Multiple GPUs with ONNX and 0.15-beta3 #15591
Unanswered
jvallery
asked this question in
Ask A Question
Replies: 2 comments
-
You need to specify the device as a string not an integer |
Beta Was this translation helpful? Give feedback.
0 replies
-
Yes! That worked. Thanks. Not obvious from the docs. The tensorrt example uses an integer. |
Beta Was this translation helpful? Give feedback.
0 replies
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
-
I'm trying to configure 0.15-beta3 to work with multiple GPUs and the onnx detector. I've tried this a few ways, but had no luck so far.
If I specify multiple detectors without a specific device ID, 2 instances are created but they are both running on gpu 0. If I specify the "device" property in the config with the integer associated with the GPU, the config fails to load as invalid.
How do I specify which GPU each detector should run on?
detectors: onnx_0: type: onnx onnx_1: type: onnx
51732 root 0 Compute 13% 480MiB 4% 26% 863MiB frigate.detector.onnx_1
51715 root 0 Compute 25% 404MiB 3% 28% 860MiB frigate.detector.onnx_0
`+---------------------------------------------------------------------------------------+
| NVIDIA-SMI 535.183.01 Driver Version: 535.183.01 CUDA Version: 12.2 |
|-----------------------------------------+----------------------+----------------------+
| GPU Name Persistence-M | Bus-Id Disp.A | Volatile Uncorr. ECC |
| Fan Temp Perf Pwr:Usage/Cap | Memory-Usage | GPU-Util Compute M. |
| | | MIG M. |
|=========================================+======================+======================|
| 0 NVIDIA GeForce RTX 3060 Off | 00000000:21:00.0 Off | N/A |
| 58% 61C P2 89W / 170W | 9454MiB / 12288MiB | 91% Default |
| | | N/A |
+-----------------------------------------+----------------------+----------------------+
| 1 NVIDIA GeForce RTX 3060 Off | 00000000:2B:00.0 Off | N/A |
| 0% 49C P8 18W / 170W | 3MiB / 12288MiB | 0% Default |
| | | N/A |
+-----------------------------------------+----------------------+----------------------+
+---------------------------------------------------------------------------------------+
| Processes: |
| GPU GI CI PID Type Process name GPU Memory |
| ID ID Usage |
|=======================================================================================|
+---------------------------------------------------------------------------------------+`
Beta Was this translation helpful? Give feedback.
All reactions