https://datasanta.net/en Wed, 16 Apr 2025 18:51:20 -0000 Wed, 16 Apr 2025 18:51:20 -0000 1440 MkDocs RSS plugin - v1.17.1 None https://datasanta.net/ Nick Ovchinnikov Dive deep into loss functions, expected risk in the MicroTorch.https://datasanta.net/2025/04/13/empirical-risk-and-cross-entropy-in-microtorch/ Sun, 13 Apr 2025 00:00:00 +0000DataSantahttps://datasanta.net/2025/04/13/empirical-risk-and-cross-entropy-in-microtorch/ Nick Ovchinnikov Learn how to build a PyTorch-like, autograd-powered Deep Learning Framework with automatic differentiation.https://datasanta.net/2025/04/03/microtorch---deep-learning-from-scratch/ Thu, 03 Apr 2025 00:00:00 +0000DataSantahttps://datasanta.net/2025/04/03/microtorch---deep-learning-from-scratch/ Nick Ovchinnikov Implement the Cross-Entropy & Softmax. Tackle the Fashion-MNIST challenge!https://datasanta.net/2025/02/05/classification---cross-entropy--softmax/ Wed, 05 Feb 2025 00:00:00 +0000DataSantahttps://datasanta.net/2025/02/05/classification---cross-entropy--softmax/ Nick Ovchinnikov Why does gradient descent struggle with exploding gradients? Make SGD stable!https://datasanta.net/2025/02/01/sgd-momentum--exploding-gradient/ Sat, 01 Feb 2025 00:00:00 +0000DataSantahttps://datasanta.net/2025/02/01/sgd-momentum--exploding-gradient/ Nick Ovchinnikov Multi-Layer neural networks solve complex, non-linear problems like the spiral datasethttps://datasanta.net/2025/01/28/solving-non-linear-patterns-with-deep-neural-network/ Tue, 28 Jan 2025 00:00:00 +0000DataSantahttps://datasanta.net/2025/01/28/solving-non-linear-patterns-with-deep-neural-network/ Nick Ovchinnikov Neural Network Training with Stochastic Gradient Descent (SGD)https://datasanta.net/2025/01/22/mastering-neural-network---linear-layer-and-sgd/ Wed, 22 Jan 2025 00:00:00 +0000DataSantahttps://datasanta.net/2025/01/22/mastering-neural-network---linear-layer-and-sgd/ Nick Ovchinnikov From Gaussian to Xavier and He methods, with math foundations and code examples.https://datasanta.net/2025/01/14/weight-initialization-methods-in-neural-networks/ Tue, 14 Jan 2025 00:00:00 +0000DataSantahttps://datasanta.net/2025/01/14/weight-initialization-methods-in-neural-networks/ Nick Ovchinnikov How the sigmoid function transforms raw model outputs (logits) into probabilitieshttps://datasanta.net/2025/01/03/the-journey-from-logits-to-probabilities/ Fri, 03 Jan 2025 00:00:00 +0000DataSantahttps://datasanta.net/2025/01/03/the-journey-from-logits-to-probabilities/ Nick Ovchinnikov A concise guide to Cross-Entropy loss with step-by-step explanations and visualizations.https://datasanta.net/2024/12/28/cross-entropy-loss/ Sat, 28 Dec 2024 00:00:00 +0000DataSantahttps://datasanta.net/2024/12/28/cross-entropy-loss/ Nick Ovchinnikov Explore the gradient descent through a practical journey from simple paraboloids to complex functions.https://datasanta.net/2024/12/12/gradient-descent-ninja-with-momentum/ Thu, 12 Dec 2024 00:00:00 +0000DataSantahttps://datasanta.net/2024/12/12/gradient-descent-ninja-with-momentum/ Nick Ovchinnikov Delving into the pitfalls of numerical differentiation, exploring its impact on gradient computation.https://datasanta.net/2024/12/10/instability-in-numerical-differentiation/ Tue, 10 Dec 2024 00:00:00 +0000DataSantahttps://datasanta.net/2024/12/10/instability-in-numerical-differentiation/ Nick Ovchinnikov Follow along with our interactive Python plot where we demonstrate the gradient descent in action.https://datasanta.net/2024/12/05/gradient-descent---downhill-to-the-minima/ Thu, 05 Dec 2024 00:00:00 +0000DataSantahttps://datasanta.net/2024/12/05/gradient-descent---downhill-to-the-minima/ Nick Ovchinnikov Explore why the gradient vector indicates the direction of steepest ascent for functions.https://datasanta.net/2024/12/04/why-does-the-gradient-point-upwards/ Wed, 04 Dec 2024 00:00:00 +0000DataSantahttps://datasanta.net/2024/12/04/why-does-the-gradient-point-upwards/ Nick Ovchinnikov Dive into the world of derivatives, from understanding the mathematical fundamentals to implementing Python code.https://datasanta.net/2024/11/30/mastering-derivatives-from-math-to-code/ Sat, 30 Nov 2024 00:00:00 +0000DataSantahttps://datasanta.net/2024/11/30/mastering-derivatives-from-math-to-code/ Nick Ovchinnikov We're diving into the realm of MNIST, a dataset that's like a treasure map for budding data scientists.https://datasanta.net/2024/11/26/dive-into-learning-from-data---mnist-video-adventure/ Tue, 26 Nov 2024 00:00:00 +0000DataSantahttps://datasanta.net/2024/11/26/dive-into-learning-from-data---mnist-video-adventure/ Nick Ovchinnikov The Heartbeat of Data Transformations - Dive into the essentials of matrix multiplication and broadcasting.https://datasanta.net/2024/11/20/matrix-multiplication-and-broadcasting/ Wed, 20 Nov 2024 00:00:00 +0000DataSantahttps://datasanta.net/2024/11/20/matrix-multiplication-and-broadcasting/ Nick Ovchinnikov Training Text-to-Speech Models - A Journey from Zero to...https://datasanta.net/2024/06/11/text-to-speech-tts-models-overview-little-theory-and-math/ Tue, 11 Jun 2024 00:00:00 +0000DataSantahttps://datasanta.net/2024/06/11/text-to-speech-tts-models-overview-little-theory-and-math/