语言:

• 韩语 标签:
• 分类 许可证: MIT 数据集:
• nsmc 小部件:
• 文本: "不朽的名作！如此感人的内容还是第一次" 示例标题: "正面"
• 文本: "时间真的浪费了。满分10分连1分都不值.." 示例标题: "负面" 指标:
• 准确率
• F1值
• 精确率
• 召回率- 准确率

情感二元分类（基于KoELECTRA-Small-v3模型和Naver Sentiment Movie Corpus数据集的微调）

使用（适用于Amazon SageMaker推理）

它直接使用了SageMaker推理工具包的接口，因此可以轻松部署到SageMaker端点。

inference_nsmc.py

import json
import sys
import logging
import torch
from torch import nn
from transformers import ElectraConfig
from transformers import ElectraModel, AutoTokenizer, ElectraTokenizer, ElectraForSequenceClassification

logging.basicConfig(
    level=logging.INFO, 
    format='[{%(filename)s:%(lineno)d} %(levelname)s - %(message)s',
    handlers=[
        logging.FileHandler(filename='tmp.log'),
        logging.StreamHandler(sys.stdout)
    ]
)
logger = logging.getLogger(__name__)

max_seq_length = 128
classes = ['负面', '正面']

tokenizer = AutoTokenizer.from_pretrained("daekeun-ml/koelectra-small-v3-nsmc")
device = torch.device("cuda" if torch.cuda.is_available() else "cpu")


def model_fn(model_path=None):
    ####
    # 如果你有自己的训练模型
    # Huggingface预训练模型: 'monologg/koelectra-small-v3-discriminator'
    ####
    #config = ElectraConfig.from_json_file(f'{model_path}/config.json')
    #model = ElectraForSequenceClassification.from_pretrained(f'{model_path}/model.pth', config=config)
    
    # 从Huggingface hub下载模型
    model = ElectraForSequenceClassification.from_pretrained('daekeun-ml/koelectra-small-v3-nsmc')   
    model.to(device)
    return model


def input_fn(input_data, content_type="application/jsonlines"): 
    data_str = input_data.decode("utf-8")
    jsonlines = data_str.split("\n")
    transformed_inputs = []

    for jsonline in jsonlines:
        text = json.loads(jsonline)["text"][0]
        logger.info("输入文本: {}".format(text))          
        encode_plus_token = tokenizer.encode_plus(
            text,
            max_length=max_seq_length,
            add_special_tokens=True,
            return_token_type_ids=False,
            padding="max_length",
            return_attention_mask=True,
            return_tensors="pt",
            truncation=True,
        )
        transformed_inputs.append(encode_plus_token)
        
    return transformed_inputs


def predict_fn(transformed_inputs, model):
    predicted_classes = []
    
    for data in transformed_inputs:
        data = data.to(device)
        output = model(**data)

        softmax_fn = nn.Softmax(dim=1)
        softmax_output = softmax_fn(output[0])
        _, prediction = torch.max(softmax_output, dim=1)

        predicted_class_idx = prediction.item()
        predicted_class = classes[predicted_class_idx]
        score = softmax_output[0][predicted_class_idx]
        logger.info("预测类别: {}".format(predicted_class))

        prediction_dict = {}
        prediction_dict["predicted_label"] = predicted_class
        prediction_dict['score'] = score.cpu().detach().numpy().tolist()

        jsonline = json.dumps(prediction_dict)
        logger.info("json行: {}".format(jsonline))        
        predicted_classes.append(jsonline)

    predicted_classes_jsonlines = "\n".join(predicted_classes)
    return predicted_classes_jsonlines


def output_fn(outputs, accept="application/jsonlines"):
    return outputs, accept

test.py

>>> from inference_nsmc import model_fn, input_fn, predict_fn, output_fn
>>> with open('samples/nsmc.txt', mode='rb') as file:
>>>     model_input_data = file.read()
>>> model = model_fn()
>>> transformed_inputs = input_fn(model_input_data)
>>> predicted_classes_jsonlines = predict_fn(transformed_inputs, model)
>>> model_outputs = output_fn(predicted_classes_jsonlines)
>>> print(model_outputs[0])    
   
[{inference_nsmc.py:47} INFO - 输入文本: 这部电影是最好的电影
[{inference_nsmc.py:47} INFO - 输入文本: 太糟糕了。演员的演技也不好，内容也太差劲了
[{inference_nsmc.py:77} INFO - 预测类别: 正面
[{inference_nsmc.py:84} INFO - json行: {"predicted_label": "正面", "score": 0.9619030952453613}
[{inference_nsmc.py:77} INFO - 预测类别: 负面
[{inference_nsmc.py:84} INFO - json行: {"predicted_label": "负面", "score": 0.9994170665740967}
{"predicted_label": "正面", "score": 0.9619030952453613}
{"predicted_label": "负面", "score": 0.9994170665740967}

示例数据 (samples/nsmc.txt)

{"text": ["这部电影是最好的电影"]}
{"text": ["太糟糕了。演员的演技也不好，内容也太差劲了"]}

参考文献

• KoELECTRA: https://github.com/monologg/KoELECTRA
• Naver Sentiment Movie Corpus数据集: https://github.com/e9t/nsmc