-
Notifications
You must be signed in to change notification settings - Fork 47
/
Copy pathtrain_monosf_selfsup_kitti_raw.sh
53 lines (47 loc) · 1.44 KB
/
train_monosf_selfsup_kitti_raw.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
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
#!/bin/bash
# experiments and datasets meta
KITTI_RAW_HOME=""
EXPERIMENTS_HOME=""
# model
MODEL=MonoSceneFlow_fullmodel
# save path
ALIAS="-kitti-"
TIME=$(date +"%Y%m%d-%H%M%S")
SAVE_PATH="$EXPERIMENTS_HOME/$MODEL$ALIAS$TIME"
CHECKPOINT=None
# Loss and Augmentation
Train_Dataset=KITTI_Raw_KittiSplit_Train_mnsf
Train_Augmentation=Augmentation_SceneFlow
Train_Loss_Function=Loss_SceneFlow_SelfSup
Valid_Dataset=KITTI_Raw_KittiSplit_Valid_mnsf
Valid_Augmentation=Augmentation_Resize_Only
Valid_Loss_Function=Loss_SceneFlow_SelfSup
# training configuration
python ../main.py \
--batch_size=4 \
--batch_size_val=1 \
--checkpoint=$CHECKPOINT \
--lr_scheduler=MultiStepLR \
--lr_scheduler_gamma=0.5 \
--lr_scheduler_milestones="[23, 39, 47, 54]" \
--model=$MODEL \
--num_workers=16 \
--optimizer=Adam \
--optimizer_lr=2e-4 \
--save=$SAVE_PATH \
--total_epochs=62 \
--training_augmentation=$Train_Augmentation \
--training_augmentation_photometric=True \
--training_dataset=$Train_Dataset \
--training_dataset_root=$KITTI_RAW_HOME \
--training_dataset_flip_augmentations=True \
--training_dataset_preprocessing_crop=True \
--training_dataset_num_examples=-1 \
--training_key=total_loss \
--training_loss=$Train_Loss_Function \
--validation_augmentation=$Valid_Augmentation \
--validation_dataset=$Valid_Dataset \
--validation_dataset_root=$KITTI_RAW_HOME \
--validation_dataset_preprocessing_crop=False \
--validation_key=total_loss \
--validation_loss=$Valid_Loss_Function \