remote_sensing_segmentation/script_river.sh at master · keillernogueira/remote_sensing_segmentation · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38

# FCN
CUDA_VISIBLE_DEVICES=0 python3 main.py --operation training --output_path /home/kno/remote_sensing_segmentation/output_arvore/fcn/256/ \
 --dataset arvore --dataset_input_path /home/kno/datasets/arvore/ --num_classes 2 --model_name fcn_50_3_2x --batch_size 16 \
 --reference_crop_size 256 --reference_stride_crop 256 --values 256 --distribution_type single_fixed

##################################################
# SEGNET
CUDA_VISIBLE_DEVICES=1 python3 main.py --operation training --output_path /home/kno/remote_sensing_segmentation/output_river/segnet/256/ \
--dataset river --dataset_input_path /home/kno/bacia/Bacia_SantoAntonio_Revisado/ --num_classes 2 --model_name segnet --batch_size 16 \
--reference_crop_size 256 --reference_stride_crop 256 --values 256 --distribution_type single_fixed

##########################################

# UNET
CUDA_VISIBLE_DEVICES=0 python3 main.py --operation training --output_path /home/kno/remote_sensing_segmentation/output/unet/32/ \
--dataset_input_path /home/kno/datasets/arvore/ \
--dataset_gt_path /home/kno/dataset_laranjal/Dataset_Laranjal/Arvore_Segmentacao\ \(Sequoia\)/sequoia_raster.tif \
--num_classes 2 --model_name unet --values 32 --reference_crop_size 32 --reference_stride_crop 20 --distribution_type single_fixed

# TGRS
CUDA_VISIBLE_DEVICES=0 python3 main.py --operation training --output_path /home/kno/remote_sensing_segmentation/output/ddcnn/25_50_v2/ \
--dataset_input_path /home/kno/datasets/arvore/ \
--dataset_gt_path /home/kno/dataset_laranjal/Dataset_Laranjal/Arvore_Segmentacao\ \(Sequoia\)/sequoia_raster.tif \
--num_classes 2 --model_name dilated_grsl_rate8 --reference_crop_size 32 --reference_stride_crop 20 --values 32,64 \
--distribution_type multi_fixed

# DEEPLAB
CUDA_VISIBLE_DEVICES=0 python3 main.py --operation training --output_path /home/kno/remote_sensing_segmentation/output/deeplab/25/ \
--dataset_input_path /home/kno/datasets/arvore/ \
--dataset_gt_path /home/kno/dataset_laranjal/Dataset_Laranjal/Arvore_Segmentacao\ \(Sequoia\)/sequoia_raster.tif \
--num_classes 2 --model_name deeplabv3+ --values 32 --reference_crop_size 32 --reference_stride_crop 20 --distribution_type single_fixed

# PIXELWISE
CUDA_VISIBLE_DEVICES=0 python3 main.py --operation training --output_path /home/kno/remote_sensing_segmentation/output/pixelwise/32_48/ \
--dataset_input_path /home/kno/datasets/arvore/ \
--dataset_gt_path /home/kno/dataset_laranjal/Dataset_Laranjal/Arvore_Segmentacao\ \(Sequoia\)/sequoia_raster.tif \
--num_classes 2 --model_name pixelwise --values 32 --reference_crop_size 32 --reference_stride_crop 1 --distribution_type single_fixed