BioFF 0.2.0

基于Forward-Forward算法的生物信息学分类工具（适配基因/蛋白质表达数据）

BioFF BioFF (Bioinformatics Forward-Forward) is a classification tool tailored for bioinformatics tasks (e.g., normal/tumor sample classification) based on the Forward-Forward algorithm, optimized for gene/protein expression data. Installation

Acknowledgements & Open Source Credits

The core implementation and code structure of this project are based on the open-source repository below. We sincerely thank the original author for their contributions:

• Repository Name: pytorch_forward_forwar

Author: mpezeshki

Repository URL: https://github.com/mpezeshki/pytorch_forward_forwar Notes This project is developed and improved based on the aforementioned open-source repository, following open-source protocols and respecting the original author's work. bash 运行

Source code installation

git clone https://github.com/smallWhite-0124/BioFF.git cd bioff pip install . Quick Start

1. Data Preparation Prepare two txt files (one sample per line, the last column is integer label): good_samples.txt: Positive samples (label 0) bad_samples.txt: Negative samples (label 1)
2. One-click Run python 运行 from bioff import run_prediction

Core call

model, results = run_prediction( good_path="good_samples.txt", bad_path="bad_samples.txt" )

Check accuracy

print("Test set accuracy:", results["accuracy"]) Notes Input txt files must be 2D matrices with consistent feature numbers per line Automatically adapts to CPU/GPU, no manual configuration required License MIT