|
--- |
|
license: apache-2.0 |
|
tags: |
|
- darknet |
|
- yolo |
|
pipeline_tag: object-detection |
|
--- |
|
# yolo-lightnet |
|
|
|
<!-- Provide a quick summary of what the model is/does. --> |
|
|
|
This is an optimized YOLO model. It is optimized for running on NVDLA. |
|
|
|
*NOTE: This is darknet format **NOT** Pytorch.* |
|
|
|
## Model Details |
|
|
|
<!-- Provide a longer summary of what this model is. --> |
|
|
|
File name is formatted like `lightnet-{name}-{resolution}.weights`. |
|
|
|
|
|
### driving |
|
|
|
- target: car, bus, person, bike, truck, motor, train, rider, traffic_sign, traffic_light |
|
- training data: BDD100K |
|
|
|
### face |
|
|
|
- target: face |
|
- train data: wider face |
|
|
|
### head_body |
|
|
|
- target: head, body(include hidden area) |
|
- train data: crowdhuman |
|
|
|
### head_body-visible |
|
|
|
- target: head, body(only visible area) |
|
- train data: crowdhuman |
|
- cfg file and label names are same as `head body` |
|
|
|
## Uses |
|
|
|
Running on: |
|
https://github.com/daniel89710/lightNet-TRT |
|
https://github.com/AlexeyAB/darknet |
|
|
|
|
|
|
