-
Notifications
You must be signed in to change notification settings - Fork 0
/
test.sh
executable file
·54 lines (31 loc) · 6.71 KB
/
test.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
# modify the variables below
model_name=bit_resnet50
coco_path=/localtmp/data/coco2017/coco_dataset/
coco_analysis_path=/localtmp/data/coco2017/coco_dataset/
oi_analysis_path=../VisionResearch/finetuneClip/openimages_dataset/val/
#TODO: Add three variables --> openimages trial path, coco_trial path, checkpoint path
# Feature 1
CUDA_VISIBLE_DEVICES=0 python train.py --model_name $model_name --dataset coco --num_classes 80 --extract_cross_analysis_features --analysis_set coco --analysis_set_path $coco_analysis_path --config_file config/coco.yaml --trial_path experiments/coco/bit_resnet50/2022-01-21\ 19\:03\:29/ --finetune --bias_analysis > /dev/null 2>&1 && echo SUCCESS || echo FAIL
CUDA_VISIBLE_DEVICES=0 python train.py --model_name $model_name --dataset coco --num_classes 80 --extract_cross_analysis_features --analysis_set openimages --analysis_set_path $oi_analysis_path --config_file config/openimages.yaml --trial_path experiments/coco/bit_resnet50/2022-01-21\ 19\:03\:29/ --finetune --bias_analysis > /dev/null 2>&1 && echo SUCCESS || echo FAIL
CUDA_VISIBLE_DEVICES=0 python train.py --model_name $model_name --dataset openimages --num_classes 601 --extract_cross_analysis_features --analysis_set coco --analysis_set_path $coco_analysis_path --config_file config/coco.yaml --trial_path experiments/openimages/bit_resnet50/2022-02-23\ 11\:44\:36/ --finetune --bias_analysis > /dev/null 2>&1 && echo SUCCESS || echo FAIL
CUDA_VISIBLE_DEVICES=0 python train.py --model_name $model_name --dataset openimages --num_classes 601 --extract_cross_analysis_features --analysis_set openimages --analysis_set_path $oi_analysis_path --config_file config/openimages.yaml --trial_path experiments/openimages/bit_resnet50/2022-02-23\ 11\:44\:36/ --finetune --bias_analysis > /dev/null 2>&1 && echo SUCCESS || echo FAIL
# Feature 2
CUDA_VISIBLE_DEVICES=0 python train.py --model_name $model_name --dataset coco --num_classes 80 --pretrained_features --analysis_set coco --analysis_set_path $coco_analysis_path --config_file config/coco.yaml --trial_path experiments/coco/bit_resnet50/2022-01-21\ 19\:03\:29/ --finetune --bias_analysis > /dev/null 2>&1 && echo SUCCESS || echo FAIL
CUDA_VISIBLE_DEVICES=0 python train.py --model_name $model_name --dataset coco --num_classes 80 --pretrained_features --analysis_set openimages --analysis_set_path $oi_analysis_path --config_file config/openimages.yaml --trial_path experiments/coco/bit_resnet50/2022-01-21\ 19\:03\:29/ --finetune --bias_analysis > /dev/null 2>&1 && echo SUCCESS || echo FAIL
CUDA_VISIBLE_DEVICES=0 python train.py --model_name $model_name --dataset openimages --num_classes 601 --pretrained_features --analysis_set coco --analysis_set_path $coco_analysis_path --config_file config/coco.yaml --trial_path experiments/openimages/bit_resnet50/2022-02-23\ 11\:44\:36/ --finetune --bias_analysis > /dev/null 2>&1 && echo SUCCESS || echo FAIL
CUDA_VISIBLE_DEVICES=0 python train.py --model_name $model_name --dataset openimages --num_classes 601 --pretrained_features --analysis_set openimages --analysis_set_path $oi_analysis_path --config_file config/openimages.yaml --trial_path experiments/openimages/bit_resnet50/2022-02-23\ 11\:44\:36/ --finetune --bias_analysis > /dev/null 2>&1 && echo SUCCESS || echo FAIL
# Feature 3
CUDA_VISIBLE_DEVICES=0 python train.py --model_name $model_name --dataset coco --num_classes 80 --load_features --analysis_set coco --analysis_set_path $coco_analysis_path --config_file config/coco.yaml --trial_path experiments/coco/bit_resnet50/2022-01-21\ 19\:03\:29/ --finetune --bias_analysis > /dev/null 2>&1 && echo SUCCESS || echo FAIL
CUDA_VISIBLE_DEVICES=0 python train.py --model_name $model_name --dataset coco --num_classes 80 --load_features --analysis_set openimages --analysis_set_path $oi_analysis_path --config_file config/openimages.yaml --trial_path experiments/coco/bit_resnet50/2022-01-21\ 19\:03\:29/ --finetune --bias_analysis > /dev/null 2>&1 && echo SUCCESS || echo FAIL
CUDA_VISIBLE_DEVICES=0 python train.py --model_name $model_name --dataset openimages --num_classes 601 --load_features --analysis_set coco --analysis_set_path $coco_analysis_path --config_file config/coco.yaml --trial_path experiments/openimages/bit_resnet50/2022-02-23\ 11\:44\:36/ --finetune --bias_analysis > /dev/null 2>&1 && echo SUCCESS || echo FAIL
CUDA_VISIBLE_DEVICES=0 python train.py --model_name $model_name --dataset openimages --num_classes 601 --load_features --analysis_set openimages --analysis_set_path $oi_analysis_path --config_file config/openimages.yaml --trial_path experiments/openimages/bit_resnet50/2022-02-23\ 11\:44\:36/ --finetune --bias_analysis > /dev/null 2>&1 && echo SUCCESS || echo FAIL
# Feature 4
CUDA_VISIBLE_DEVICES=0 python train.py --model_name $model_name --dataset coco --num_classes 80 --load_features --multiple_trials --analysis_set coco --analysis_set_path $coco_analysis_path --config_file config/coco.yaml --finetune --bias_analysis > /dev/null 2>&1 && echo SUCCESS || echo FAIL
CUDA_VISIBLE_DEVICES=0 python train.py --model_name $model_name --dataset coco --num_classes 80 --load_features --multiple_trials --analysis_set openimages --analysis_set_path $oi_analysis_path --config_file config/openimages.yaml --finetune --bias_analysis > /dev/null 2>&1 && echo SUCCESS || echo FAIL
CUDA_VISIBLE_DEVICES=0 python train.py --model_name $model_name --dataset openimages --num_classes 601 --load_features --multiple_trials --analysis_set coco --analysis_set_path $coco_analysis_path --config_file config/coco.yaml --finetune --bias_analysis > /dev/null 2>&1 && echo SUCCESS || echo FAIL
CUDA_VISIBLE_DEVICES=0 python train.py --model_name $model_name --dataset openimages --num_classes 601 --load_features --multiple_trials --analysis_set openimages --analysis_set_path $oi_analysis_path --config_file config/openimages.yaml --finetune --bias_analysis > /dev/null 2>&1 && echo SUCCESS_DONE || echo FAIL
# Feature 5 --> supports testing finetuning on COCO
#CUDA_VISIBLE_DEVICES=0 python train.py --model_name $model_name --dataset coco --dataset_path $coco_path --batch_size 64 --epochs 1 --num_classes 80 --analysis_set coco --analysis_set_path $coco_analysis_path --config_file config/coco.yaml --bias_analysis --finetune > /dev/null 2>&1 && echo SUCCESS || echo FAIL
# Feature 6 --> supports testing resuming training on COCO
#CUDA_VISIBLE_DEVICES=0 python train.py --model_name $model_name --dataset coco --dataset_path $coco_path --batch_size 64 --epochs 1 --num_classes 80 --analysis_set coco --analysis_set_path $coco_analysis_path --config_file config/coco.yaml --bias_analysis --finetune --resume_training --checkpoint experiments/coco/bit_resnet50/2022-04-04\ 16\:50\:19/model/bit_resnet50/version_0/checkpoints/epoch\=0-step\=1848.ckpt > /dev/null 2>&1 && echo SUCCESS || echo FAIL