-
Notifications
You must be signed in to change notification settings - Fork 0
/
run_antwerp_count.sh
executable file
·28 lines (21 loc) · 1.42 KB
/
run_antwerp_count.sh
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
#!/bin/bash
trap "exit" INT
camera=${1?Error: which camera? CAM11_1, CAM14_4}
date=${2?Error: which date will i do the detection? Nov_28_2020}
opt=${3?Error: which stage do I want to evaluate? detect, track_count, detect_track_count}
# ***************************************************************************#
# User Defined Variables #
# ***************************************************************************#
# ckpt=checkpoints/antwerp/CAM11_1_tr_Oct_02_2020_subtractBG_False_head_False_lr_0.0010_efficientdet-d2_24_22425.pth
ckpt=/home/jovyan/bo/exp_data/student_model_D7/D2/CAM11_1_Nov_27_28_2020_lr_0.0050_decay_cosine_optim_adamw_subtractBG_False_head_False_version_102/efficientdet-d2_44_18270.pth # can either be a pre-trained Efficientdet model or the best student model
framepath=/home/jovyan/bo/dataset
compound_coef=2 # used component coefficient
skip=2 # read every %d frame
ooi=ped_car # object of interest, ped_car/person/car/
# ***************************************************************************#
# Run Detection, Tracking, Counting #
# ***************************************************************************#
for seq in {0..12}
do
python3 inference_antwerpen.py --frame_path $framepath --camera $camera --date $date --sequence $seq --compound_coef $compound_coef --skip $skip --ooi $ooi --program $opt
done