Skip to content

Darknet based yolov3 training and Inference in Tensorflow for single class (Kangaroo)

License

Notifications You must be signed in to change notification settings

pucha48/object_detection_kangaroo

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

19 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

object_detection_kangaroo

Darknet based yolov3 training and Inference in Tensorflow for single class (Kangaroo)

Credits:

I referred and used these fantastic repos during the implementation:

AlexeyAB/darknet

wizyoung/YOLOv3_TensorFlow

Training:

1. Dataset preparation and data augmentation. I have used free Roboflow for format conversion and augmentation
2. Training with 291 training dataset and 28 test dataset.
Training Parameters:

learning_rate=0.001 max_batches = 6000 steps=4000,5400

Inference:

There are two steps:

1. Weight Conversion from darknet weights to Tensorflow checkpoint
2. Testing on sample dataset taken from Kaggle sample_data

#Some results:

Sample 1

Sample 2

About

Darknet based yolov3 training and Inference in Tensorflow for single class (Kangaroo)

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published