thibaud-perrin Goto Github PK

archi-pypwa

PyScript Demo https://thibaud-perrin.github.io/archi-pypwa/

art-generation-nst

This Python notebook presents a Neural Style Transfer project, an image processing technique based on neural networks. It allows you to combine the styles of a source image (style reference image) with the content of a target image to generate a new image.

assistant-llm

bauhaus-generator

Generate bauhaus wallpaper from an uploaded image

bear-sheep-hunter-env

This project aims to implement a Q-learning algorithm to solve an environment with 3 blobs: a bear, a sheep, and a hunter. The bear has to catch the sheep while avoiding the hunter, both of which move randomly. The bear moves based on Q-learning.

box2d

This project contains the implementation of reinforcement learning algorithms to solve the Lunar Lander and Bipedal Walker environments using the DQN and DDPG algorithms respectively.

car-detection-yolo

This project applies YOLOV2 for car detection using the Keras version from allanzelener/YAD2K.

champagne-champion-ouy-directus

This project is the CMS headless used to monitor the website champagne champion ouy

classic-control

This project demonstrates the use of Q-learning and Deep Q-Networks (DQN) to solve several classic control environments provided by OpenAI Gym. The project includes the following Jupyter notebooks

commonlit-evaluate-student-summaries

diffusion-model

hibo-mistral-7b-fc

Dataset and model fine-tuning for function calling

kaggle-titanic

Jupyter notebook for the Titanic Kaggle ML competition. https://www.kaggle.com/competitions/titanic/data

langchain-demo

mini-gpt

The goal of this project was to implement the encoder only transformer in order to recreate a mini version of GPT.

neural_networks

pyscript-kmeans

https://thibaud-perrin.github.io/PyScript-Kmeans/

qlearning

This project contains two Jupyter notebooks that demonstrate how to implement Q-learning in different environments.

transformer

Transformer model developed from scratch for a translation task. The design is heavily inspired by the original transformer model described in the seminal paper "Attention is All You Need".