-
Notifications
You must be signed in to change notification settings - Fork 0
/
cls_enhanced_base.py
36 lines (33 loc) · 1.43 KB
/
cls_enhanced_base.py
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
import random
import pandas as pd
from configuration import Configuration
from train import train_cls_enhanced
from configuration import CONSTANTS as C
from eval_test import test_enhanced_cls
import os
print(C.DEVICE)
# baseline without adapter
n_epochs_cls = 10
model_size_list = ['base']
with_adapter_list = ['yes', 'no']
lr_cls = 1e-7
wd_cls = 1e-3
mlm_adapter_name_list = ['model_1657682495', 'model_1657692000', 'model_1657659496']
dict_list_model = {'n_epochs_cls':[], 'lr_cls':[], 'wd_cls':[], 'model_size':[], 'test_model_dir':[], 'test_loss':[], 'mlm_adapter_name':[]}
for i in range(len(model_size_list)):
model_size = model_size_list[i]
mlm_adapter_name = mlm_adapter_name_list[i]
print(model_size)
dict_cls = {'n_epochs_cls': n_epochs_cls, 'lr_cls': lr_cls, 'wd_cls': wd_cls, 'mlm_adapter_name': mlm_adapter_name, 'model_size': model_size}
config = Configuration(dict_cls)
model_dir = train_cls_enhanced(config)
dict_test = {'model_size': model_size, 'test_model_dir': model_dir}
config = Configuration(dict_test)
test_loss = test_enhanced_cls(config)
for key in dict_cls.keys():
dict_list_model[key].append(dict_cls[key])
dict_list_model['test_model_dir'].append(model_dir)
dict_list_model['test_loss'].append(test_loss)
if not os.path.exists('models_result/'):
os.makedirs('models_result/')
pd.DataFrame(dict_list_model).to_csv('models_result/quick_cls_summary_enhanced.csv')