OBUP Model Weights

Welcome to the OBUP model repository! This repository hosts the pre-trained models and weight files designed to be used with the open-source GitHub project OBUP.

Introduction

This repository contains the necessary model files to run the OBUP project. The core files are stored in the model_merged directory, which includes the best weights trained on specific datasets, such as the Stage 1 model trained on ABCD, HCP_A, and HCP_D data.

• GitHub Repository: ZhibinHe/OBUP
• Framework: PyTorch

File Structure

The core file structure of this repository is as follows:

OBUP/ └── model_merged/ ├── best_stage1_model_train_ABCD,HCP_A,HCP_D.pth # Best Stage 1 training weights └── [Other potential model or config files]

How to Use

The core file structure of this repository is as follows:

To properly run the code, please follow these steps to integrate the model files from this repository with the source code on GitHub:

1. Clone the GitHub Repository First, clone the official code repository from GitHub: git clone [https://github.com/ZhibinHe/OBUP.git](https://github.com/ZhibinHe/OBUP.git) cd OBUP

2. Download the Model Files You can download the model_merged folder directly via the web interface, or use the huggingface-cli for an efficient download: pip install -U huggingface_hub huggingface-cli download zhibinhe/OBUP --local-dir ./OBUP_models

3. Place the Model Files Copy or move the entire downloaded model_merged folder directly into the root directory of the cloned OBUP repository. Your final directory structure should look like this: OBUP/ <-- GitHub code root directory ├── model_merged/ <-- Folder downloaded from Hugging Face │ └── best_stage1_model_train_ABCD,HCP_A,HCP_D.pth ├── [Source code files] ├── README.md └── requirements.txt

4. Run the Code Once the models are in place, you can follow the instructions provided in the OBUP GitHub README to run the experiments.

Training Data

The Stage 1 model was trained using the following neuroimaging datasets:

1. ABCD (Adolescent Brain Cognitive Development Study)

2. HCP-A (Human Connectome Project - Aging)

3. HCP-D (Human Connectome Project - Development)

4. HCP-YA (Human Connectome Project - Young Adults)