-
Notifications
You must be signed in to change notification settings - Fork 1
/
tests.py
185 lines (149 loc) · 7.05 KB
/
tests.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
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
179
180
181
182
183
184
185
import CodemixedNLP as csnlp
DATASET_FOLDER = "./datasets"
def test_bert_only_classification(pretrained_name_or_path, dataset="dummysail2017/Hinglish"):
dataset_folder = f"{DATASET_FOLDER}/{dataset}"
csnlp.benchmarks.run_unified(
dataset_folder=dataset_folder,
encoders="bert",
encoder_attributes={"pretrained_name_or_path": pretrained_name_or_path},
task_attributes={"name": "classification"},
target_label_fields="label",
mode="train",
max_epochs=1)
return
def test_bert_only_classification_tagging(pretrained_name_or_path):
dataset_folder = f"{DATASET_FOLDER}/dummysail2017/Hinglish"
csnlp.benchmarks.run_unified(
dataset_folder=dataset_folder,
encoders="bert",
encoder_attributes={"pretrained_name_or_path": pretrained_name_or_path},
task_attributes=[{"name": "classification"}, {"name": "seq_tagging"}],
target_label_fields=["label", "langids"],
mode="train")
return
def test_bert_only_tagging(pretrained_name_or_path, label_field):
dataset_folder = f"{DATASET_FOLDER}/dummysail2017/Hinglish"
csnlp.benchmarks.run_unified(
dataset_folder=dataset_folder,
encoders="bert",
encoder_attributes={"pretrained_name_or_path": pretrained_name_or_path},
task_attributes={"name": "seq_tagging"},
target_label_fields=label_field,
mode="train")
return
def test_bert_plus_lstm_classification(pretrained_name_or_path,
lstm_input_representation,
encodings_merge_type="concat"):
dataset_folder = f"{DATASET_FOLDER}/dummysail2017/Hinglish"
csnlp.benchmarks.run_unified(
dataset_folder=dataset_folder,
encoders=["bert", "lstm"],
encoder_attributes=[
{"pretrained_name_or_path": pretrained_name_or_path},
{"input_representation": lstm_input_representation}],
task_attributes={"name": "classification"},
target_label_fields="label",
mode="train",
encodings_merge_type=encodings_merge_type,
debug=True)
return
def test_bert_only_classification_with_xxx(xxx_input_label_field):
dataset_folder = f"{DATASET_FOLDER}/dummysail2017/Hinglish"
csnlp.benchmarks.run_unified(
dataset_folder=dataset_folder,
encoders="bert",
encoder_attributes={
"pretrained_name_or_path": "/Users/muralidhar/Education/CMU/11927Capstone/Codemixed/checkpoints/pretrained/bert-base-multilingual-cased"},
task_attributes={"name": "classification"},
target_label_fields="label",
mode="train",
xxx_input_label_field=xxx_input_label_field)
return
def test_bert_only_classification_with_fusion(pretrained_name_or_path):
dataset_folder = f"{DATASET_FOLDER}/dummysail2017/Hinglish"
csnlp.benchmarks.run_unified(
dataset_folder=dataset_folder,
encoders="bert",
encoder_attributes={"pretrained_name_or_path": pretrained_name_or_path},
task_attributes={"name": "classification"},
target_label_fields="label",
mode="train",
fusion_text_fields=["text", "text_pp"])
return
def test_bert_only_classification_from_checkpoint(pretrained_name_or_path, ckpt_path):
dataset_folder = f"{DATASET_FOLDER}/dummysail2017/Hinglish"
csnlp.benchmarks.run_unified(
dataset_folder=dataset_folder,
encoders="bert",
encoder_attributes={"pretrained_name_or_path": pretrained_name_or_path},
task_attributes={"name": "classification"},
target_label_fields="label",
mode="test",
eval_ckpt_path=ckpt_path)
return
def test_bert_only_classification_orgdata(pretrained_name_or_path):
dataset_folder = f"../datasets/sail2017/Hinglish"
csnlp.benchmarks.run_unified(
dataset_folder=dataset_folder,
encoders="bert",
encoder_attributes={"pretrained_name_or_path": pretrained_name_or_path},
task_attributes={"name": "classification"},
target_label_fields="label",
mode=["train", "test"])
return
def test_two_berts_based_classification():
dataset_folder = f"{DATASET_FOLDER}/dummysail2017/Hinglish"
csnlp.benchmarks.run_unified(
dataset_folder=dataset_folder,
encoders=["bert", "bert"],
encoder_attributes=[{"pretrained_name_or_path": "bert-base-multilingual-cased"},
{"pretrained_name_or_path": "xlm-roberta-base"}],
task_attributes={"name": "classification"},
target_label_fields="label",
mode="train",
max_epochs=5,
encodings_merge_type="weighted_ensemble")
return
def test_bert_only_tagging_orgdata(pretrained_name_or_path):
dataset_folder = f"../datasets/gluecos_ner/Hinglish"
csnlp.benchmarks.run_unified(
dataset_folder=dataset_folder,
encoders="bert",
encoder_attributes={"pretrained_name_or_path": pretrained_name_or_path},
task_attributes={"name": "seq_tagging"},
target_label_fields="nertags",
mode=["train", "test"])
return
if __name__ == "__main__":
""" training (dummy datasets) """
# TEST 1
# test_bert_only_classification(pretrained_name_or_path="bert-base-multilingual-cased")
# test_bert_only_classification(pretrained_name_or_path="xlm-roberta-base")
# TEST 2
# test_bert_only_tagging(pretrained_name_or_path="xlm-roberta-base", label_field="langids")
# test_bert_only_classification_tagging(pretrained_name_or_path="xlm-roberta-base")
# TEST 3
# test_bert_plus_lstm_classification(pretrained_name_or_path="xlm-roberta-base", lstm_input_representation="sc")
# # test_bert_plus_lstm_classification(pretrained_name_or_path="xlm-roberta-base",
# # lstm_input_representation="fasttext")
# TEST 4
# test_bert_only_classification_with_xxx(xxx_input_label_field="langids")
# TEST 5
# test_bert_only_classification_with_fusion(pretrained_name_or_path="xlm-roberta-base")
# TOEST 6 - binary classification
# test_bert_only_classification(pretrained_name_or_path="xlm-roberta-base",
# dataset="dummyhatespeech/Hinglish")
""" weighted ensemble concat models """
# test_bert_plus_lstm_classification(pretrained_name_or_path="xlm-roberta-base",
# lstm_input_representation="sc",
# encodings_merge_type="weighted_ensemble")
# test_two_berts_based_classification()
""" from checkpoints (dummy datasets) """
# test_bert_only_classification_from_checkpoint(
# pretrained_name_or_path="bert-base-multilingual-cased",
# ckpt_path="/Users/muralidhar/Education/CMU/11927Capstone/Codemixed/CodemixedNLP/datasets/dummysail2017/Hinglish/checkpoints/2021-03-27_06:32:02.754301"
# )
""" training (org. datasets) """
# test_bert_only_classification_orgdata(pretrained_name_or_path="bert-base-multilingual-cased")
# test_bert_only_tagging_orgdata(pretrained_name_or_path="xlm-roberta-base")
print("complete")