📦 iamstarlee / BiMemNet

This Repo is designed to continuously learn grid data

★ 0 stars ⑂ 0 forks 👁 0 watching

📥 Clone https://github.com/iamstarlee/BiMemNet.git

HTTPS git clone https://github.com/iamstarlee/BiMemNet.git

SSH git clone git@github.com:iamstarlee/BiMemNet.git

CLI gh repo clone iamstarlee/BiMemNet

ZIP Desktop VS Code

Xinxin Li Add files via upload 1e783b8 4 months ago 📝 History

📂 main View all commits →

📁 model_data

📁 nets

📁 utils

📄 eval.py

📄 extract_data.py

📄 get_map.py

📄 get_the_classes.py

📄 get_the_map.py

📄 predict.py

📄 README.md

📄 train.py

📄 voc_annotation.py

📄 yolo.py

📄 README.md

0. 多卡并行训练和推理

CUDA_VISIBLE_DEVICES=2,3,4,5,6,7 python -m torch.distributed.launch --nproc_per_node=6 --master_port=25641 train.py

CUDA_VISIBLE_DEVI=0 python predict.py

需要对网络添加一定数量的生成模块，具体体现在注释掉train.py中333-342行的代码，修改490-491行优化器的参数pg0→reg_params，注释掉493行和494行。