How to Train SSD MOBILENET for detecting dump trucks | Tutorial | Accubits

preview_player
Показать описание
This tutorial shows how you can train an object detector neural network to detect custom objects of your choice in videos. It utilizes the TensorFlow object detection API to train an SSD MobileNet V2 to detect dump trucks in videos. You can see dataset sampling, data annotation and training strategies to be employed when creating the model that detects any specific objects (dump trucks in our case). Watch this space for links to the sample code and data, so you can build the model of your choice.
Рекомендации по теме
Комментарии
Автор

Brilliant tutorial! Exactly what I was looking for, gave it a thumbs up. Could you PLEASE share the repo for this?

tahrirsiddiqui
Автор

The git hub repo doesn't have train py can you please help with it, it has model_main and difference in versions and due to depricated lib I am unable to go further

Rinkakumawatbvoc
Автор

Can you do a tutorial with multi class?

crickshowshorts
Автор

Can you share mobile net have trained ?

Sugar
Автор

Great tutorial! Would you mind letting us know where you're getting the number of at least 300 pics for training to detect a class?

tahrirsiddiqui
Автор

Is there any way to measure the performance of the mode mAP + the error loss curve

digitalzoul
Автор

Nice tutorial. How do i train for counting the number of trucks?

silasshaibubg
Автор

May i question for you?
I confused with loss in ssd, loss in ssd have a formula so how to implementation to code in ssd? Please help me.

bambangsolehudin
Автор

would you please do the same thing for .txt label files (YOLO format). I want to convert .txt to xml....

NaqqashDilshad
Автор

You use pycharm ?
I need for this tutorial tensorflow gpu?
PLEASE HELP ME PLEASE

fatimaslim
Автор

you did a good job! but it will be better if you share the code

hichamm
Автор

Great explanation. Could you share us the code?

JRichardsz
Автор

From where the support text file can be found ? For windows.

mehreenkhan
Автор

Hey man, great tutorial.. please can you give me repository link of your code? Or tell me where i can find it? Please

yudaskata
Автор

Sir caan you please provide the code used here

yeshwantkumar
Автор

your tutorial is help me so much! if I want your project, how can I got it?

sorpiseth
Автор

how to extract the parameter of ssd mobilenet v2?

Yourself
Автор

What version of Tensorflow are you using?

jimmytimes
Автор

hai, than you, how to make a frozen graph from this inference graph ? and my requirement is i want to make a model which detect human and conveyor belt machine only. if i used transfer learning we get a model trained almost 90 classes from that i need only one class which is human and again i need to train that model with conveyor belt machine.what is the best option for me? is human and conveyor belt machine train from scratch or transfer learning

samilajuneed
Автор

Awesome tutorial Btw! I have trained using Faster R-CNN, displayed the losses on tensor board but you know it shows only the losses, how can we generate the mAP of the training? Thanks.

shawkatali