Vjekoslav Drvar (vdrvar)

vdrvar

Geek Repo

Company:CROZ AI

Location:Zagreb, Croatia

Home Page:https://croz.ai/

Twitter:@v_drvar

Github PK Tool:Github PK Tool

Vjekoslav Drvar's repositories

bitcoin_fraud_detection

Fraud detection in Bitcoin transactions using GCNs. Uses C++ for data preprocessing and Python for training. Features include data processing, graph creation, and visualization with PyTorch Geometric.

Language:Jupyter NotebookLicense:MITStargazers:8Issues:0Issues:0

euro2024_nlp_predictor

NLP project forecasting Euro 2024 outcomes using sentiment analysis of football news. Includes automated scraping and sentiment analysis. Utilizes Python, Beautiful Soup, and Transformers.

Language:Jupyter NotebookStargazers:7Issues:1Issues:0

java_ml_text_processor

A Java program demonstrating natural language processing with Apache OpenNLP and JavaFX, offering an intuitive GUI designed for educational use and interaction.

Language:JavaStargazers:4Issues:0Issues:0

spotify_recommender_system

Flask app offering personalized song recommendations using cosine and RBF similarity. Includes song exploration, tailored recommendations, a favorites list, Redis caching, and Dockerization.

Language:PythonStargazers:4Issues:0Issues:0

comp_vision_for_soda_bottles

Soda bottle recognition using transfer learning. Achieves 94% accuracy with a fine-tuned Xception model. Features a FastAPI interface for image uploads and predictions. Uses Python and TensorFlow.

Language:Jupyter NotebookStargazers:3Issues:0Issues:0

julia_image_autoencoder

A Julia application using Flux.jl to construct an autoencoder for image compression, processing multiple images to enhance feature recognition and reconstruction quality.

Language:JuliaStargazers:3Issues:0Issues:0

seir_spatial_outbreak_simulation

Designed and implemented an epidemiological model in C++ using stochastic modelling. Such models generalize classical SEIR ones that are only defined via ODEs, without considering spatial relations.

Language:C++Stargazers:1Issues:1Issues:0

signature_randomization_differential-equations

Theoretically improved and implemented a method for learning rough path trajectories using multi-dimensional analysis, neural differential equations and randomization.

Language:Jupyter NotebookStargazers:1Issues:0Issues:0

sport_certs

In this repository, you can find my certificates and achievements in marathons and half marathons.

Stargazers:0Issues:0Issues:0