README
My solution files for identifying dance form challenge- https://www.hackerearth.com/challenges/competitive/hackerearth-deep-learning-challenge-identify-dance-form/problems/
What is this repository for?
The problem is a classification problem. But I framed the challenge as object detection and recognition problem, and solved it using a Darknet based system- YOLOv4 You can find all details related to this system in below url- https://github.com/AlexeyAB/darknet
How do I get set up?
This project has been created in following environment
- Anaconda(Python 3.7.3 version)
- Windows 10
- CUDA 10.1
- cuDNN 7.6.5
- Visual Studio 2017
How to train data from scratch
Follow below post- http://dsbyprateekg.blogspot.com/2020/06/identify-eight-types-of-indian.html
How to run
- Install and build Darkent in your system by following below post- http://dsbyprateekg.blogspot.com/2020/05/how-to-install-and-compile-yolo-v4-with.html
- Once build successfully put the obj folder inside the vcpkg-master\installed\x64-windows\tools\darknet\data folder
- Copy obj.data, obj.names, train.txt and test.txt files also inside the vcpkg-master\installed\x64-windows\tools\darknet\data folder
- Download the trained weights from below url and put inside the vcpkg-master\installed\x64-windows\tools\darknet\backup folder- https://www.kaggle.com/iamprateek/yolo4dance-weights
- Open windows powershell, go to the Darknet root directory and run below command- ./darknet detector test data/obj.data cfg/yolo-obj.cfg backup/yolo-obj_final.weights -ext_output -dont_show -out data/dance_result.json < data/test.txt
- Above command will generate prediction in json format, now to prepare the submit file run the json_csv_submission2.py file
My Github Repo link
https://github.com/dsbyprateekg/hackerearth-deep-learning-challenge-identify-dance-form
My Blog Post
http://dsbyprateekg.blogspot.com/2020/07/my-solution-to-hackerearths-identify.html
