kittiviewer/backend/main.py main -port=xxxx in your server/local. pytorch/train.py evaluate -config_path=./configs/car.config -model_dir=/root/model/car Nvidia-docker run -it -rm -v /media/yy/960evo/datasets/:/root/data -v $HOME/pretrained_models:/root/model -ipc=host second-pytorch:latest
Training process use batchsize=6 as default for 1080Ti, you need to reduce batchsize if your GPU has less memory.Ĭurrently only support single GPU training, but train a model only needs 20 hours (165 epoch) in a single 1080Ti and only needs 50 epoch to reach 78.3 AP with super converge in car moderate 3D in Kitti validation dateset. A new directory will be created if the model_dir doesn't exist, otherwise will read checkpoints in it. Make sure "/path/to/model_dir" doesn't exist if you want to train new model. Modify config file, set enable_mixed_precision to true. For example, 50 epochs = 15500 steps for and single GPU, if you use 4 GPUs, you need to divide steps and steps_per_eval by 4. You need to modify total step in config file. Note: The batch_size and num_workers in config file is per-GPU, if you use multi-gpu, they will be multiplied by number of GPUs. pytorch/train.py train -config_path=./configs/ -model_dir=/path/to/model_dir -multi_gpu=True
0 kommentar(er)
