Skip to content

Code for undergraduate thesis "Active Learning for Deep Object Detection".

Notifications You must be signed in to change notification settings

CtfChan/pytorch_ssd_active_learning

Repository files navigation

Active Learning for Deep Object Detection

This repo contains the code that I used to write my undergraduate thesis on active learning for deep object detection. The code in the repo was adapted from sgrvinod's tutorial on object detection. If you find any errors or room for improvements please do not hesitate to leave a pull request.

Setup Instructions

Clone and install requirements

git clone https://github.com/CtfChan/pytorch_ssd_active_learning.git
cd pytorch_ssd_active_learning
conda env create -f environment.yml
conda activate tf_gpu

Prepare dataset

cd pytorch_ssd_active_learning/voc
bash download_voc.bash #this will take a while
python3 create_data_lists.py #this will save the paths of the images to a json file

Run Experiments

python3 train_baseline.py # this will take a while
bash generate_plots.bash # this takes even longer

Data visualizations

The experiments will be saved in to the results/ directory. You can run any of the .ipynb files to recreate the plots that were used in my thesis.

About

Code for undergraduate thesis "Active Learning for Deep Object Detection".

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published