-
Notifications
You must be signed in to change notification settings - Fork 0
/
run_conv_image_alignment_finetuning.sh
executable file
·154 lines (150 loc) · 6.19 KB
/
run_conv_image_alignment_finetuning.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
# CUDA_VISIBLE_DEVICES=1 python conv_image_alignment_finetuning.py \
# --output_dir="./save/coref_candidates|mm/conv-im-align/clip" \
# --cache_dir="./cache/coref_candidates|mm/conv-im-align/clip" \
# --train_dataset_path="./preprocessed_data/coref_candidates|mm/simmc2.1_coref_candidates_dstc11_train.json" \
# --dev_dataset_path="./preprocessed_data/coref_candidates|mm/simmc2.1_coref_candidates_dstc11_dev.json" \
# --devtest_dataset_path="./preprocessed_data/coref_candidates|mm/simmc2.1_coref_candidates_dstc11_devtest.json" \
# --additional_special_token_path "./preprocessed_data/coref_candidates|mm/simmc2_special_tokens.json" \
# --model_name_or_path="openai/clip-vit-base-patch32" \
# --include_other_similar_objects=False \
# --include_other_referred_objects=False \
# --per_device_train_batch_size=32 \
# --per_device_eval_batch_size=32 \
# --num_train_epochs=200 \
# --max_seq_length=77 \
# --fp16=True \
# --save_strategy="epoch" \
# --save_steps=1 \
# --logging_strategy="epoch" \
# --logging_steps=1 \
# --evaluation_strategy="epoch" \
# --eval_steps=1 \
# --gradient_accumulation_steps=8 \
# --eval_accumulation_steps=8 \
# --learning_rate=1e-4 \
# --save_total_limit=1 \
# --remove_unused_columns=False \
# --report_to="tensorboard" \
# --seed=42 \
# --data_seed=42 \
# --load_best_model_at_end=True \
# --dataloader_num_workers=4 \
# --overwrite_output_dir=True
CUDA_VISIBLE_DEVICES=1 python conv_image_alignment_finetuning.py \
--output_dir="./save/ambiguous_candidates/conv-im-align/clipper_v3" \
--cache_dir="./cache/ambiguous_candidates/conv-im-align/clipper_v3" \
--train_dataset_path="./preprocessed_data/ambiguous_candidates/simmc2.1_ambiguous_candidates_dstc11_train.json" \
--dev_dataset_path="./preprocessed_data/ambiguous_candidates/simmc2.1_ambiguous_candidates_dstc11_dev.json" \
--devtest_dataset_path="./preprocessed_data/ambiguous_candidates/simmc2.1_ambiguous_candidates_dstc11_devtest.json" \
--model_name_or_path="openai/clip-vit-base-patch32" \
--include_other_similar_objects=True \
--include_other_referred_objects=True \
--per_device_train_batch_size=32 \
--per_device_eval_batch_size=32 \
--num_train_epochs=200 \
--max_seq_length=77 \
--fp16=True \
--save_strategy="epoch" \
--save_steps=1 \
--logging_strategy="epoch" \
--logging_steps=1 \
--evaluation_strategy="epoch" \
--eval_steps=1 \
--gradient_accumulation_steps=1 \
--eval_accumulation_steps=1 \
--learning_rate=1e-5 \
--save_total_limit=1 \
--remove_unused_columns=False \
--report_to="tensorboard" \
--seed=42 \
--data_seed=42 \
--load_best_model_at_end=True \
--dataloader_num_workers=4 \
--overwrite_output_dir=True
# CUDA_VISIBLE_DEVICES=1 python conv_image_alignment_finetuning.py \
# --output_dir="./save/coref_candidates|no_mm/conv-im-align/clipper_v3" \
# --cache_dir="./cache/coref_candidates|no_mm/conv-im-align/clipper_v3" \
# --train_dataset_path="./preprocessed_data/coref_candidates|no_mm/simmc2.1_coref_candidates_dstc11_train.json" \
# --dev_dataset_path="./preprocessed_data/coref_candidates|no_mm/simmc2.1_coref_candidates_dstc11_dev.json" \
# --devtest_dataset_path="./preprocessed_data/coref_candidates|no_mm/simmc2.1_coref_candidates_dstc11_devtest.json" \
# --model_name_or_path="openai/clip-vit-base-patch32" \
# --include_other_similar_objects=True \
# --include_other_referred_objects=True \
# --per_device_train_batch_size=32 \
# --per_device_eval_batch_size=32 \
# --num_train_epochs=200 \
# --max_seq_length=77 \
# --fp16=True \
# --save_strategy="epoch" \
# --save_steps=1 \
# --logging_strategy="epoch" \
# --logging_steps=1 \
# --evaluation_strategy="epoch" \
# --eval_steps=1 \
# --gradient_accumulation_steps=8 \
# --eval_accumulation_steps=8 \
# --learning_rate=1e-4 \
# --save_total_limit=1 \
# --remove_unused_columns=False \
# --report_to="tensorboard" \
# --seed=42 \
# --data_seed=42 \
# --load_best_model_at_end=True \
# --dataloader_num_workers=4 \
# --overwrite_output_dir=True
# CUDA_VISIBLE_DEVICES=0 python conv_image_alignment_finetuning.py \
# --output_dir="./save/conv-im-align/clip_sigmoid" \
# --cache_dir="./cache/conv-im-align/clip_sigmoid" \
# --model_name_or_path="openai/clip-vit-base-patch32" \
# --vision_model_name_or_path="openai/clip-vit-base-patch32" \
# --text_model_name_or_path="roberta-base" \
# --per_device_train_batch_size=32 \
# --per_device_eval_batch_size=32 \
# --num_train_epochs=200 \
# --max_seq_length=77 \
# --fp16=True \
# --save_strategy="epoch" \
# --save_steps=1 \
# --logging_strategy="epoch" \
# --logging_steps=1 \
# --evaluation_strategy="epoch" \
# --eval_steps=1 \
# --gradient_accumulation_steps=8 \
# --eval_accumulation_steps=8 \
# --learning_rate=1e-4 \
# --save_total_limit=1 \
# --remove_unused_columns=False \
# --report_to="tensorboard" \
# --seed=42 \
# --data_seed=42 \
# --load_best_model_at_end=True \
# --dataloader_num_workers=4 \
# --overwrite_output_dir=True
# # DUMMY FOR TESTING
# CUDA_VISIBLE_DEVICES=0 python conv_image_alignment_finetuning.py \
# --output_dir="./save/test" \
# --cache_dir="./cache/test" \
# --model_name_or_path="openai/clip-vit-base-patch32" \
# --vision_model_name_or_path="openai/clip-vit-base-patch32" \
# --text_model_name_or_path="roberta-base" \
# --per_device_train_batch_size=32 \
# --per_device_eval_batch_size=32 \
# --num_train_epochs=10 \
# --fp16=True \
# --save_strategy="epoch" \
# --save_steps=1 \
# --logging_strategy="epoch" \
# --logging_steps=1 \
# --evaluation_strategy="epoch" \
# --eval_steps=1 \
# --gradient_accumulation_steps=8 \
# --eval_accumulation_steps=8 \
# --learning_rate=1e-4 \
# --save_total_limit=1 \
# --remove_unused_columns=False \
# --report_to="tensorboard" \
# --seed=42 \
# --data_seed=42 \
# --load_best_model_at_end=True \
# --dataloader_num_workers=4 \
# --overwrite_output_dir=True