Pranjal Khadka·Jul 16Scaling over RotationA matrix, in general, does two things to a vector: it rotates it and it scales it. Most of the time these two effects are tangled together…A response icon2A response icon2
Pranjal Khadka·Feb 26Training Dynamics: A Practical Guide to Training Deep Neural NetworksThere is a version of deep learning that exists in tutorials. You load a clean dataset, define a standard architecture and watch the…
Pranjal Khadka·Jan 14A History of InferenceAt a fundamental level, machine learning asks just one question: what is likely, given what I know? In probability theory, this question…
Pranjal Khadka·Dec 25, 2025The One Assumption That Makes Machine Learning PossibleIf you have ever worked with machine learning algorithms like neural networks or logistic regression, you have likely trained models…
Pranjal Khadka·Sep 15, 2025Build AI Agentic Systems That Actually WorkAgentic systems are rapidly gaining popularity as LLMs become more capable of executing tasks beyond simple conversation. However, AI…
InTowards AIbyPranjal Khadka·Jan 29, 2025How to Fine-Tune Any Large Language Model (LLM)Fine-tuning large language models (LLMs) has become an easier task today thanks to the availability of low-code/no-code tools that allow…
InTowards AIbyPranjal Khadka·Nov 10, 2024Why does Batch Normalization work ?Batch Normalization is a widely used technique for faster and stable training of deep neural networks. While the reason for the…
InTowards AIbyPranjal Khadka·Oct 11, 2024How to Deploy Machine Learning Models Using Docker and AWS: A Step-by-Step GuideMachine learning models are often developed in controlled environments where dependencies, libraries, and configurations are manually set…A response icon1A response icon1
InTowards AIbyPranjal Khadka·Sep 21, 2024Decoding CLIP: How Large-Scale Contrastive Learning Transformed Vision-Language ModelsMotivation
Pranjal Khadka·Jul 24, 2024Paper Breakdown : U-Net (Convolutional Networks for Biomedical Image Segmentation)U-Net is one of the most influential paper in the field of Image segmentation. It came out around 2015 by researchers from University of…