Since 2017, transformers have demonstrated their superiority in performance and computational efficiency, surpassing recurrent neural networks (RNNs). The attention mechanism introduced in the paper ‘Attention is All You Need’ and their ability to parallelize training—a feat traditional RNNs struggled with—attribute this superiority. However, transformers come with a challenge: the memory and inference costs associated with … Continued

Introduction Google -particularly its DeepMind squad- has launched a set of lightweights models, called Gemma, the same that were involved in Gemini creation. It is available in two sizes, 2B and 7B. It comes with an outstanding Responsible Generative AI Toolkit, an SFT for several frameworks and ready-to-use libs and colabs. Pre-trained models or customized … Continued

Model Merging 1. Introduction Model composition is a well-known problem in the Machine Learning community. Its aim is to extend the capabilities of a model, without forgetting what it already knows. Let’s consider a situation in which we have a good performing model in a certain task (e.g. Text-to-SQL) but it was fine-tuned for that … Continued

Introduction Making a prototype of a Machine Learning solution is crucial, it acts as a cornerstone for effective problem-solving. Because it lets you quickly test and validate ideas, which is key in a field such as Machine Learning, characterized by complex algorithms and unpredictable data patterns. Through this process, teams can quickly assess the viability … Continued

1. Introduction 1.1 Motivation Have you ever found yourself tirelessly scouring the internet for that one image that perfectly conveys your creative vision, only to come up short? Perhaps you’re a content creator on the quest for visuals that align seamlessly with your ideas. But hours of web surfing yield little more than frustration. Imagine … Continued

Introduction Diffusion models have earned a special place in the AI visual Content Generation landscape, dethroning GANs and positioning themselves as the go-to approach when creating realistic content. As technologies like LoRAs and Latent Consistency Models arrived, these models started to be less restrictive in terms of time and computing resources, and new possibilities and … Continued

Machine learning and data projects, in contrast with traditional software engineering projects, are governed by a resource characterized by unpredictable patterns, distributions, and biases – namely, the data itself. For a successful implementation of projects that give value through inference over datasets, we often consider: Does the collaboration between Agile methodologies and Machine Learning projects … Continued

Imagine you have to develop a computer vision application that must run in an industrial or agricultural environment with limited connectivity. Think, for example, of an application that detects plagues using a camera mounted on agricultural machinery.  Now imagine an application that monitors some machine in an industrial plant and needs to raise a real-time … Continued

Going on with our blog series on Large Language Models (LLMs) today we will talk about a successful finetuning case. While a well-crafted prompt can get the job done in many scenarios, there are times where it might not be enough. Finetuning steps in when we need results of higher quality compared to using prompts … Continued