🧠 Medical Image Reconstruction on BraTS Dataset

This repository provides an implementation for medical image reconstruction based on the BraTS dataset. The framework is adapted from MTrans, with support for both single-GPU and multi-GPU training.

---

Installation

# 1. Clone the repository
git https://github.com/script-Yang/BraTS-MTrans.git
cd BraTS-MTrans

# 2. Create and activate environment (recommended)
conda create -n recon python=3.8 -y
conda activate recon

# 3. Install required packages
pip install -r requirements.txt

---

Training Instructions

Single-GPU Training

python train.py --experiment reconstruction_single

Multi-GPU Training

CUDA_VISIBLE_DEVICES=1,2,3 python -m torch.distributed.launch --nproc_per_node=3 train.py

Adjust CUDA_VISIBLE_DEVICES and --nproc_per_node based on your setup.

---

📌 Acknowledgment

This project is based on MTrans. Please cite their work if used.