docs: update OpenVINO D-FINE configuration default device (#21231)

* docs: remove OpenVINO D-FINE configuration device

* docs: change D-FINE model detectors default device
This commit is contained in:
GuoQing Liu 2025-12-11 21:31:52 +08:00 committed by GitHub
parent 1de7519d1a
commit 0a293aebab
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194

View File

@ -395,7 +395,7 @@ After placing the downloaded onnx model in your config/model_cache folder, you c
detectors:
ov:
type: openvino
device: GPU
device: CPU
model:
model_type: dfine
@ -431,10 +431,10 @@ When using Docker Compose:
```yaml
services:
frigate:
---
devices:
- /dev/dri
- /dev/kfd
...
devices:
- /dev/dri
- /dev/kfd
```
For reference on recommended settings see [running ROCm/pytorch in Docker](https://rocm.docs.amd.com/projects/install-on-linux/en/develop/how-to/3rd-party/pytorch-install.html#using-docker-with-pytorch-pre-installed).
@ -462,9 +462,9 @@ When using Docker Compose:
```yaml
services:
frigate:
environment:
HSA_OVERRIDE_GFX_VERSION: "10.0.0"
...
environment:
HSA_OVERRIDE_GFX_VERSION: "10.0.0"
```
Figuring out what version you need can be complicated as you can't tell the chipset name and driver from the AMD brand name.