Priyanshu mauryaGetting to Know Jensen-Shannon Divergence (JSD)JSD or rather Jensen-Shannon Divergence is A distance measure built on the idea of Kullback-Leibler (KL) Divergence, which is the most…5d ago5d ago
Priyanshu mauryaBeyond Overfitting: Understanding Grokking in Model TrainingLet’s first understand the term “overfitting” and its causes. Overfitting occurs during training when a model learns the training data too…Jul 21Jul 21
Priyanshu mauryaMaximizing RAG Efficiency: Token Reduction Strategies and Cost SavingsWhy use many token when few token do trick?Jul 15Jul 15
Priyanshu mauryaUnderstanding Shannon Entropy and KL-Divergence through Information Theory“ Information theory gives us precise language for describing a lot of things. How uncertain am I? How much does knowing the answer to…Jul 7Jul 7
Priyanshu mauryaAlternative to Cosine Similarity Function for Self-Supervised TasksIn self-supervised training, we use contrastive loss to learn the feature embeddings of structures or text. For contrastive loss, we often…Jun 29Jun 29
Priyanshu mauryaWhy Use InfoNCE Loss in Self-supervised LearningTo better understand InfoNCE loss, let’s assume you need to train a model that converts textual or pictorial representations into…Jun 261Jun 261
Priyanshu mauryaGet Free GPU with more than 24GB Vram, to train your modelI was on the hunt for a GPU with more than 16GB of VRAM to train my model. I explored several providers like Colab, Kaggle, and Gradient…Jun 25Jun 25