Top AI Repos — open-source AI, indexed and scored
Top AI Repos tracks AI repositories on GitHub and answers two different questions about each one: is it moving right now, and would you bet a product on it.
Top AI Repos tracks AI repositories on GitHub and answers two different questions about each one: is it moving right now, and would you bet a product on it.
I transfer the backend of yolov3 into Mobilenetv1，VGG16，ResNet101 and ResNeXt101
| Date | Stars |
|---|---|
| 2026-07-24 | 561 |
| 2026-07-25 | 561 |
| 2026-07-28 | 561 |
| 2026-07-30 | 561 |
| 2026-08-06 | 561 |
Today
— stars today
This week
— stars this week
This month
— stars this month
Momentum
0.0
growth rate 0.00%/day
# keras-yolo3-Mobilenet
[](LICENSE)
## Introduction
A Keras implementation of YOLOv3 (Tensorflow backend) inspired by [allanzelener/YAD2K](https://github.com/allanzelener/YAD2K).
#### And I change the backend of darknet53 into
- [x] Mobilenet
- [x] VGG16
- [x] ResNet101
- [x] ResNeXt101
## Experiment on open datasets
| Model name | InputSize | TrainSet | TestSet | mAP | Speed | Ps |
| ----- | ------ | ------ | ------ | ----- | ----- | ----- |
| YOLOv3-Mobilenet | 320x320 | VOC07 | VOC07 | 64.22% | 29fps | Keras on 1080Ti |
| YOLOv3-Mobilenet | 320x320 | VOC07+12 | VOC07 | 74.56% | 29fps | Keras on 1080Ti |
| YOLOv3-Mobilenet | 416x416 | VOC07+12 | VOC07 | 76.82% | 25fps | Keras on 1080Ti |
| [MobileNet-SSD](https://github.com/chuanqi305/MobileNet-SSD) | 300x300 | VOC07+12+coco | VOC07 | 72.7% | (unknown) ||
| [MobileNet-SSD](https://github.com/chuanqi305/MobileNet-SSD) | 300x300 | VOC07+12 | VOC07 | 68% | (unknown) ||
| [Faster RCNN, VGG-16](https://github.com/ShaoqingRen/faster_rcnn)| ~1000x600 | VOC07+12| VOC07 | 73.2% | 151ms | Caffe on Titan X |
|[SSD,VGG-16](https://github.com/pierluigiferrari/ssd_keras) | 300x300 | VOC07+12 | VOC07 | 77.5% | 39fps | Keras on Titan X |
#### PS:
1. Compared with MobileNet-SSD, YOLOv3-Mobilenet is much better on VOC2007 test, even without pre-training on Ms-COCO
2. I use the default anchor size that the author cluster on COCO with inputsize of 416\*416, whereas the anchors for VOC 320 input should be smaller. The change of anchor size could gain performance improvement.
3. Evaluation on https://github.com/Adamdad/Object-Detection-Metrics.git
4. I only use the pure model of YOLOv3-Mobilenet with no additional tricks.
# Guide of keras-yolov3-Mobilenet
1.train_Mobilenet.py
> * **Code for training**
> * I change some of the code to read in the annotaions seperately (train.txt and val.txt), remember to change that, and the .txt file are in the same form descibed below
2.yolo3/model_Mobilenet.py
> * **Model_Mobilenet is the yolo model based on Mobilenet**
> * If you want to go through the source code,ignore the other function,please see the yolo_body
(I extract three layers from the Mobilenet to make the prediction)
3.yolo_Mobilenet.py
> * **Testing on images**
###### Be sure that you do not load pretrained model when training because I did it on keras_applications,and the keras will load the pretrained model for you
##### if you find anything tricky, contact me as you wish
---
# Evaluation
#### Please use this repo to draw the RP curve calculate the MAP https://github.com/Adamdad/Object-Detection-Metrics.git
---
# Guide of keras-yolov3
[this is the guide for darknet53 not mobilenet]
## Quick Start
1. Download YOLOv3 weights from [YOLO website](http://pjreddie.com/darknet/yolo/).
2. Convert the Darknet YOLO model to a Keras model.
3. Run YOLO detection.
```
wget https://pjreddie.com/media/files/yolov3.weights
python convert.py yolov3.cfg yolov3.weights model_data/yolo.h5
python yolo_video.py [OPTIONS...] --image, for image detection mode, OR
python yolo_video.py [video_path] [output_path (optional)]
```
For Tiny YOLOv3, just do in a similar way, just specify model path and anchor path with `--model model_file` and `--anchors anchor_file`.
### Usage
Use --help to see usage of yolo_video.py:
```
usage: yolo_video.py [-h] [--model MODEL] [--anchors ANCHORS]
[--classes CLASSES] [--gpu_num GPU_NUM] [--image]
[--input] [--output]
positional arguments:
--input Video input path
--output Video output path
optional arguments:
-h, --help show this help message and exit
--model MODEL path to model weight file, default model_data/yolo.h5
--anchors ANCHORS path to anchor definitions, default
model_data/yolo_anchors.txt
--classes CLASSES path to class definitions, default
model_datExcerpt of 7,218 characters
Read on GitHubWould you bet a product on this? Bounded 0–100 and slow moving.
matched fp:67b98969efe3196c, topic:object-detection, topic:yolo, readme:object detection
matched fp:67b98969efe3196c, topic:tensorflow, readme:pre-training
