assuming the budget is not a big problem, what would be the best hardware for BI? I think I know what would be a good GPU for CPAI, but what about BI itself? The goal is to handle as many...

No progress, I don't think CPAI can use neural engines atm.

It's hosted publically on Github and has Server Side Public License, so anyone can fork it and continue to develop it even if the current team stops updating it. Here is the list of all the current forks, btw Forks · codeproject/CodeProject.AI-Server

I'm running BI on pretty small windows machine, that doesn't have room for a good videocard. I want to start playing with CodeProject.AI and thinking about buying new Windows box with CUDA capable card or low end Mac mini M2 to run CP.AI on it with BI connecting to it remotely. What would be...

I tried to configure IVS on IPC-Color4K-T180 (V3.120.0000000.2.R, Build Date: 2022-10-21) using MacOS Chrome/FF/Safari and it's not working, failing after line drawing. I tested it on Windows and everything is working fine there, so it's definitely something MacOS specific :( I suspect it...

https://tailscale.com/

Hmm, should I test if it can detect a human crawling, for example? :) In general, I think there should be multiple way to deceive camera AI.

I just noticed that my T5442T-ZE that is configured to use IVS and SMD with object filtering is completely ignoring tracked construction machines, such as excavator and front loader. It detects humans and wheeled vehicles just fine. I think that could be an interesting way to deceive AI...

I'm using Chrome on MacOS all the time to connect to my Dahua cameras. It works without any problems. No plugins. I just tested Firefox and it also works ok.