-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathrun_train.sh
68 lines (66 loc) · 2.09 KB
/
run_train.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
# EmoRoBERTa
#CUDA_VISIBLE_DEVICES=0 python3 run_ememe_main.py \
# --dataset_name goemotion \
# --model_name_or_path arpanghoshal/EmoRoBERTa \
# --do_train \
# --do_eval \
# --max_seq_length 128 \
# --per_device_train_batch_size 32 \
# --per_device_eval_batch_size 32 \
# --seed 123 \
# --learning_rate 1e-5 \
# --num_train_epochs 3 \
# --evaluation_strategy epoch \
# --save_strategy epoch \
# --load_best_model_at_end \
# --metric_for_best_model accuracy \
# --greater_is_better true \
# --output_dir outputs/test/ \
# --overwrite_cache
# ViLT
#CUDA_VISIBLE_DEVICES=0 python3 run_ememe_main.py \
# --dataset_name memotion \
# --model_name_or_path dandelin/vilt-b32-mlm \
# --do_train \
# --do_eval \
# --max_seq_length 128 \
# --per_device_train_batch_size 32 \
# --per_device_eval_batch_size 32 \
# --seed 123 \
# --learning_rate 1e-5 \
# --num_train_epochs 3 \
# --evaluation_strategy epoch \
# --save_strategy epoch \
# --load_best_model_at_end \
# --metric_for_best_model accuracy \
# --greater_is_better true \
# --output_dir outputs/test/ \
# --overwrite_cache
# Ememe Model
CUDA_VISIBLE_DEVICES=0 python3 run_ememe_main.py \
--dataset_name ememe \
--cached_dataset cached_data \
--model_name_or_path ememe \
--emoroberta_model_ckpt emoroberta_pytorch_model.bin \
--emoroberta_model_name_or_path tae898/emoberta-base \
--vilt_model_name_or_path dandelin/vilt-b32-mlm \
--do_train \
--do_eval \
--evaluation_strategy epoch \
--save_strategy epoch \
--load_best_model_at_end \
--metric_for_best_model accuracy \
--greater_is_better true \
--output_dir outputs/epoch_50ememe-loss_c_0.5-logic_scale_2.6592-proj_dim_512/ \
--overwrite_cache \
--max_seq_length 128 \
--seed 123 \
--learning_rate 1e-5 \
--num_train_epochs 50 \
--loss_c 0.5 \
--contrastive_logit_scale 2.6592 \
--per_device_train_batch_size 64 \
--per_device_eval_batch_size 64 \
--projection_dim 512 \
--overwrite_output_dir \
--train_w_huggingface_trainer False