Downloads · 30 days
0
Yolov7_Visual_Pollution_Detection
Aalaa/Yolov7_Visual_Pollution_Detection
Yolov7_Visual_Pollution_Detection is a object detection model from Aalaa. Use it when you need objects located in an image.
The dataset was referenced in the Smartathon competition.It's consist of 7874 images annontated with 11 classes:
Access
Public
Updated Jan 22, 2023
Repo size
74.9 MB
Likes
3
Public
Repo makeup
Click a slice to open those files.
.pt74.9 MB · 100%
At a glance
- Task
- Object Detection
- Access
- Public
- Created
- Jan 22, 2023
- Updated
- Jan 22, 2023
- SHA
- c59154f4
Model card
From the Hugging Face model README
Dataset
The dataset was referenced in the Smartathon competition.It's consist of 7874 images annontated with 11 classes:
- GARBAGE 8597
- CONSTRUCTION_ROAD 2730
- POTHOLES 2625
- CLUTTER_SIDEWALK 2253
- BAD_BILLBOARD 1555
- GRAFFITI 1124
- SAND_ON_ROAD 748
- UNKEPT_FACADE 127
- FADED_SIGNAGE 107
- BROKEN_SIGNAGE 83
- BAD_STREETLIGHT 1
The dataset highly imbalanced and contain some humman errors.
Our SEE Team Solution
- Convert from Pascal VOC to YOLO format
- Model Hyperparamter tuning
- Train the data on Yolov7
- Evaluate the model
- Expalin Different techniques to Automation of Data Annotation
For our solution detials: notebook
How to use
- You can just download file weights from the files section
- clone yolov7 repo
!git clone https://github.com/WongKinYiu/yolov7 - ensure your current working directory is yolov7 then run
! pip install -r requirements.txt - then run the detector script
! python detect.py --weights " model.pt path" --img 736 --conf 0.27 --source "testing image path" --save-txt
- Task
- Object Detection
- Created
- Jan 22, 2023
- Updated
- Jan 22, 2023