|
|
--- |
|
|
title: Yolo V3 |
|
|
emoji: 📈 |
|
|
colorFrom: purple |
|
|
colorTo: yellow |
|
|
sdk: gradio |
|
|
sdk_version: 4.38.1 |
|
|
app_file: app.py |
|
|
pinned: false |
|
|
license: mit |
|
|
--- |
|
|
|
|
|
|
|
|
|
|
|
# Yolov3 Trained on Pascal VOC |
|
|
|
|
|
- Model is trained on [PASCAL VOC](https://www.kaggle.com/datasets/aladdinpersson/pascal-voc-dataset-used-in-yolov3-video) dataset |
|
|
- The model is trained for 40 epochs using one cycle policy |
|
|
- Model has used Mosaic technique to speed up the training part |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
## Links |
|
|
|
|
|
- Training Repository Link: https://github.com/Shilpaj1994/ERA/tree/master/Session13 |
|
|
|
|
|
- Model File Link: https://huggingface.co/spaces/Shilpaj/yolo_v3/resolve/main/epoch%3D39-step%3D41400.ckpt |
|
|
|
|
|
- Application Link: https://huggingface.co/spaces/Shilpaj/yolo_v3 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
## Usage |
|
|
|
|
|
- Select the IOU threshold which internally modifies the non-max suppression values to display the bounding box output |
|
|
- Select the threshold value of the class. Objects with class confidence above the threshold value will be displayed with bounding box |
|
|
- Select if GradCam output is to be displayed or not |
|
|
- Select the opacity of the GradCam output over the input image |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
 |
