Kostya Numan’s Post

Today is our last day of the ML live workshop. And I am proud to see what we were able to build with our students in 1 week! Follow Kostya Numan to learn more about Machine Learning fundamentals. 💪🏼

Through our workshops you get the most in-demand skills. This is how we build them 👇

#DeepLearning #Asphalt_pavment #Custom_CNN Recently developed a custom CNN architecture for a dataset previously worked on with VGG16. By optimizing pixel width, I achieved better accuracy and performance. Excited to see the continued potential of deep learning in computer vision! 🎯🔍 GitHub repo : https://lnkd.in/gQe_Da_3 Main paper link:https://lnkd.in/gjJyEcpi

Abstractions of a Dataset in Pattern Recognition

🚀 Exploring the Power of Transfer Learning with Pytorch! 🌟 Just completed hands-on training with ResNet-18 using the STL-10 dataset. From data preprocessing to visualizations, and diving deep into ResNet's architecture, it's been an incredible journey into image classification. 📊📸 Github Link: "https://lnkd.in/g-csKmGY" Excited about the potential of leveraging pretrained models for real-world applications. 💡 #DeepLearning #TransferLearning #ResNet #MachineLearning #AI

Just finished Advanced Computer Vision with TensorFlow

I was a guest for an episode of "Software Engineering Radio" recently, and it's now available to listen to here: https://lnkd.in/gV7Bys6X I really enjoyed chatting with Philip Winston about a variety of topics including the role of a data scientist, the difference between running experiments in notebooks and building an automated pipeline for production, machine learning vs. AI, the typical pipeline steps for machine learning, and the role of software engineering in data science. Hope you enjoy it and let me know your comments or questions!

Day 185 of My Data Science Journey 🚀 Today was all about diving deep into the fascinating world of Support Vector Machines (SVMs). Here's what I covered: 1. Geometric Intuition of SVMs: Understanding how SVMs find the optimal hyper-plane that best separates the classes in the feature space. 2. Equation of a Hyper-plane in N Dimensions: Delving into the mathematics behind defining hyper-planes in multi-dimensional spaces. 3. Mathematics of SVM | Hard Margin SVM: Exploring the concept of hard margin SVMs and the conditions under which they work best. 4. Support Vector Machine | Code Sample: Implementing SVMs in code to see theory in action. 5. Mathematics of Support Vector Machine | Soft Margin SVM: Studying soft margin SVMs to handle non-linearly separable data and adding flexibility with slack variables. 6. Kernel Trick in SVM | Geometric Intuition: Understanding how the kernel trick maps data into higher dimensions to make it linearly separable. 7. Kernel Trick in SVM | Code Example: Applying different kernel functions in code and observing their impact on SVM performance. Excited to continue this journey and see how these powerful tools can solve complex real-world problems. Onward and upward! 📈🔍💡 #DataScience #MachineLearning #SupportVectorMachines #SVM #KernelTrick #Coding #Mathematics #AI #100DaysOfCode

Hello Connections! Just finished the course “Build Your Own GPTs” by Alina Zhang!. 🌟 This course has been an incredible journey, diving deep into the intricacies of creating and customizing GPT models. From understanding the architecture to fine-tuning models for specific applications, I've learned how to harness the power of AI to build more sophisticated and tailored solutions.💡 Check it out: https://lnkd.in/eZcTDuwj #chatbotdevelopment.

🚀 Just wrapped up a deep learning project using CNN architectures like VGG16, InceptionV3, and InceptionResNetV2 to classify house types—detached, semi-detached, terraced, and flats—from images. InceptionResNetV2 achieved an impressive 92.48% accuracy! This work highlights the power of transfer learning in complex image classification tasks and its potential for applications in real estate analytics. Excited to explore more in this field! #DeepLearning #CNN #TransferLearning #RealEstateTech https://lnkd.in/eZE6qqNT