Spaces:
Sleeping
Sleeping
File size: 7,385 Bytes
4a1df2e |
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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 |
import copy
from anonymous_demo.functional.config.config_manager import ConfigManager
from anonymous_demo.core.tad.classic.__bert__.models import TADBERT
_tad_config_template = {
"model": TADBERT,
"optimizer": "adamw",
"learning_rate": 0.00002,
"patience": 99999,
"pretrained_bert": "microsoft/mdeberta-v3-base",
"cache_dataset": True,
"warmup_step": -1,
"show_metric": False,
"max_seq_len": 80,
"dropout": 0,
"l2reg": 0.000001,
"num_epoch": 10,
"batch_size": 16,
"initializer": "xavier_uniform_",
"seed": 52,
"polarities_dim": 3,
"log_step": 10,
"evaluate_begin": 0,
"cross_validate_fold": -1,
"use_amp": False,
# split train and test datasets into 5 folds and repeat 3 training
}
_tad_config_base = {
"model": TADBERT,
"optimizer": "adamw",
"learning_rate": 0.00002,
"pretrained_bert": "microsoft/deberta-v3-base",
"cache_dataset": True,
"warmup_step": -1,
"show_metric": False,
"max_seq_len": 80,
"patience": 99999,
"dropout": 0,
"l2reg": 0.000001,
"num_epoch": 10,
"batch_size": 16,
"initializer": "xavier_uniform_",
"seed": 52,
"polarities_dim": 3,
"log_step": 10,
"evaluate_begin": 0,
"cross_validate_fold": -1
# split train and test datasets into 5 folds and repeat 3 training
}
_tad_config_english = {
"model": TADBERT,
"optimizer": "adamw",
"learning_rate": 0.00002,
"patience": 99999,
"pretrained_bert": "microsoft/deberta-v3-base",
"cache_dataset": True,
"warmup_step": -1,
"show_metric": False,
"max_seq_len": 80,
"dropout": 0,
"l2reg": 0.000001,
"num_epoch": 10,
"batch_size": 16,
"initializer": "xavier_uniform_",
"seed": 52,
"polarities_dim": 3,
"log_step": 10,
"evaluate_begin": 0,
"cross_validate_fold": -1
# split train and test datasets into 5 folds and repeat 3 training
}
_tad_config_multilingual = {
"model": TADBERT,
"optimizer": "adamw",
"learning_rate": 0.00002,
"patience": 99999,
"pretrained_bert": "microsoft/mdeberta-v3-base",
"cache_dataset": True,
"warmup_step": -1,
"show_metric": False,
"max_seq_len": 80,
"dropout": 0,
"l2reg": 0.000001,
"num_epoch": 10,
"batch_size": 16,
"initializer": "xavier_uniform_",
"seed": 52,
"polarities_dim": 3,
"log_step": 10,
"evaluate_begin": 0,
"cross_validate_fold": -1
# split train and test datasets into 5 folds and repeat 3 training
}
_tad_config_chinese = {
"model": TADBERT,
"optimizer": "adamw",
"learning_rate": 0.00002,
"patience": 99999,
"cache_dataset": True,
"warmup_step": -1,
"show_metric": False,
"pretrained_bert": "bert-base-chinese",
"max_seq_len": 80,
"dropout": 0,
"l2reg": 0.000001,
"num_epoch": 10,
"batch_size": 16,
"initializer": "xavier_uniform_",
"seed": 52,
"polarities_dim": 3,
"log_step": 10,
"evaluate_begin": 0,
"cross_validate_fold": -1
# split train and test datasets into 5 folds and repeat 3 training
}
class TADConfigManager(ConfigManager):
def __init__(self, args, **kwargs):
"""
Available Params: {'model': BERT,
'optimizer': "adamw",
'learning_rate': 0.00002,
'pretrained_bert': "roberta-base",
'cache_dataset': True,
'warmup_step': -1,
'show_metric': False,
'max_seq_len': 80,
'patience': 99999,
'dropout': 0,
'l2reg': 0.000001,
'num_epoch': 10,
'batch_size': 16,
'initializer': 'xavier_uniform_',
'seed': {52, 25}
'embed_dim': 768,
'hidden_dim': 768,
'polarities_dim': 3,
'log_step': 10,
'evaluate_begin': 0,
'cross_validate_fold': -1 # split train and test datasets into 5 folds and repeat 3 training
}
:param args:
:param kwargs:
"""
super().__init__(args, **kwargs)
@staticmethod
def set_tad_config(configType: str, newitem: dict):
if isinstance(newitem, dict):
if configType == "template":
_tad_config_template.update(newitem)
elif configType == "base":
_tad_config_base.update(newitem)
elif configType == "english":
_tad_config_english.update(newitem)
elif configType == "chinese":
_tad_config_chinese.update(newitem)
elif configType == "multilingual":
_tad_config_multilingual.update(newitem)
elif configType == "glove":
_tad_config_glove.update(newitem)
else:
raise ValueError(
"Wrong value of config type supplied, please use one from following type: template, base, english, chinese, multilingual, glove"
)
else:
raise TypeError(
"Wrong type of new config item supplied, please use dict e.g.{'NewConfig': NewValue}"
)
@staticmethod
def set_tad_config_template(newitem):
TADConfigManager.set_tad_config("template", newitem)
@staticmethod
def set_tad_config_base(newitem):
TADConfigManager.set_tad_config("base", newitem)
@staticmethod
def set_tad_config_english(newitem):
TADConfigManager.set_tad_config("english", newitem)
@staticmethod
def set_tad_config_chinese(newitem):
TADConfigManager.set_tad_config("chinese", newitem)
@staticmethod
def set_tad_config_multilingual(newitem):
TADConfigManager.set_tad_config("multilingual", newitem)
@staticmethod
def set_tad_config_glove(newitem):
TADConfigManager.set_tad_config("glove", newitem)
@staticmethod
def get_tad_config_template() -> ConfigManager:
_tad_config_template.update(_tad_config_template)
return TADConfigManager(copy.deepcopy(_tad_config_template))
@staticmethod
def get_tad_config_base() -> ConfigManager:
_tad_config_template.update(_tad_config_base)
return TADConfigManager(copy.deepcopy(_tad_config_template))
@staticmethod
def get_tad_config_english() -> ConfigManager:
_tad_config_template.update(_tad_config_english)
return TADConfigManager(copy.deepcopy(_tad_config_template))
@staticmethod
def get_tad_config_chinese() -> ConfigManager:
_tad_config_template.update(_tad_config_chinese)
return TADConfigManager(copy.deepcopy(_tad_config_template))
@staticmethod
def get_tad_config_multilingual() -> ConfigManager:
_tad_config_template.update(_tad_config_multilingual)
return TADConfigManager(copy.deepcopy(_tad_config_template))
@staticmethod
def get_tad_config_glove() -> ConfigManager:
_tad_config_template.update(_tad_config_glove)
return TADConfigManager(copy.deepcopy(_tad_config_template))
|