Mirrors/blakeblackshear.frigate
1
0
Fork 0
mirror of https://github.com/blakeblackshear/frigate.git synced 2025-08-13 13:47:36 +02:00
Code Issues Projects Releases Wiki Activity

Revamp RF-DETR Export Docs (#19341)

Browse Source 
* Revamp RFDETR docs

* Clarify model size

* Specify model size
This commit is contained in:
Nicolas Mowen 2025-07-31 08:11:46 -06:00 committed by GitHub
parent d18f2282c8
commit c3410cd13e
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
2 changed files with 24 additions and 25 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

37
docs/docs/configuration/object_detectors.md
View File

@ -365,8 +365,8 @@ detectors:
model: model:
model_type: rfdetr model_type: rfdetr
width: 560 width: 320
height: 560 height: 320
input_tensor: nchw input_tensor: nchw
input_dtype: float input_dtype: float
path: /config/model_cache/rfdetr.onnx path: /config/model_cache/rfdetr.onnx
@ -616,8 +616,8 @@ detectors:
model: model:
model_type: rfdetr model_type: rfdetr
width: 560 width: 320
height: 560 height: 320
input_tensor: nchw input_tensor: nchw
input_dtype: float input_dtype: float
path: /config/model_cache/rfdetr.onnx path: /config/model_cache/rfdetr.onnx
@ -983,22 +983,21 @@ Make sure you change the batch size to 1 before exporting.
### Download RF-DETR Model ### Download RF-DETR Model
To export as ONNX: RF-DETR can be exported as ONNX by running the command below. You can copy and paste the whole thing to your terminal and execute, altering `MODEL_SIZE=Nano` in the first line to `Nano`, `Small`, or `Medium` size.
1. `pip3 install rfdetr` ```sh
2. `python3` docker build . --build-arg MODEL_SIZE=Nano --output . -f- <<'EOF'
3. `from rfdetr import RFDETRBase` FROM python:3.11 AS build
4. `x = RFDETRBase()` RUN apt-get update && apt-get install --no-install-recommends -y libgl1 && rm -rf /var/lib/apt/lists/*
5. `x.export()` COPY --from=ghcr.io/astral-sh/uv:0.8.0 /uv /bin/
WORKDIR /rfdetr
#### Additional Configuration RUN uv pip install --system rfdetr onnx onnxruntime onnxsim onnx-graphsurgeon
ARG MODEL_SIZE
The input tensor resolution can be customized: RUN python3 -c "from rfdetr import RFDETR${MODEL_SIZE}; x = RFDETR${MODEL_SIZE}(resolution=320); x.export()"
FROM scratch
```python ARG MODEL_SIZE
from rfdetr import RFDETRBase COPY --from=build /rfdetr/output/inference_model.onnx /rfdetr-${MODEL_SIZE}.onnx
x = RFDETRBase(resolution=560) # resolution must be a multiple of 56 EOF
x.export()
``` ```
### Downloading YOLO-NAS Model ### Downloading YOLO-NAS Model

12
docs/docs/frigate/hardware.md
View File

@ -166,12 +166,12 @@ There are improved capabilities in newer GPU architectures that TensorRT can ben
Inference speeds will vary greatly depending on the GPU and the model used. Inference speeds will vary greatly depending on the GPU and the model used.
`tiny` variants are faster than the equivalent non-tiny model, some known examples are below: `tiny` variants are faster than the equivalent non-tiny model, some known examples are below:
| Name | YOLOv9 Inference Time | YOLO-NAS Inference Time | RF-DETR Inference Time | | Name | YOLOv9 Inference Time | YOLO-NAS Inference Time | RF-DETR Inference Time |
| --------------- | --------------------- | ------------------------- | ------------------------- | | --------------- | --------------------- | ------------------------- | ---------------------- |
| RTX 3050 | 320: 15 ms | 320: ~ 10 ms 640: ~ 16 ms | 336: ~ 16 ms 560: ~ 40 ms | | RTX 3050 | t-320: 15 ms | 320: ~ 10 ms 640: ~ 16 ms | Nano-320: ~ 12 ms |
| RTX 3070 | 320: 11 ms | 320: ~ 8 ms 640: ~ 14 ms | 336: ~ 14 ms 560: ~ 36 ms | | RTX 3070 | t-320: 11 ms | 320: ~ 8 ms 640: ~ 14 ms | Nano-320: ~ 9 ms |
| RTX A4000 | | 320: ~ 15 ms | | | RTX A4000 | | 320: ~ 15 ms | |
| Tesla P40 | | 320: ~ 105 ms | | | Tesla P40 | | 320: ~ 105 ms | |
### ROCm - AMD GPU ### ROCm - AMD GPU