https://github.com/chenliu-1996/brainfm

https://github.com/chenliu-1996/brainfm

Science Score: 26.0%

This score indicates how likely this project is to be science-related based on various indicators:

  • CITATION.cff file
  • codemeta.json file
    Found codemeta.json file
  • .zenodo.json file
    Found .zenodo.json file
  • DOI references
  • Academic publication links
  • Academic email domains
  • Institutional organization owner
  • JOSS paper metadata
  • Scientific vocabulary similarity
    Low similarity (7.6%) to scientific vocabulary
Last synced: 10 months ago · JSON representation

Repository

Basic Info
  • Host: GitHub
  • Owner: ChenLiu-1996
  • Language: Jupyter Notebook
  • Default Branch: main
  • Size: 585 MB
Statistics
  • Stars: 1
  • Watchers: 1
  • Forks: 0
  • Open Issues: 0
  • Releases: 0
Created over 1 year ago · Last pushed 10 months ago
Metadata Files
Readme

Readme.md

Brain Foundation Model

Krishnaswamy Lab, Yale University

Twitter

Steps to reproduce

  1. Data preparation. If you work on Misha, you can find the preprocessed data at /gpfs/radev/home/cl2482/project/BrainFM/data/Dynamic_Natural_Vision.

1.1 Download Dynamic Natural Vision data. cd src/data_download python download_DNV.py

1.2 Unzip (sorry this step is messy) The desired structure is: data/DynamicNaturalVision/{subject1,subject2,subject3,video} Under {subject1,subject2,subject3} there are {fmri/smri} folders. Under {video} there are {seg1.mp4,seg2.mp4,...,test5.mp4}.

1.3 Preprocess the video. This will create a {videoframes} folder. ``` cd src/preprocessing python preprocessDNV_videos.py ```

1.4 Preprocess the fMRI. This will create a {fMRIscans} folder. ``` cd src/preprocessing python preprocessDNV_fmri.py ```

  1. Train. cd src/ python main.py --batch-size 4 --max-training-iters 64 --max-validation-iters 64

Dependencies

We developed the codebase in a miniconda environment. How we created the conda environment: ```

Optional: Update to libmamba solver.

conda update -n base conda conda install -n base conda-libmamba-solver conda config --set solver libmamba

conda create --name brainfm pytorch==2.1.0 torchvision torchaudio pytorch-cuda=12.1 -c pytorch -c nvidia -c anaconda -c conda-forge -y conda activate brainfm conda install scikit-learn scikit-image pillow matplotlib seaborn tqdm -c pytorch -c anaconda -c conda-forge -y conda install accelerate -c conda-forge -y conda install nibabel -y python -m pip install opencv-python python -m pip install torchgeometric einops python -m pip install git+https://github.com/openai/CLIP.git python -m pip install kornia open-clip-torch omegaconf python -m pip install diffusers["torch"]==0.21.4 transformers huggingfacehub==0.25.2 python -m pip install xformers==0.0.22.post7 python -m pip install dalle2-pytorch==1.15.6 python -m pip install natsort python -m pip install lpips

For NeuroClips

python -m pip install webdataset pytorch-lightning ```

Owner

  • Name: Chen Liu
  • Login: ChenLiu-1996
  • Kind: user
  • Location: New Haven
  • Company: Yale University

CS PhD student at @KrishnaswamyLab, @YaleUniversity. Reviewing Committee member at NeurIPS, ICLR, ICML.

GitHub Events

Total
  • Watch event: 1
  • Push event: 46
  • Public event: 1
Last Year
  • Watch event: 1
  • Push event: 46
  • Public event: 1