Bijay Kumar    I’ve found that a single line chart often hides the most important parts of a story. When you are dealing with time series data, showing just the average or the trend line isn’t enough to see the full picture. You often need to see the spread or the volatility of that data over specific intervals, ...

Bijay Kumar    Visualizing data over time is a task I perform almost daily as a developer. While line charts are the standard, there are many times when a scatter plot is actually the better choice. In my experience, scatter plots are perfect for identifying outliers in time-based data or showing specific events...

Bijay Kumar    In my years of building data dashboards for US-based financial firms, I’ve realized that Matplotlib’s default scaling isn’t always perfect. Sometimes, the auto-scaling feature hides the small differences between data points that actually matter for your analysis. In this tutorial, I will show you...

Bijay Kumar    Visualizing data with massive variations can be quite a headache. I often find that a standard linear scale makes small values disappear while large values hog the entire chart. In my experience, switching to a logarithmic scale is the best way to solve this. It allows you to see the percentage...

Bijay Kumar    Dynamic data visualization is a skill I frequently utilize, particularly when tracking real-time trends or simulating movements. In my experience, static plots often fail to tell the whole story, particularly when you need to see how data points evolve. I have spent many hours refining the process...

Bijay Kumar    I have spent the last four years building deep learning models, and if there is one thing I have learned, it is that standard Convolutional Neural Networks (ConvNets) sometimes miss the big picture. While convolutions are great at picking up local patterns, they often struggle to understand which...

Bijay Kumar    I’ve found that scaling Vision Transformers (ViT) often leads to significant training instability. Standard ViT architectures tend to saturate or diverge when you add too many layers, which can be quite frustrating during model development. Recently, I started using Class Attention Image...

Bijay Kumar    I have often noticed a frustrating drop in accuracy when deploying models. You train a model on 224 times 224 images, but the real-world performance only peaks when you feed it larger images during inference. This phenomenon is known as the train-test resolution discrepancy, and it occurs because...

Bijay Kumar    I have spent a significant amount of time building complex deep learning models that perform brilliantly but are far too heavy for mobile devices. In my experience, Knowledge Distillation is the most effective way to shrink a massive Teacher model into a compact Student model while keeping the...

Bijay Kumar    In my four years of working with Keras, I’ve realized that moving from traditional CNNs to Vision Transformers (ViT) is a massive shift. The most confusing part for many developers I mentor is how we actually turn a standard image into a sequence of tokens that a Transformer can understand. In this...