Vicente Rodríguez

Vicente Rodríguez Blog

Deep Learning, Machine Learning, and Learning

Testing and workflow to train PyTorch models.

A summary of PyTorch training. From tips to wrap blocks of code, to testing ideas and good approaches.

Sept. 17, 2025

A quick recap on Reinforcement Learning

Now days, Reinforcement Learning has found a resurrection by the hand of Large Language Models. Given that reinforcement learning can be employed to train and further improve LLMs; Companies and researches are heavily investing in this area of training. Being the most recent advancement Group relative policy optimization, a policy approach to guide LLM to answer better responses.

Sept. 17, 2025

Triton library, GPUs, and Flash Attention

In the modern world of Deep Learning, it becomes more and more important to understand how GPUs work given their status as the core tool to develop and deploy massive models. Leading to a necessity of learning about GPUs inner design and about related libraries such as Triton.

July 3, 2025

Serving Deep Learning Models using the Publish Subscribe Pattern

Model serving can be achieved using different approaches, let's see how the Publish-Subscribe Pattern can be used to create an infrastructure to serve Deep Learning models along with Spot Instances to reduce the costs of this infrastructure

Feb. 16, 2022

Thin Plate Splines and its implementation as data augmentation technique

Thin-Plate Splines is a technique for data interpolation and smoothing. We can use this technique to transform or augment data like images to increase the number of training images or change the image content without impacting its style.

Dec. 8, 2021

Face Detection for low-end hardware using the BlazeFace Architecture

Let's see how we can implement the BlazeFace architecture for face detection that works in low-end hardware using TensorFlow 2.0

May 17, 2021

BlazeFace in iOS using TensorFlow Lite

The BlazeFace architecture is truly suitable for mobile devices, let's see how we can implement the face detection task in iOS using this architecture along with TensorFlow Lite

May 17, 2021

BlazeFace in iOS using CoreML

The BlazeFace architecture is truly suitable for mobile devices, let's see how we can implement the face detection task in iOS, this time using CoreML

May 17, 2021