Category Archives: PyTorch

I was at a conference where I was scheduled to give a presentation “Introduction to Neural Networks Using C#”. I figured I’d create a PyTorch version of my from-scratch C# demo, just in case I had lots of extra time … Continue reading →

I’ve always known that, in theory, to create a PyTorch binary classification prediction model, you can just pretend the problem is a multi-class classification problem. Bottom line: the multi-class approach works but has no advantage over the standard binary approach. … Continue reading →

Quantile regression is a machine learning technique to predict a single numeric value in situations where you care about under-prediction (or sometimes over-prediction). I’ll phrase the rest of this blog post in terms of scenarios where you mostly care about … Continue reading →

Poisson distributed data is often counts that arise when “things arrive” per some fixed unit of time. For example, if you observe the count of cars arriving at a traffic intersection each Monday between 11:00 and 11:05 PM, you might … Continue reading →

Bottom line: I did a few experiments that compared predicting multivariate Poisson distributed target data using 1.) a standard PyTorch neural network regression system and 2.) using the scikit-learn PoissonRegressor() module. The results were roughly the same. Several weeks ago, … Continue reading →

The DeepSeek R1 large language model, announced on January 20 of this year, shocked the deep learning community because it was much less expensive to create and is much less expensive to use, compared to all other current LLMs. According … Continue reading →

I was preparing to teach a class on PyTorch neural networks at the large tech company I work for. To prepare, I wanted to mentally review PyTorch neural network basics, including the mysterious NLLLoss function. 1. Early Days: One-Hot Targets, … Continue reading →