Fine-tuning a large language model depends on optimizing several key hyperparameters to unlock its full potential. In this discussion, we focus on the three most important hyperparameters – batch s...
In the ever-changing world of machine learning, keeping our code simple and easy to maintain is more important than ever. That’s where TrlParser comes in—a smart extension of Hugging Face’s HfArgu...
Let’s recap how LLMs are created, it will help to better understand the whole fine-tuning pipeline. The training lifecycle typically involves three stages: Pre-training The model is initially...
Adding TensorBoard as a data visualization tool to your Keras model is easier than you might think. With just a few extra lines of code, you can gain valuable insights into your model’s training pr...
Weights & Biases (W&B) is not just a logging tool—it’s a robust platform that makes your machine learning experiments organized, reproducible, and visually compelling. In this article, we d...
Keeping your data in-house In today’s data-driven world, companies are increasingly concerned about sensitive data leaving their boundaries. Deploying your own Large Language Model (LLM) isn’t jus...
Surgical Fine-Tuning: Precision Adjustments for Language Models with PyReFT Fine-tuning large language models (LMs) can be a challenging task, especially when trying to achieve high performance wi...
As organizations increasingly demand on‑premises solutions for data security and control, Milvus—a leading open‑source vector database—offers attractive features for managing high‑dimensional data....
CogniX is an enterprise-level Retrieval-Augmented Generation (RAG) system, designed to manage and semantically analyze millions of documents with precision and efficiency. It represents the forefro...
Welcome to this exciting journey where we merge machine learning with the skies! In this post, we’ll show you how we created a dataset from the FAA manual to fly a hot air balloon—and now it’s time...