Dual - perspective mamba unmixing network
本项目现已集成网格搜索与贝叶斯优化的超参数搜索方案,支持自动寻找最优模型配置。
This project now includes integrated grid search and Bayesian optimization for hyperparameter tuning, supporting automatic discovery of optimal model configurations.
# 创建默认配置文件 / Create default configuration files
python hyperopt.py
# 网格搜索 / Grid Search
python run_hyperopt.py --method grid --dataset jasper
# 贝叶斯优化 / Bayesian Optimization
python run_hyperopt.py --method bayes --trials 50 --dataset jasper
# Optuna优化 / Optuna Optimization
python run_hyperopt.py --method optuna --trials 50 --dataset jasper请查看 HYPEROPT_GUIDE.md 获取完整的使用指南。
See HYPEROPT_GUIDE.md for complete usage guide.
快速演示:
Quick demo:
bash quick_demo.sh完整示例:
Full example:
bash example_hyperopt.sh原始训练方式依然可用:
Original training method is still available:
python train.py --dataset jasper --epochs 50 --batch_size 64