def test(self, input_filepath, args=QATestArgs()):
"""
Tests the question answering model. Used to obtain results
input_filepath: a string that contains the location of a csv file
for training. Contains the following header values:
context, question
args: Either a QATestArgs() object or a dictionary that contains all of the same keys as ARGS_QA_TEST
return: A list of dictionaries. Each dictionary
contains the keys: "score", "start", "end" and "answer"
"""
if type(args) == dict:
method_dataclass_args = create_args_dataclass(
default_dic_args=ARGS_QA_TEST,
input_dic_args=args,
method_dataclass_args=QATestArgs)
elif type(args) == QATestArgs:
method_dataclass_args = args
else:
raise ValueError(
"Invalid args type. Use a QATestArgs object or a dictionary")
return self._trainer.test(input_filepath=input_filepath,
solve=self.answer_question,
dataclass_args=method_dataclass_args)
def train(self, input_filepath, args=QATrainArgs()):
"""
Trains the question answering model
input_filepath: a string that contains the location of a csv file
for training. Contains the following header values: context,
question, answer_text, answer_start
args: Either a QATrainArgs() object or a dictionary that contains all of the same keys as ARGS_QA_TRAIN
return: None
"""
if type(args) == dict:
method_dataclass_args = create_args_dataclass(
default_dic_args=ARGS_QA_TRAIN,
input_dic_args=args,
method_dataclass_args=QATrainArgs)
elif type(args) == QATrainArgs:
method_dataclass_args = args
else:
raise ValueError(
"Invalid args type. Use a QATrainArgs object or a dictionary")
self._trainer.train(input_filepath=input_filepath,
dataclass_args=method_dataclass_args)
def eval(self, input_filepath, args=QAEvalArgs()) -> EvalResult:
"""
Trains the question answering model
input_filepath: a string that contains the location of a csv file
for training. Contains the following header values:
context, question, answer_text, answer_start
args: Either a QAEvalArgs() object or a dictionary that contains all of the same keys as ARGS_QA_EVAl
return: A dictionary that contains a key called "eval_loss"
"""
if type(args) == dict:
method_dataclass_args = create_args_dataclass(
default_dic_args=ARGS_QA_EVAl,
input_dic_args=args,
method_dataclass_args=QAEvalArgs)
elif type(args) == QAEvalArgs:
method_dataclass_args = args
else:
raise ValueError(
"Invalid args type. Use a QAEvalArgs object or a dictionary")
return self._trainer.eval(input_filepath=input_filepath,
dataclass_args=method_dataclass_args)
def train(self, input_filepath, args=ARGS_WP_TRAIN):
if type(args) == dict:
method_dataclass_args = create_args_dataclass(
default_dic_args=ARGS_WP_TRAIN,
input_dic_args=args,
method_dataclass_args=WPTrainArgs)
elif type(args) == WPTrainArgs:
method_dataclass_args = args
else:
raise ValueError(
"Invalid args type. Use a WPTrainArgs object or a dictionary")
self._trainer.train(input_filepath=input_filepath,
dataclass_args=method_dataclass_args)
def eval(self, input_filepath, args=ARGS_WP_EVAl) -> EvalResult:
if type(args) == dict:
method_dataclass_args = create_args_dataclass(
default_dic_args=ARGS_WP_EVAl,
input_dic_args=args,
method_dataclass_args=WPEvalArgs)
elif type(args) == WPEvalArgs:
method_dataclass_args = args
else:
raise ValueError(
"Invalid args type. Use a ARGS_WP_EVAl object or a dictionary")
return self._trainer.eval(input_filepath=input_filepath,
dataclass_args=method_dataclass_args)
def eval(self, input_filepath: str, args=GENEvalArgs()) -> EvalResult:
"""
:param input_filepath:a file path to a text file that contains nothing but evaluating data
:param args: either a GENEvalArgs() object or a dictionary that contains all of the same keys as ARGS_GEN_EVAl
:return: None
"""
if type(args) == dict:
method_dataclass_args = create_args_dataclass(
default_dic_args=ARGS_GEN_EVAl,
input_dic_args=args,
method_dataclass_args=GENEvalArgs)
elif type(args) == GENEvalArgs:
method_dataclass_args = args
else:
raise ValueError(
"Invalid args type. Use a GENEvalArgs object or a dictionary")
return self._trainer.eval(input_filepath=input_filepath,
dataclass_args=method_dataclass_args)
def train(self, input_filepath: str, args=GENTrainArgs()):
"""
:param input_filepath:a file path to a text file that contains nothing but training data
:param args: either a GENTrainArgs() object or a dictionary that contains all of the same keys as ARGS_GEN_TRAIN
:return: None
"""
if type(args) == dict:
method_dataclass_args = create_args_dataclass(
default_dic_args=ARGS_GEN_TRAIN,
input_dic_args=args,
method_dataclass_args=GENTrainArgs)
elif type(args) == GENTrainArgs:
method_dataclass_args = args
else:
raise ValueError(
"Invalid args type. Use a GENTrainArgs object or a dictionary")
self._trainer.train(input_filepath=input_filepath,
dataclass_args=method_dataclass_args)
def eval(self, input_filepath, args=TCEvalArgs()) -> EvalResult:
"""
Evaluated the text classification answering model
input_filepath: a string that contains the location of a csv file
for training. Contains the following header values:
text, label
return: an EvalResult() object
"""
if type(args) == dict:
method_dataclass_args = create_args_dataclass(
default_dic_args=ARGS_TC_EVAL,
input_dic_args=args,
method_dataclass_args=TCEvalArgs)
elif type(args) == TCEvalArgs:
method_dataclass_args = args
else:
raise ValueError(
"Invalid args type. Use a TCEvalArgs object or a dictionary")
return self._trainer.eval(input_filepath=input_filepath,
dataclass_args=method_dataclass_args)
def test(self, input_filepath, args=TCTestArgs()):
"""
Tests the text classification model. Used to obtain results
input_filepath: a string that contains the location of a csv file
for training. Contains the following header value:
text
return: A list of TextClassificationResult() objects
"""
if type(args) == dict:
method_dataclass_args = create_args_dataclass(
default_dic_args=ARGS_TC_TEST,
input_dic_args=args,
method_dataclass_args=TCTestArgs)
elif type(args) == TCTestArgs:
method_dataclass_args = args
else:
raise ValueError(
"Invalid args type. Use a TCTestArgs() object or a dictionary")
return self._trainer.test(input_filepath=input_filepath,
solve=self.classify_text,
dataclass_args=method_dataclass_args)