YOLOv3

Update!!! Now you can use Simpsons font when drawing boxes. Just change data/font/FiraMono-Medium.otf by data/font/font_Simpsonfont.ttf in dutils.py.

You can RUN YoloDetectorBot by following to next steps:

* Install `telebot` library https://github.com/eternnoir/pyTelegramBotAPI

1) Download Repository and `yolov3(tiny).h5` and put model into `models/yolov3(tiny).h5`.

Link to the models .h5: https://www.kaggle.com/kanametov/yolov3

2) RUN `bot.py` using Powershell and command `python bot.py`

3) or RUN `Bot.ipynb`

4) Finaly open Telegram Bot: https://web.telegram.org/@YoloDetectorBot and use command `/start`

Before to RUN model :

1) Download `yolov3.h5` and `yolov3(tiny).h5` models and put them into the folder `models/` as `models/yolov3.h5` and `models/yolov3(tiny).h5`.

* You can download just `yolov3(tiny).h5` and run all on Tiny version of YOLOv3.

Link to the models .h5: https://www.kaggle.com/kanametov/yolov3

2** If you are going to train your own model on VOC Dataset download the dataset using the link below and put it into the folder `dataset/` as follows `dataset/VOC2007`.

Link to the VOC Dataset: https://www.kaggle.com/kanametov/vocdataset

RUN model :

To RUN Object Detector on picture:

1) With Notebook: use PictureDetection.ipynb

2) With Windows Powershell: use command: `python detect_picture.py`

Parameters of `detect_picture.py`

Parameters	Short Commands	Full Commands	Default value
Image	`-i`	`--input`	`dog.jpg`
Image size	`-is`	`--input_size`	`416`
Score Threshold	`-st`	`--score_threshold`	`0.25`
IoUnion Threshold	`-it`	`--iou_threshold`	`0.5`

To RUN Object Detector on web-camera:

1) With Notebook: use WebDetection.ipynb

2) With Windows Powershell: use command: `python detect_web.py`

Parameters of `detect_web.py`

Parameters	Short Commands	Full Commands	Default value
Image size	`-is`	`--input_size`	`256`
Score Threshold	`-st`	`--score_threshold`	`0.25`
IoUnion Threshold	`-it`	`--iou_threshold`	`0.5`

Press `q` to exit from Web-Detector.

Navigation Menu

Repository files navigation

YOLOv3

You can RUN YoloDetectorBot by following to next steps:

* Install `telebot` library https://github.com/eternnoir/pyTelegramBotAPI

1) Download Repository and `yolov3(tiny).h5` and put model into `models/yolov3(tiny).h5`.

Link to the models .h5: https://www.kaggle.com/kanametov/yolov3

2) RUN `bot.py` using Powershell and command `python bot.py`

3) or RUN `Bot.ipynb`

4) Finaly open Telegram Bot: https://web.telegram.org/@YoloDetectorBot and use command `/start`

Before to RUN model :

1) Download `yolov3.h5` and `yolov3(tiny).h5` models and put them into the folder `models/` as `models/yolov3.h5` and `models/yolov3(tiny).h5`.

* You can download just `yolov3(tiny).h5` and run all on Tiny version of YOLOv3.

Link to the models .h5: https://www.kaggle.com/kanametov/yolov3

2** If you are going to train your own model on VOC Dataset download the dataset using the link below and put it into the folder `dataset/` as follows `dataset/VOC2007`.

Link to the VOC Dataset: https://www.kaggle.com/kanametov/vocdataset

RUN model :

To RUN Object Detector on picture:

1) With Notebook: use PictureDetection.ipynb

2) With Windows Powershell: use command: `python detect_picture.py`

Parameters of `detect_picture.py`

To RUN Object Detector on web-camera:

1) With Notebook: use WebDetection.ipynb

2) With Windows Powershell: use command: `python detect_web.py`

Parameters of `detect_web.py`

Press `q` to exit from Web-Detector.

About

Releases

Packages

Used by

Contributors

Languages

Name		Name	Last commit message	Last commit date
data		data
dataset		dataset
demo		demo
images		images
model		model
yolov3		yolov3
README.md		README.md
bot.py		bot.py
detect_picture.py		detect_picture.py
detect_web.py		detect_web.py
dutils.py		dutils.py
requirements.txt		requirements.txt
voc_utils.py		voc_utils.py

Latest commit

History

Folders and files

Repository files navigation

YOLOv3

You can RUN YoloDetectorBot by following to next steps:

* Install telebot library https://github.com/eternnoir/pyTelegramBotAPI

1) Download Repository and yolov3(tiny).h5 and put model into models/yolov3(tiny).h5.

Link to the models .h5: https://www.kaggle.com/kanametov/yolov3

2) RUN bot.py using Powershell and command python bot.py

3) or RUN Bot.ipynb

4) Finaly open Telegram Bot: https://web.telegram.org/@YoloDetectorBot and use command /start

Before to RUN model :

1) Download yolov3.h5 and yolov3(tiny).h5 models and put them into the folder models/ as models/yolov3.h5 and models/yolov3(tiny).h5.

* You can download just yolov3(tiny).h5 and run all on Tiny version of YOLOv3.

Link to the models .h5: https://www.kaggle.com/kanametov/yolov3

2** If you are going to train your own model on VOC Dataset download the dataset using the link below and put it into the folder dataset/ as follows dataset/VOC2007.

Link to the VOC Dataset: https://www.kaggle.com/kanametov/vocdataset

RUN model :

To RUN Object Detector on picture:

1) With Notebook: use PictureDetection.ipynb

2) With Windows Powershell: use command: python detect_picture.py

Parameters of detect_picture.py

To RUN Object Detector on web-camera:

1) With Notebook: use WebDetection.ipynb

2) With Windows Powershell: use command: python detect_web.py

Parameters of detect_web.py

Press q to exit from Web-Detector.

About

Topics

Resources

Stars

Watchers

Forks

Releases

Packages

Used by

Contributors

Languages

* Install `telebot` library https://github.com/eternnoir/pyTelegramBotAPI

1) Download Repository and `yolov3(tiny).h5` and put model into `models/yolov3(tiny).h5`.

2) RUN `bot.py` using Powershell and command `python bot.py`

3) or RUN `Bot.ipynb`

4) Finaly open Telegram Bot: https://web.telegram.org/@YoloDetectorBot and use command `/start`

1) Download `yolov3.h5` and `yolov3(tiny).h5` models and put them into the folder `models/` as `models/yolov3.h5` and `models/yolov3(tiny).h5`.

* You can download just `yolov3(tiny).h5` and run all on Tiny version of YOLOv3.

2** If you are going to train your own model on VOC Dataset download the dataset using the link below and put it into the folder `dataset/` as follows `dataset/VOC2007`.

2) With Windows Powershell: use command: `python detect_picture.py`

Parameters of `detect_picture.py`

2) With Windows Powershell: use command: `python detect_web.py`

Parameters of `detect_web.py`

Press `q` to exit from Web-Detector.