Deep Learning Object Detection by Training & Deploying YOLOX

Name: Deep Learning Object Detection by Training & Deploying YOLOX
Rating: 4.3 (9 reviews)

Finetuning and testing a YOLOX model on custom built dataset. Creating and deploying object detection API to cloud

Created byNeuralearn Dot AI

Last updated 5/2023

English

English [Auto],

What you'll learn

Master the basics of Object detection
Understanding pre-deep learning algorithms like haarcascades
Understanding deep learning algorithms like YOLO and YOLOX
Create your own dataset with Remo
Understanding the Pascal VOC dataset
Convert your custom dataset to Pascal VOC Format
Testing and training YOLOX model on custom dataset
Integrating Wandb for experiment tracking
Converting trained model to Onnx format
Understanding how APIs work
Building Object detection API with Fastapi
Deploying API to the Cloud
Load testing the API with Locust
Running the object detection model in c++

Course content

6 sections • 23 lectures • 6h 58m total length

Welcome1:04
Explore state-of-the-art deep learning object detection models, from theory to practice, by preparing datasets, fine-tuning models for your tasks, and deploying end-to-end systems in the cloud.
General Introduction6:47
Learn deep learning object detection with YOLOX, localizing and classifying objects in images and videos, with applications in medicine, parking systems, robotics harvesting, sports analytics, and autonomous vehicles.
About this Course2:57
Explore object detection from haar cascades to YOLOX, train on Pascal VOC with Remote AI, fine-tune, deploy via FastAPI, and test inference in ONNX and C++.
Link to code0:01

Haar Cascades and Histogram of gradients36:53
Discover how Haar cascades and histogram of oriented gradients enable grayscale-based feature extraction and subsequent machine learning for object detection, using OpenCV to process images and deploy detection.
Convolutional Neural Networks24:06
Explore how convolutional neural networks learn from image data by using convolutional and pooling layers for feature extraction, then classify with fully connected layers trained via gradient descent.
RCNN,FastRCNN, FaterRCNN7:06
Explore convolutional neural networks for object detection, from RCNN and Fast RCNN to Faster RCNN, using selective search, ROI pooling, and YOLO.

Understanding YOLO (You Only look once)25:05
Explore the single-stage YOLO object detection approach, from grid-based encoding and bounding boxes to non-max suppression, IOU scoring, and the shift toward anchor-free YOLOX.
Understanding YOLOX14:06
Explore yolox, an anchor-free detector with a decoupled head, data augmentation, and multi-positives to improve object detection. Learn the machine learning development life cycle from data gathering to deployment.

Pascal VOC dataset15:43
this lecture explains organizing the fashion dataset in Pascal VOC format with annotations, image sets, and JPEG images. it compares Pascal VOC to COCO and shows bounding boxes for detection.
Preparing a custom dataset with Remo20:12
Learn to build a custom fashion object-detection dataset with remote, annotate images, export remote json, convert to Pascal VOC in Colab, and split data for training and evaluation.
Preparing a custom dataset with LabelImg14:58
Learn to create custom datasets with label image by annotating images, generating Pascal VOC and YOLO XML files, and organizing train, val, and test splits for model training.
Assignment0:09

Testing and FInetuning on Custom Dataset1:00:56
Test and finetune the yolox model on a custom fashion dataset, training and evaluating in Colab with Google Drive access, and compare yolox variants for deployment.
Wandb integration17:05
Track YOLO X Nano experiments with one DB, log dataset and model artifacts, and visualize training metrics for edge-device deployment.
Running inference on Onnx model8:16
Convert a trained PyTorch model to the ONNX format and run cross-framework inference with the ONNX runtime, enabling deployment across TensorFlow, CAFE, and diverse hardware.
Assignment0:04

Understanding how APIs work19:54
Explore how APIs enable web and mobile apps to access models using http, rest, and json, with get, post, put, delete methods, status codes, headers, and Postman testing.
Building an API with Fastapi1:37:15
Learn to build a FastAPI-based API for object detection with YOLOX, including virtual environments, uvicorn, swagger docs, schemas, and deployment with gunicorn.
Deploying on heroku18:24
Deploy a YOLOX nano object detection API to Heroku, configure headless OpenCV, CPU PyTorch, and a clean requirements.txt, then test endpoints with Postman.
Load testing with Locust24:06
Learn to perform load testing of a YOLOX-based API using Locust, simulating millions of users locally and in the cloud, and diagnose with logs to optimize deployment.
Integration with C++3:26
Explore running an onnx-based object detection system in C++, using OpenCV and Onnx runtime, with a CMake build and the Basilio Nano model on a demo image.
Assignment0:02

Requirements

Basic Knowledge of Python
Access to an internet connection, as we shall be using Google Colab (free version)

Description

Object detection algorithms are everywhere. With creation of much more efficient models from the early 2010s, these algorithms which now are built using deep learning models are achieving unprecedented performances.

In this course, we shall take you through an amazing journey in which you'll master different concepts with a step by step approach. We shall start from understanding how object detection algorithms work, to deploying them to the cloud, while observing best practices.

You will learn:

Pre-deep learning object detection algorithms like Haarcascades
Deep Learning algorithms like Convolutional neural networks, YOLO and YOLOX
Object detection labeling formats like Pascal VOC.
Creation of a custom dataset with Remo
Conversion of our custom dataset to the Pascal VOC format.
Finetuning and testing YOLOX model with custom dataset
Conversion of finetuned model to Onnx format
Experiment tracking with Wandb
How APIs work and building your own API with Fastapi
Deploying an API to the Cloud
Load testing a deployed API with Locust
Running object detection model in c++

If you are willing to move a step further in your career, this course is destined for you and we are super excited to help achieve your goals!

This course is offered to you by Neuralearn. And just like every other course by Neuralearn, we lay much emphasis on feedback. Your reviews and questions in the forum, will help us better this course. Feel free to ask as many questions as possible on the forum. We do our very best to reply in the shortest possible time.

Enjoy!!!

Who this course is for:

Beginner Python Developers curious about applying deep learning techniques like YOLO
Software developers interested in using A.I and deep learning for object detection
Students interested in learning about object detection and how it can be applied practically
AI Practitioners wanting to master how to deploy AI Models to the cloud very easily
Software developers who want to learn how state of art object detection models are built and trained using deep learning.
Students who study different Object Detection Algorithms and want to Train YOLO with Custom Data.
Students who study Computer Vision and want to know how to use YOLO and its variants like YOLOX for Object Detection

Deep Learning Object Detection by Training & Deploying YOLOX

What you'll learn

Explore related topics

Course content

Introduction4 lectures • 11min

Theoritical background3 lectures • 1hr 8min

Single Stage Algorithms2 lectures • 39min

Dataset Preparation4 lectures • 51min

Finetuning and Testing4 lectures • 1hr 26min

Deployment6 lectures • 2hr 43min

Requirements

Description

Who this course is for: