-
Notifications
You must be signed in to change notification settings - Fork 1
/
config.json
76 lines (76 loc) · 2.54 KB
/
config.json
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
{
"mode": "finetune",
"base_args": {
"output_dir": "checkpoints/",
"overwrite_output_dir": true,
"fp16": true,
"load_best_model_at_end": true,
"save_safetensors": true,
"save_total_limit": 2,
"report_to": "wandb"
},
"pretrain": {
"pretrain_args": {
"num_train_epochs": 10,
"per_device_train_batch_size": 64,
"per_device_eval_batch_size": 64,
"evaluation_strategy": "steps",
"save_steps": 10000,
"eval_steps": 10000,
"learning_rate": 0.00002,
"metric_for_best_model": "eval_loss",
"greater_is_better": false,
"warmup_ratio": 0.2
},
"gradual_unfreeze": false,
"discriminative_lr": false,
"output_dir": "output/FinTwitBERT",
"pretrained_model": "yiyanghkust/finbert-pretrain",
"pretrained_tokenizer": "yiyanghkust/finbert-pretrain"
},
"pre-finetune": {
"pre-finetune_args": {
"num_train_epochs": 10,
"per_device_train_batch_size": 64,
"per_device_eval_batch_size": 64,
"evaluation_strategy": "steps",
"save_steps": 250,
"eval_steps": 250,
"logging_steps": 250,
"learning_rate": 0.00002,
"metric_for_best_model": "accuracy",
"greater_is_better": true,
"warmup_ratio": 0.2,
"weight_decay": 0.01
},
"gradual_unfreeze": false,
"discriminative_lr": true,
"dft_rate": 1.2,
"output_dir": "output/FinTwitBERT-tweeteval",
"pretrained_model": "output/FinTwitBERT",
"pretrained_tokenizer": "output/FinTwitBERT"
},
"finetune": {
"finetune_args": {
"num_train_epochs": 10,
"per_device_train_batch_size": 64,
"per_device_eval_batch_size": 64,
"evaluation_strategy": "steps",
"save_steps": 500,
"eval_steps": 500,
"logging_steps": 500,
"learning_rate": 0.00002,
"metric_for_best_model": "accuracy",
"greater_is_better": true,
"warmup_ratio": 0.2,
"weight_decay": 0.01
},
"gradual_unfreeze": false,
"discriminative_lr": true,
"dft_rate": 1.2,
"oversampling": "simple",
"output_dir": "output/FinTwitBERT-sentiment",
"pretrained_model": "output/FinTwitBERT-tweeteval",
"pretrained_tokenizer": "output/FinTwitBERT-tweeteval"
}
}