[HW Accel Support]: #18756
Replies: 5 comments 17 replies
-
To be clear this is CPU usage not TPU usage, the inference speed is slow which suggests there may be some inefficiencies leading to slowdown. Is this connected via USB3? Can you try a different full speed USB cable? |
Beta Was this translation helpful? Give feedback.
-
Thanks. It's connected via USB3. I can try a different cable but the same cable worked fine on the other machine
…On Tue, Jun 17, 2025, 11:16 AM Nicolas Mowen ***@***.***> wrote:
To be clear this is CPU usage not TPU usage, the inference speed is slow
which suggests there may be some inefficiencies leading to slowdown. Is
this connected via USB3? Can you try a different full speed USB cable?
—
Reply to this email directly, view it on GitHub
<#18756 (comment)>,
or unsubscribe
<https://github.com/notifications/unsubscribe-auth/AA7LMIQM22AB5AAFRMJ75IL3D5T53AVCNFSM6AAAAAB7OVZALOVHI2DSMVQWIX3LMV43URDJONRXK43TNFXW4Q3PNVWWK3TUHMYTGNBZGAYTAOI>
.
You are receiving this because you authored the thread.Message ID:
***@***.***
com>
|
Beta Was this translation helpful? Give feedback.
-
ok discovered the issue - the gpu was passed thgrough to a different vm. i have moved frigate to the vm with the gpu and update my config as follows. Frigate starts but i'm seeing this in the logs logs
config
compose:
any advice on what i'm doing wrong would be greatly appreciated. Note plex is on the same vm and can accsess the gpu fine I'm keen to still use my TPU for main detections to reduce load on gpu for other tasks if possible thank you |
Beta Was this translation helpful? Give feedback.
-
Beta Was this translation helpful? Give feedback.
-
ah i see it coudl be cause i'm not using ghcr.io/blakeblackshear/frigate:stable-tensorrt - trying this now edit: no luck its using 78% instead of 99% tho but the ui is very unresponssive Logs now:
|
Beta Was this translation helpful? Give feedback.
Uh oh!
There was an error while loading. Please reload this page.
Uh oh!
There was an error while loading. Please reload this page.
-
Describe the problem you are having
hello - i have had my system running well for a while. I am running it now on proxmox vm on a very powerful server but i am wanting to use my tpu still (I do have a 5090 i can use tho)
I have the tpu passed through fine and frigate starts fine but the tpu cpu usage is 99% constantly. I have an amd threadripper cpu (4 cores assigned so far)
I am not using the 5090 as I plan to use it to run some large llms. but I guess i could move over to it .
appreciate any help thank you
Version
0.15.0-6cb5cfb
Frigate config file
docker-compose file or Docker CLI command
Relevant Frigate log output
Relevant go2rtc log output
FFprobe output from your camera
Install method
Proxmox via Docker
Object Detector
Coral
Network connection
Wired
Camera make and model
unifi and duahua
Screenshots of the Frigate UI's System metrics pages
Any other information that may be helpful
No response
Beta Was this translation helpful? Give feedback.
All reactions