Skip to content

Latest commit

 

History

History
53 lines (41 loc) · 1.83 KB

STARTED.md

File metadata and controls

53 lines (41 loc) · 1.83 KB
Raw

Getting Started

The dataset configs are located within tools/cfgs/det_dataset_configs, and the model configs are located within tools/cfgs/det_model_configs for different datasets.

Training & Testing

Test and evaluate the pretrained models

  • Test with a pretrained model:
python test.py --cfg_file ${CONFIG_FILE} --batch_size ${BATCH_SIZE} --ckpt ${CKPT}
  • To test all the saved checkpoints of a specific training setting and draw the performance curve on the Tensorboard, add the --eval_all argument:
python test.py --cfg_file ${CONFIG_FILE} --batch_size ${BATCH_SIZE} --eval_all
  • To test with multiple GPUs:
sh scripts/dist_test.sh ${NUM_GPUS} \
    --cfg_file ${CONFIG_FILE} --batch_size ${BATCH_SIZE}

# or

sh scripts/slurm_test_mgpu.sh ${PARTITION} ${NUM_GPUS} \
    --cfg_file ${CONFIG_FILE} --batch_size ${BATCH_SIZE}

Train a model

You could optionally add extra command line parameters --batch_size ${BATCH_SIZE} and --epochs ${EPOCHS} to specify your preferred parameters.

  • KITTI
sh scripts/dist_train.sh ${NUM_GPUS} --cfg_file ${CONFIG_FILE}

# or 

sh scripts/slurm_train.sh ${PARTITION} ${JOB_NAME} ${NUM_GPUS} --cfg_file ${CONFIG_FILE}
  • Waymo (Two stage training schame)
#LiDAR branch
sh scripts/dist_train.sh ${NGPUSLIST} ${NUM_GPUS} --cfg_file ${CONFIG_FILE}
#LoGoNet branch
sh scripts/dist_train_mm.sh ${NGPUSLIST} ${NUM_GPUS} --cfg_file ${CONFIG_FILE} --pretrained_model ${PRETRAINED_MODEL_PATH}
# or 

#LiDAR branch
sh scripts/slurm_train.sh ${PARTITION} ${JOB_NAME} ${NUM_GPUS} --cfg_file ${CONFIG_FILE}
#LoGoNet branch
sh scripts/slurm_train_mm.sh ${PARTITION} ${JOB_NAME} ${NUM_GPUS} --cfg_file ${CONFIG_FILE} --pretrained_model ${PRETRAINED_MODEL_PATH}