ai-vision is a more advanced solution and with a few VS basics relatively easy to integrate into BI. But it is a relatively expensive solution if you do not use it for specific tasks (e.g. occupational safety).
https://azure.microsoft.com/en-us/products/ai-services/ai-vision
not directly what you ask since they include a complete Android OS, but they are great to run smarthome dashboards or nvr apps:
liyama ProLite TW2424AS-B1
Allnet ALLSD24RK3588A12B4G
This is a quite common issue. The usb version is intended for developers, not for 24/7 operation. You can read about various sources of interference in relevant forums like homeassitent or frigate (usb power supply, interfering wlan/bt modules, bad cables,...) - maybe there is something simple...
since an 8tb datacenter nvme starts at 400$ (eg. Kingston DC1500M or Samsung PM9A3), i would not waste money on spinning rust. Put a u.2 pcie card for 20$ on the list and install everything on a single device.
viewing/fastforward is MUCH faster with nvme storage.
apart from the license costs, there is no big difference it you install the "desktop experience" on windows server => Server Core vs Server with Desktop Experience install options
A setup without the desktop experience is possible, but i would only recommend it in a bigger environment with...
Your contractor has to figure out options. There are enough easier jobs, so every craftsman will currently prefer these or set a high quote. And yes, like in germany they make good money because "everybody" want´s something combined with a huge labor shortage.
And i would not cheap out on the...
You should tell that this is only the cheap part. He´ll need to spend quitea lot to be on code with wind saftey and fire protection regulation like for the DC cables and battery storage. And he is from florida where you may need to prove that it can withstand +150mph storms and many roofs where...
We use a simple battery powerd GSM Modul like the GA01P GSM what is powerd by the same source as our servers/network. In case of an power outage or ups failure, it is able to call you or send an message via sms...
did you change the cpu profile to high performance? The vmware default for a whitelabel box will be "balanced" => you cpu would not boost up to full speed. More cores won´t help a lot in this case since you have only a single camera.
In the end the xeon v3 is a ~10 year old plattform and does not boost very high to todays standards. Also a small number of old sas disks with raid 6 isn´t the best choice for lot´s of small writes. But it should be able to handle a single camera in a vm setup for sure. What exactly does not...
if you simple want to try it out or do some benchmarks, there are multiple images out there with Tensorflow and a BI compatible API => https://raw.githubusercontent.com/runningman84/docker-coral-rest-server/master/Dockerfile
Check the SSD Firmware, the Evo 840 series ist about 10 years old and the controller was quite buggy:
https://pcper.com/2016/06/samsung-magician-4-9-7-and-samsung-840-dxt0ab0q-firmware-tested-read-speed-issues-finally-addressed/
Glad to see that they go multi-engine as well.
@Chicken: it won´t matter as long you run it on the same host without any layer in between (like docker)
I would recommend to use a CDN in front of your cameras (cdn77k, vbrick, cloudflare,..). Then you only need the bandwidth for a single stream and do not need to expose your cameras to the public.
The driver provides access on the dedicated TPU chips. That means it is no problem to use 2 or more USB TPU´s or one or more of the dual TPU version. There are also some quite expensive m.2 e-key pcie adapter with an active cooling solution on the market to handle up to 16 TPU chips (e.g. from...
mouser.com was able to deliver my orders from november (dual TPU Version). I would recommend the usb version only for testing/development since it will add some latency and flaws on high load.
as far i can tell codeproject does only support pytorch => so no coral support for now?
https://pytorch.org/blog/compromised-nightly-dependency/#what-we-know
as far i can tell the actual windows builds from codeproject do not include the compromised package. If you have installed/updated a local linux/docker pyTorch instance in the last december week, i would recommend to have a look.
if you want more easy disk access => HPE ML30 Gen10 Plus (4x LFF Hotswap bays).
In principle, we always recommend servers with appropriate maintenance contracts - then a hardware defect is significantly less stress. Without this, I would even prefer whitelabel boxes because you can install...
BI asked my in my last deployment if it should install Codeproject AI, i guess they already build in a "1-click" installer?
There are multiple Tensorflow based projects with coral support, they provide the same api as deepstack and can be used in BI without a problem. But all of them need quite...
Since there are already multiple API wrapper for Tensorflow, it would be also interessting how many users already moved to solutions with hardware support for Coral or non-CUDA solutions?
sorry, but what do you want? I would understand if you would tell me that an R350 can´t handle a GPU, because i was meaning the T350.
1. if he takes the dell model and configure it, he´ll see.He asked for a direction, not for a sales quote.
2. where did i recommend 16gb for Storage? ;) I...
He "literally" asked for a server and specified a budget.
The CPU is defined with the server model and it does not matter if you choose it with 2.4 or 2.8ghz.
1. stop the app
2. open a cmd window and enter "netstat -aon | findstr :80" => check if anything is listening on port 80 => notice the "pid" on the right
3. enter the follow in the cmd window: tasklist /svc /FI "PID eq [your PID]"
i used the port 81 in my example, so just change it to 80.