def __init__(self, configs: dict):
# 加载参数字典
self.configs = configs
# 设置模型以及路径
if self.configs["model_select"] == "fasttext":
self.model_path = self.configs["model_path"]["fasttext"]
elif self.configs["model_select"] == "lstm_base":
self.model_path = self.configs["model_path"]["lstm_base"]
elif self.configs["model_select"] == "lstm_pack":
self.model_path = self.configs["model_path"]["lstm_pack"]
elif self.configs["model_select"] == "textcnn":
self.model_path = self.configs["model_path"]["textcnn"]
elif self.configs["model_select"] == "bert":
self.model_path = self.configs["model_path"]["bert"]
elif self.configs["model_select"] == "electra":
self.model_path = self.configs["model_path"]["electra"]
elif self.configs["model_select"] == "xlnet":
self.model_path = self.configs["model_path"]["xlnet"]
# 设置模型名称
self.model_name = self.configs["model_select"]
# 设置评估推断的数据文件
self.eval_data_file = self.configs["eval_data_file"]
# 设置标签转换路径，需要将预测的label转为真实标签
self.label2index_json_path = self.configs["eval_label_transfer_file"]
# 设置token映射词表，在自定义模型中的tokenizer使用
self.token2index_json_path = self.configs["eval_token_transfer_file"]
# 设置分词器
if self.model_name in ["fasttext", "lstm_base", "lstm_pack", "textcnn"]:
self.tokenizer = SequenceTokenizer(load_json(self.token2index_json_path))
elif self.model_name == "bert":
self.tokenizer = BertTokenizer.from_pretrained(self.configs["pretrained_model_path"]["bert"])
elif self.model_name == "electra":
self.tokenizer = ElectraTokenizer.from_pretrained(self.configs["pretrained_model_path"]["electra"])
elif self.model_name == "xlnet":
self.tokenizer = XLNetTokenizer.from_pretrained(self.configs["pretrained_model_path"]["xlnet"])
# 设置label转换器
self.label_tokenizer = ClassificationLabelTokenizer(load_json(self.label2index_json_path))
# 加载模型
self.model = torch.load(self.model_path)
def predict(self):
eval_df = load_xlsx(self.eval_data_file)
data_x = list(eval_df["text"])
data_y = list(eval_df["intent"])
d = load_json(self.label2index_json_path)
# 加载干净的数据集
data_x_clear, data_y_clear = list(), list()
for index in range(len(data_x)):
if data_y[index] != "[]" and "," not in data_y[index] and data_y[index] in d.keys():
data_x_clear.append(data_x[index])
data_y_clear.append(data_y[index])
# print(len(data_x_clear))
# return data_x_clear, data_y_clear
predict_result = list()
for index in tqdm(range(len(data_x_clear))):
predict_result.append(self.predict_one(data_x_clear[index], data_y_clear[index]))
save_df = pd.DataFrame({
"text":data_x_clear,
"intent":data_y_clear,
"predict":predict_result
})
save_df.to_excel("final_.xlsx")
def __init__(self, label2index: dict):
if isinstance(label2index, dict):
self.label2index = label2index
else:
self.label2index = load_json(label2index)
self.index2label = {
value: key
for key, value in self.label2index.items()
}
def __init__(self, configs):
self.configs = configs
self.predicted_dataset_path = self.configs["predicted_dataset_path"]
self.label2index_json_path = self.configs["label2index_json_path"]
# dataframe
self.eval_data = load_xlsx(self.predicted_dataset_path)
# dict，可视化时将索引转为label
self.label2index = load_json(self.label2index_json_path)
self.index2label = {value: key for key, value in self.label2index.items()}
# confusion_matrix
self.confusion_matrix = torch.zeros((len(self.label2index), len(self.label2index)))
# 初始化混淆矩阵
self.confusion_matrix_init()
# 记录测试集中不存在的类别（list）
self.not_in_test = self.confusion_matrix_clear()
def get_eval_data(self):
data_all_dataframe = load_xlsx(self.eval_path)
data_x = list(data_all_dataframe["text"])
data_y = list(data_all_dataframe["intent"])
# for index in range(len(data_x) - 1, -1, -1):
# # 将多标签的数据替换为第一个标签
# if "," in data_y[index]:
# label_split = data_y[index].split(",")
# label = label_split[0][2:-1]
# data_y[index] = "['" + label + "']"
d = load_json(self.label2index_json_path)
data_x_clear, data_y_clear = list(), list()
for index in range(len(data_x)):
if data_y[index] != "[]" and "," not in data_y[index] and data_y[
index] in d.keys():
data_x_clear.append(data_x[index])
data_y_clear.append(data_y[index])
# print(len(data_x_clear))
return data_x_clear, data_y_clear
def __init__(self, train_record_json):
# 加载训练过程中产生的数据
self.train_params = load_json(train_record_json)