Skip to content

VinceJnz/object_detection_using_tensorflow

This branch is 34 commits ahead of, 69 commits behind jackliu333/object_detection_using_tensorflow:main.

Folders and files

NameName
Last commit message
Last commit date

Latest commit

659b94b · Mar 1, 2023
Jan 20, 2023
Jan 20, 2023
Jan 20, 2023
Oct 11, 2022
Dec 3, 2022
Dec 4, 2022
Dec 3, 2022
Dec 4, 2022
Jun 14, 2021
Feb 28, 2023
Feb 28, 2023
Sep 19, 2022
Jan 31, 2023
Mar 1, 2023
Feb 3, 2023
Jan 20, 2023
Feb 8, 2023
Feb 28, 2023
Jan 27, 2023
Dec 19, 2022
Feb 7, 2022
Feb 7, 2022
Jan 26, 2023
Jan 26, 2023
Feb 7, 2022
Feb 7, 2022
Feb 28, 2023
Feb 7, 2022

Repository files navigation

Training customized object detection model using Tensorflow Object Detection API

This repository provides hassle-free tutorial on training customizd object detection model via transfer learning using Tensorflow Object Detection API, based on Google Colab.

This fork has the following changes

  1. Works on Windows 10
  2. COCO to TF Record: Import COCO files to TF Record files (that are suitable for use with object detection training)
  3. COCO tiling utility: Used to break up large images (with annotations) into smaller images (tiles), and fix up the annotations to line up with the smaller images

label studio

http://localhost:8080/projects/6/data/local-storage/6/9ee06b40-Day_1_Images_Flight_1_DSC09209_geotag_1.JPG

Model training

Training details

https://www.tensorflow.org/guide/distributed_training https://www.tensorflow.org/guide/checkpoint

Execution flow for training

[model_main.py] main [model_lib_v2.py] train_loop [inputs.py] train_input [image_resizer_builder.py] build

????????? train_step_fn() eager_train_step _compute_losses_and_predictions_dicts(...) #model prediction is done here

??????????

About

No description, website, or topics provided.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages

  • Jupyter Notebook 99.3%
  • Other 0.7%