search

happinesslz / EPNetV2

EPNet++: Cascade Bi-directional Fusion for Multi-Modal 3D Object Detection (TPAMI-2022)
MIT License
54 stars 5 forks source link

Is the train mask necessary for evaluation purposes? #7

Open shayari21 opened 1 year ago

shayari21 commented 1 year ago

Hello,

i just wish to use the code for inference. Do i also need to download the train_mask for the inference?

happinesslz commented 1 year ago

@shayari21 Hi, we do not use the train_mask or test_mask for the inference. The command for test is following: CUDA_VISIBLE_DEVICES=0 python eval_rcnn.py --cfg_file cfgs/CAR_EPNet_plus_plus_for_test.yaml --eval_mode rcnn_online \ --output_dir ./epnet_plus_plus_released_trained_models/CAR/eval_results/ \ --data_path ../data/ --ckpt ./epnet_plus_plus_released_trained_models/CAR/checkpoint_epoch_43.pth --test \ --set LI_FUSION.ENABLED True LI_FUSION.ADD_Image_Attention True CROSS_FUSION True USE_P2I_GATE True \ DEEP_RCNN_FUSION False USE_IMAGE_LOSS True IMAGE_WEIGHT 1.0 USE_IMAGE_SCORE True

Please remember to modify: 1)TEST.SPLIT: (val ---> test) in configure file 2) add '--test' in the command 3) remove the line 317 "img_seg_mask = self.get_KINS_car_mask(sample_id)" in the file of 'kitti_rcnn_dataset.py' to the above line of 526.

shayari21 commented 1 year ago

Hi, thanks alot for your response. So if i wish to evaluate, can the above change work too?