-
Notifications
You must be signed in to change notification settings - Fork 0
/
run_experiment_ablation_positions_as_target.sh
executable file
·178 lines (161 loc) · 13.3 KB
/
run_experiment_ablation_positions_as_target.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
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
#!/bin/bash
# arg1: (Default:venv3) name of the virtual environment (will be stored in the main directory of this project). Default is stored in the file ./venv . if problems occur, change the name in the file
# arg2: (Default:0) the number of the gpu to use. This script does not support multiple gpus
# get the virtual environment name from the first argument to the script. if none is given, use the last virtual environment.
# change the first line in the ../venv file to match the default virtual environment name if needed
virtual_env_name=${1:-$(<.././venv)}
rm .././venv
echo "$virtual_env_name"
echo "$virtual_env_name">>.././venv
echo "Requested virtual environment: $virtual_env_name"
### DEPRECATED: In this experiment, we ablate the retinanet and transformer auxiliary output branches (with the target as the x,y,z and joint angles as block classes)
# get dir
dir=`dirname "$BASH_SOURCE"`
scripts_dir=`pwd`
project_dir="$(dirname "$scripts_dir")"
echo "The project directory: " $project_dir
cd $project_dir
echo "Activating virtualenv"
source $virtual_env_name"/bin/activate"
# inside the virtual environment, run all setups
if [[ $(basename "$VIRTUAL_ENV") == "$virtual_env_name" ]]; then
echo "Running experiments"
# export gpu
export CUDA_VISIBLE_DEVICES=${2:-0}
# setup all the packages
cd $project_dir/modules/data_fusion/fusionnet/keras_fusionnet/bin/
# this is a test script
python3 train.py --steps=10 --val-steps=5 --batch-size=1 --epochs=1 --experiment-tag="SANITY_CHECK_TEST" \
--comet-api-key="$COMET_API_KEY" --comet-project-name="ir_ablation_pos_exp" --comet-workspace="$COMET_WORKSPACE" \
--no-snapshots \
fusionnet $DS_PATH/GeneratedData/FusionMultimodalCSV/virtual/Annotations/train.csv \
--val-annotations $DS_PATH/GeneratedData/FusionMultimodalCSV/virtual/Annotations/val.csv \
--output-filter="retinanet_regression,retinanet_classification,transformer_classification,fusionnet_regression" \
vision retinanet $DS_PATH/GeneratedData/VisionMultimodalCSV/virtual/Annotations/train.csv \
$DS_PATH/GeneratedData/VisionMultimodalCSV/virtual/Annotations/classmaps.csv \
--val-annotations $DS_PATH/GeneratedData/VisionMultimodalCSV/virtual/Annotations/val.csv \
language_translation transformer $DS_PATH/GeneratedData/LanguageTranslationMultimodalCSV/virtual/Annotations/train.csv \
--vocab $DS_PATH/GeneratedData/LanguageTranslationMultimodalCSV/virtual/Annotations/vocab.lst \
--val-annotations $DS_PATH/GeneratedData/LanguageTranslationMultimodalCSV/virtual/Annotations/val.csv \
--val-golden-set $DS_PATH/GeneratedData/LanguageTranslationMultimodalCSV/virtual/Annotations/val_golden.txt
# perform 3 rounds
for i in {1..3}
do
echo "ROUND $i \n\n"
# the full network
python3 train.py --steps=5000 --val-steps=500 --epochs=10 --batch-size=1 --experiment-tag="FULL" \
--comet-api-key="$COMET_API_KEY" --comet-project-name="ir_ablation_pos_exp" --comet-workspace="$COMET_WORKSPACE" \
--no-snapshots \
fusionnet $DS_PATH/GeneratedData/FusionMultimodalCSV/virtual/Annotations/train.csv \
--val-annotations $DS_PATH/GeneratedData/FusionMultimodalCSV/virtual/Annotations/val.csv \
--output-filter="retinanet_regression,retinanet_classification,transformer_classification,fusionnet_regression" \
vision retinanet $DS_PATH/GeneratedData/VisionMultimodalCSV/virtual/Annotations/train.csv \
$DS_PATH/GeneratedData/VisionMultimodalCSV/virtual/Annotations/classmaps.csv \
--val-annotations $DS_PATH/GeneratedData/VisionMultimodalCSV/virtual/Annotations/val.csv \
language_translation transformer $DS_PATH/GeneratedData/LanguageTranslationMultimodalCSV/virtual/Annotations/train.csv \
--vocab $DS_PATH/GeneratedData/LanguageTranslationMultimodalCSV/virtual/Annotations/vocab.lst \
--val-annotations $DS_PATH/GeneratedData/LanguageTranslationMultimodalCSV/virtual/Annotations/val.csv \
--val-golden-set $DS_PATH/GeneratedData/LanguageTranslationMultimodalCSV/virtual/Annotations/val_golden.txt
# no transformer output, no transformer validation
python3 train.py --steps=5000 --val-steps=500 --epochs=10 --batch-size=1 --experiment-tag="NO_TRANS" \
--comet-api-key="$COMET_API_KEY" --comet-project-name="ir_ablation_pos_exp" --comet-workspace="$COMET_WORKSPACE" \
--no-snapshots \
fusionnet $DS_PATH/GeneratedData/FusionMultimodalCSV/virtual/Annotations/train.csv \
--val-annotations $DS_PATH/GeneratedData/FusionMultimodalCSV/virtual/Annotations/val.csv \
--output-filter="retinanet_regression,retinanet_classification,fusionnet_regression" \
vision retinanet $DS_PATH/GeneratedData/VisionMultimodalCSV/virtual/Annotations/train.csv \
$DS_PATH/GeneratedData/VisionMultimodalCSV/virtual/Annotations/classmaps.csv \
--val-annotations $DS_PATH/GeneratedData/VisionMultimodalCSV/virtual/Annotations/val.csv \
language_translation transformer $DS_PATH/GeneratedData/LanguageTranslationMultimodalCSV/virtual/Annotations/train.csv \
--vocab $DS_PATH/GeneratedData/LanguageTranslationMultimodalCSV/virtual/Annotations/vocab.lst \
--val-annotations $DS_PATH/GeneratedData/LanguageTranslationMultimodalCSV/virtual/Annotations/val.csv \
--val-golden-set $DS_PATH/GeneratedData/LanguageTranslationMultimodalCSV/virtual/Annotations/val_golden.txt --no-evaluation
# no transformer output, no transformer validation, no retinanet classification, no retinanet validation
python3 train.py --steps=5000 --val-steps=500 --epochs=10 --batch-size=1 --experiment-tag="NO_TRANS_NO_RET_CLAS" \
--comet-api-key="$COMET_API_KEY" --comet-project-name="ir_ablation_pos_exp" --comet-workspace="$COMET_WORKSPACE" \
--no-snapshots \
fusionnet $DS_PATH/GeneratedData/FusionMultimodalCSV/virtual/Annotations/train.csv \
--val-annotations $DS_PATH/GeneratedData/FusionMultimodalCSV/virtual/Annotations/val.csv \
--output-filter="retinanet_regression,fusionnet_regression" \
vision retinanet $DS_PATH/GeneratedData/VisionMultimodalCSV/virtual/Annotations/train.csv \
$DS_PATH/GeneratedData/VisionMultimodalCSV/virtual/Annotations/classmaps.csv \
--val-annotations $DS_PATH/GeneratedData/VisionMultimodalCSV/virtual/Annotations/val.csv --no-evaluation \
language_translation transformer $DS_PATH/GeneratedData/LanguageTranslationMultimodalCSV/virtual/Annotations/train.csv \
--vocab $DS_PATH/GeneratedData/LanguageTranslationMultimodalCSV/virtual/Annotations/vocab.lst \
--val-annotations $DS_PATH/GeneratedData/LanguageTranslationMultimodalCSV/virtual/Annotations/val.csv \
--val-golden-set $DS_PATH/GeneratedData/LanguageTranslationMultimodalCSV/virtual/Annotations/val_golden.txt --no-evaluation
# no transformer output, no transformer validation, no retinanet regression, no retinanet validation
python3 train.py --steps=5000 --val-steps=500 --epochs=10 --batch-size=1 --experiment-tag="NO_TRANS_NO_RET_REG" \
--comet-api-key="$COMET_API_KEY" --comet-project-name="ir_ablation_pos_exp" --comet-workspace="$COMET_WORKSPACE" \
--no-snapshots \
fusionnet $DS_PATH/GeneratedData/FusionMultimodalCSV/virtual/Annotations/train.csv \
--val-annotations $DS_PATH/GeneratedData/FusionMultimodalCSV/virtual/Annotations/val.csv \
--output-filter="retinanet_classification,fusionnet_regression" \
vision retinanet $DS_PATH/GeneratedData/VisionMultimodalCSV/virtual/Annotations/train.csv \
$DS_PATH/GeneratedData/VisionMultimodalCSV/virtual/Annotations/classmaps.csv \
--val-annotations $DS_PATH/GeneratedData/VisionMultimodalCSV/virtual/Annotations/val.csv --no-evaluation \
language_translation transformer $DS_PATH/GeneratedData/LanguageTranslationMultimodalCSV/virtual/Annotations/train.csv \
--vocab $DS_PATH/GeneratedData/LanguageTranslationMultimodalCSV/virtual/Annotations/vocab.lst \
--val-annotations $DS_PATH/GeneratedData/LanguageTranslationMultimodalCSV/virtual/Annotations/val.csv \
--val-golden-set $DS_PATH/GeneratedData/LanguageTranslationMultimodalCSV/virtual/Annotations/val_golden.txt --no-evaluation
# no retinanet classification, no retinanet validation
python3 train.py --steps=5000 --val-steps=500 --epochs=10 --batch-size=1 --experiment-tag="NO_RET_CLAS" \
--comet-api-key="$COMET_API_KEY" --comet-project-name="ir_ablation_pos_exp" --comet-workspace="$COMET_WORKSPACE" \
--no-snapshots \
fusionnet $DS_PATH/GeneratedData/FusionMultimodalCSV/virtual/Annotations/train.csv \
--val-annotations $DS_PATH/GeneratedData/FusionMultimodalCSV/virtual/Annotations/val.csv \
--output-filter="transformer_classification,retinanet_regression,fusionnet_regression" \
vision retinanet $DS_PATH/GeneratedData/VisionMultimodalCSV/virtual/Annotations/train.csv \
$DS_PATH/GeneratedData/VisionMultimodalCSV/virtual/Annotations/classmaps.csv \
--val-annotations $DS_PATH/GeneratedData/VisionMultimodalCSV/virtual/Annotations/val.csv --no-evaluation \
language_translation transformer $DS_PATH/GeneratedData/LanguageTranslationMultimodalCSV/virtual/Annotations/train.csv \
--vocab $DS_PATH/GeneratedData/LanguageTranslationMultimodalCSV/virtual/Annotations/vocab.lst \
--val-annotations $DS_PATH/GeneratedData/LanguageTranslationMultimodalCSV/virtual/Annotations/val.csv \
--val-golden-set $DS_PATH/GeneratedData/LanguageTranslationMultimodalCSV/virtual/Annotations/val_golden.txt
# no retinanet regression, no retinanet validation
python3 train.py --steps=5000 --val-steps=500 --epochs=10 --batch-size=1 --experiment-tag="NO_RET_REG" \
--comet-api-key="$COMET_API_KEY" --comet-project-name="ir_ablation_pos_exp" --comet-workspace="$COMET_WORKSPACE" \
--no-snapshots \
fusionnet $DS_PATH/GeneratedData/FusionMultimodalCSV/virtual/Annotations/train.csv \
--val-annotations $DS_PATH/GeneratedData/FusionMultimodalCSV/virtual/Annotations/val.csv \
--output-filter="transformer_classification,retinanet_classification,fusionnet_regression" \
vision retinanet $DS_PATH/GeneratedData/VisionMultimodalCSV/virtual/Annotations/train.csv \
$DS_PATH/GeneratedData/VisionMultimodalCSV/virtual/Annotations/classmaps.csv \
--val-annotations $DS_PATH/GeneratedData/VisionMultimodalCSV/virtual/Annotations/val.csv --no-evaluation \
language_translation transformer $DS_PATH/GeneratedData/LanguageTranslationMultimodalCSV/virtual/Annotations/train.csv \
--vocab $DS_PATH/GeneratedData/LanguageTranslationMultimodalCSV/virtual/Annotations/vocab.lst \
--val-annotations $DS_PATH/GeneratedData/LanguageTranslationMultimodalCSV/virtual/Annotations/val.csv \
--val-golden-set $DS_PATH/GeneratedData/LanguageTranslationMultimodalCSV/virtual/Annotations/val_golden.txt
# no retinanet regression, no retinanet classification, no retinanet validation
python3 train.py --steps=5000 --val-steps=500 --epochs=10 --batch-size=1 --experiment-tag="NO_RET_REG_NO_RET_CLAS" \
--comet-api-key="$COMET_API_KEY" --comet-project-name="ir_ablation_pos_exp" --comet-workspace="$COMET_WORKSPACE" \
--no-snapshots \
fusionnet $DS_PATH/GeneratedData/FusionMultimodalCSV/virtual/Annotations/train.csv \
--val-annotations $DS_PATH/GeneratedData/FusionMultimodalCSV/virtual/Annotations/val.csv \
--output-filter="transformer_classification,fusionnet_regression" \
vision retinanet $DS_PATH/GeneratedData/VisionMultimodalCSV/virtual/Annotations/train.csv \
$DS_PATH/GeneratedData/VisionMultimodalCSV/virtual/Annotations/classmaps.csv \
--val-annotations $DS_PATH/GeneratedData/VisionMultimodalCSV/virtual/Annotations/val.csv --no-evaluation \
language_translation transformer $DS_PATH/GeneratedData/LanguageTranslationMultimodalCSV/virtual/Annotations/train.csv \
--vocab $DS_PATH/GeneratedData/LanguageTranslationMultimodalCSV/virtual/Annotations/vocab.lst \
--val-annotations $DS_PATH/GeneratedData/LanguageTranslationMultimodalCSV/virtual/Annotations/val.csv \
--val-golden-set $DS_PATH/GeneratedData/LanguageTranslationMultimodalCSV/virtual/Annotations/val_golden.txt
# no transformer output, no transformer validation, no retinanet classification, no retinanet regression, no retinanet validation
python3 train.py --steps=5000 --val-steps=500 --epochs=10 --batch-size=1 --experiment-tag="NO_TRANS_NO_RET_CLAS_NO_RET_REG" \
--comet-api-key="$COMET_API_KEY" --comet-project-name="ir_ablation_pos_exp" --comet-workspace="$COMET_WORKSPACE" \
--no-snapshots \
fusionnet $DS_PATH/GeneratedData/FusionMultimodalCSV/virtual/Annotations/train.csv \
--val-annotations $DS_PATH/GeneratedData/FusionMultimodalCSV/virtual/Annotations/val.csv \
--output-filter="fusionnet_regression" \
vision retinanet $DS_PATH/GeneratedData/VisionMultimodalCSV/virtual/Annotations/train.csv \
$DS_PATH/GeneratedData/VisionMultimodalCSV/virtual/Annotations/classmaps.csv \
--val-annotations $DS_PATH/GeneratedData/VisionMultimodalCSV/virtual/Annotations/val.csv --no-evaluation \
language_translation transformer $DS_PATH/GeneratedData/LanguageTranslationMultimodalCSV/virtual/Annotations/train.csv \
--vocab $DS_PATH/GeneratedData/LanguageTranslationMultimodalCSV/virtual/Annotations/vocab.lst \
--val-annotations $DS_PATH/GeneratedData/LanguageTranslationMultimodalCSV/virtual/Annotations/val.csv \
--val-golden-set $DS_PATH/GeneratedData/LanguageTranslationMultimodalCSV/virtual/Annotations/val_golden.txt --no-evaluation
done
else
echo "Activation failed - skipping python package installations. Expected \"$virtual_env_name\" but got $(if [[ -z "$VIRTUAL_ENV" ]]; then echo "nothing"; else basename "$VIRTUAL_ENV"; fi) instead"
fi
cd $scripts_dir