2.Train Custom Object Detection Model Using Detectron 2

Video Explains about the training of object detection model on custom dataset from scratch.

✅Recommended Gaming Laptops For Machine Learning and Deep Learning :

✅ Best Work From Home utilities to Purchase for Data Scientist :

✅ Recommended Books to Read on Machine Learning And Deep Learning:

Connect with me on :

#datascience #deeplearning #machinelearning

[fernandolee]

Great content sir, I tried and it worked thanks 😀, btw is there a way to show the accuracy of the train and test results? Thank you once again!

[lazzyboyhunter]

thanks for such informative video. can you make video on how to train custom text recognition model using tensorflow.

[yhnshkn]

Is it possible for you to create a tutorial on how to deploy trained models? Im kinda stuck on what to do next after training a model using detectron2

[muddassarjilani]

great video, how we can deploy this model using flask

[THEULTIMATEPRASHANT]

hi boss, i trained the model using the same steps... now i am looking for two things..
1. a python file to load this .pth model and do the detection
2. how to segment the detected object as this functionality can be achieved by tensorflow object detection.. we used detectron to segment (create mask over detected object) right?


thanks bro

[robr]

Any suggestions for me? I detected an object but I want to register or align it automatically. Any thoughts on how I can do this?

[projectcapstone]

At 20:32 , you are inferencing for a single image and the result is displayed as an image with box-predictions. Is there any way to get the result as a json with bounding box coordinates (for a single test_image
)?

[yhnshkn]

hi sir, thanks for the tutorial, it worked for me, however i have a question. How can i use the trained model if i want to use this in raspberry pi 4 64 bit? Should i just download the pth file?

[amolwankhade]

Sir, how can i get testing accuracy graph using tensorboard for detectron2??

[mayanksingh]

Sir i have 100*100 pixel images. I applied augmentation to them and have 21k images in train set.
I did everything according to your steps
Yet I am not getting any bounding boxes on my predictions. Please help sir

[yhnshkn]

hi sir, it's me again. I also want to ask how to convert it to onnx format so that it can run on raspberry pi as well? if it's possible can you do a tutorial for that? thank you very much for your help

[manuthvann]

Idk dear, but there is always an error with trainer.train function, and it’s related with the cuda version . Looking forward to hearing back from u

[deidy]

Man, I created a customized model with 3 classes (A, B and C). And I would like to take especific masks for a especific object. And I use this code: mask_A = <- but not always object A is 0, sometimes is 1 or 2. This is for pred_masks only, the object identification is always correct. Is there a way to fix pred_masks ID's?

[TheLearningBachpan]

Hi, I can get prediction with 2 class of labels however, if I trained model with 3 or more than 3 labels then model is not predicting anything.
I am using the same code for the same. So can you please guide me that how can train model with multiple labels?

[sneha]

Sir can we do the training on multilabel dataset also ??

[nichuffnoob]

Sir my saved file is as txt and not json sir

[ajaymane]

hello how to reuse pretrain model, i mean how to load pretrain detection model.