InThe ML IntuitionbyPraneeta Mallela·Dec 1, 2025The Logistics of Large-Scale Training: DP, TP, PP, and the Role of CommunicationWhen I first started working on distributed training, I didn’t realize how much was happening behind the scenes. To put it simply — I was…A response icon1A response icon1
InThe ML IntuitionbyPraneeta Mallela·Oct 10, 2025Tricks to Fine-Tuning Without Model Tweaks: Achieving Generalization Through DataA guide to data-driven techniques like that improve model performance without changing architecture across ML domains.A response icon2A response icon2
InThe ML IntuitionbyPraneeta Mallela·Aug 21, 2025L1 vs L2 Regularization vs Weight Decay: Key Differences Every ML Engineer Must KnowOverfitting manifests differently based on the training pipeline one follows and its artifacts. Training loss dropping with a increase in…
InThe ML IntuitionbyPraneeta Mallela·Jun 12, 2025You’ve been using Negative Log Loss — Here’s what it actually meansThe machine learning field is full of tools we use daily but rarely stop to question. Negative Log Loss (NLL) is one of them. Whether…
InThe ML IntuitionbyPraneeta Mallela·Mar 10, 2025How Does MoCo Learn Without Labels? The Magic Behind Momentum Contrast ExplainedThis article explains how momentum contrast revolutionizes self-supervised learning
InThe ML IntuitionbyPraneeta Mallela·Mar 5, 2025Why CNNs Aren’t Truly ‘Convolutional’Leveraging mirroring Source: OpenArt
Praneeta Mallela·Oct 5, 2024Celestial Battles and Divine Grace: Unforgettable Verses from the Devi MahatmyaThe Durga Saptashati, or Devi Mahatmya, originates from the Markandeya Purana. In 700 shlokas across 13 chapters, it recounts how The…