Pinned Repositories
Unveiling-Gene-Disease-Connections
This study uses Graph Neural Networks (GNNs) to identify gene-disease associations from a large Knowledge Graph, enabling the prediction of novel links and providing insights into the genetic foundations of health conditions.
ApprAuto2023
BCG-Gamma-Challenge
We're proud to announce that our team secured the First Prize in the BCG GAMMA Challenge for Data Science and Consulting. This GitHub repository hosts all the code, models, and analytical techniques we employed during the competition.
Hydrid-AI-Course
This repository contains the final assessment for the Hybrid AI course at École CentraleSupélec in Paris, France. The final assessment involves a review of the paper titled "Physics-Informed Deep Neural Operator Networks," available at https://arxiv.org/pdf/2207.05748.
ImproveLLMStorytellingWithFineTuning
This workshop, led by Lucas José for the Google Developer Student Clubs, focuses on developing practical skills in fine-tuning smaller Language Models (LLMs) to enhance their storytelling capabilities. The aim is to transfer knowledge from a larger model (Meta Llama 2-70B) to a smaller one (Microsoft Phi 1-1B), ensuring high-quality narrative.
IsVoldemortAGorilla
MangroveAI
This repository contains the code and data for the paper titled "A Deep Learning-Based Approach for Mangrove Monitoring". The paper was accepted for presentation at the European Conference on Machine Learning (ECML), specifically at the Machine Learning for Earth Observation Workshop.
MODA
MODA (Multimodal Object Description Assistant) is an AI model that describes fashion items by combining visual and textual data. Powered by FashionCLIP and OPT, MODA provides descriptions of various fashion objects.
Scanipy
Scanipy stands for "scan it with Python"—it's your smart Python library for scanning and parsing complex PDF files like books, reports, articles, and academic papers. Utilizing cutting-edge Deep Learning algorithms, Scanipy transforms your PDFs into a treasure trove of extractable information: tables, images, equations, and text.
ViT-for-Dark-Matter-Morphology
Under the international Google Summer of Code program, the project "Transformers for Dark Matter Morphology with Strong Gravitational Lensing" was developed to improve the Machine Learning for Science pipeline using transformers, contributing to both astrophysics and machine learning research.
SVJLucas's Repositories
SVJLucas/Scanipy
Scanipy stands for "scan it with Python"—it's your smart Python library for scanning and parsing complex PDF files like books, reports, articles, and academic papers. Utilizing cutting-edge Deep Learning algorithms, Scanipy transforms your PDFs into a treasure trove of extractable information: tables, images, equations, and text.
SVJLucas/BCG-Gamma-Challenge
We're proud to announce that our team secured the First Prize in the BCG GAMMA Challenge for Data Science and Consulting. This GitHub repository hosts all the code, models, and analytical techniques we employed during the competition.
SVJLucas/ViT-for-Dark-Matter-Morphology
Under the international Google Summer of Code program, the project "Transformers for Dark Matter Morphology with Strong Gravitational Lensing" was developed to improve the Machine Learning for Science pipeline using transformers, contributing to both astrophysics and machine learning research.
SVJLucas/ApprAuto2023
SVJLucas/MangroveAI
This repository contains the code and data for the paper titled "A Deep Learning-Based Approach for Mangrove Monitoring". The paper was accepted for presentation at the European Conference on Machine Learning (ECML), specifically at the Machine Learning for Earth Observation Workshop.
SVJLucas/Hydrid-AI-Course
This repository contains the final assessment for the Hybrid AI course at École CentraleSupélec in Paris, France. The final assessment involves a review of the paper titled "Physics-Informed Deep Neural Operator Networks," available at https://arxiv.org/pdf/2207.05748.
SVJLucas/MODA
MODA (Multimodal Object Description Assistant) is an AI model that describes fashion items by combining visual and textual data. Powered by FashionCLIP and OPT, MODA provides descriptions of various fashion objects.
SVJLucas/Physics-Informed-Features-For-Dark-Matter-Morphology
The Physics Informed Features module, created under the international Google Summer of Code program, is a complex tool designed for the generation and manipulation of data in relativistic gravitational lensing studies.
SVJLucas/TrainingSmallVisionLanguageModels
This repository offers a guide and example for training small-scale vision-language models, with a focus on implementing and training MODA (Multimodal Object Description Assistant). MODA is an advanced AI model designed to describe fashion items by effectively combining visual and textual data.
SVJLucas/Cinematch-Backend
Cinematch is a Google Developers Student Club Initiative. Rooted in the heart of collaboration and innovation, our project is not merely about recommending movies. It's about kindling passion in student developers and furnishing them with a hands-on learning experience in backend development.
SVJLucas/Cinematch-Frontend
SVJLucas/classic-watermark-analysis
Test & compare classic frequency domain watermarking techniques: DCT, DWT and DWT-SVD. Analyze strengths, weaknesses, robustness against attacks.
SVJLucas/ImproveLLMStorytellingWithFineTuning
This workshop, led by Lucas José for the Google Developer Student Clubs, focuses on developing practical skills in fine-tuning smaller Language Models (LLMs) to enhance their storytelling capabilities. The aim is to transfer knowledge from a larger model (Meta Llama 2-70B) to a smaller one (Microsoft Phi 1-1B), ensuring high-quality narrative.
SVJLucas/IsVoldemortAGorilla
SVJLucas/ML4SCI.github.io
Machine Learning for Science github site
SVJLucas/OCR-To-Solve-Linear-Systems
The system works by taking a image of a set of linear equations and using OCR technology to extract the equations from the images. The system then applies machine learning algorithms to parse and analyze the equations, determining the type of equation and the appropriate method to solve it.
SVJLucas/self-supervised-watermarking
Analyzing Self-Supervised Models: CLIP, DiNOv2, and ViTMAE in the Context of Zero-Watermarking
SVJLucas/Cracking-the-coding-interview
This repository contains my Python solutions to problems from "Cracking the Coding Interview" by Gayle Laakmann McDowell. Aimed at helping software engineers prepare for interviews, each solution includes code comments and time complexity analysis.
SVJLucas/Decision-Systems
Practical Work for the Decision Systems and Preferences class of Msc AI at CentraleSupélec
SVJLucas/DeepLense
SVJLucas/GDSC-Mastering-AI-Data-Fine-Tuning-Techniques-Workshop
Are you interested in Large Language Models (LLMs) like Bard or ChatGPT? Have you ever wondered how to make them understand your data? Whether you're a novice or an AI enthusiast, this workshop is the perfect introduction to fine-tuning techniques such as PEFT, LoRA, and QLoRA, and how to use these techniques to create your own customizable model.
SVJLucas/GraphMining
SVJLucas/IsVoldemortAGorilla-
SVJLucas/mangrove-world-analysis
SVJLucas/MTP
SVJLucas/NeuralSchrodinger-Solving-the-Schrodinger-Equation-using-Neural-Networks
This repository offers a collection of Jupyter notebooks and scripts for utilizing neural networks to solve the Schrödinger equation, a critical concept in quantum mechanics that describes the behavior of quantum systems.
SVJLucas/SegundaEntregaCES22
Primeira atividade do segundo bimestre CES22/2021
SVJLucas/SVJLucas