albert-news-classification

Traditional Chinese news classification

繁體中文新聞分類任務，使用ckiplab/albert-base-chinese預訓練模型，資料集只有2.6萬筆，做為課程的範例模型。

from transformers import BertTokenizer, AlbertForSequenceClassification
model_path = "clhuang/albert-news-classification"
model = AlbertForSequenceClassification.from_pretrained(model_path)
tokenizer = BertTokenizer.from_pretrained("bert-base-chinese")
# Category index
news_categories=['政治','科技','運動','證卷','產經','娛樂','生活','國際','社會','文化','兩岸']
idx2cate = { i : item for i, item in enumerate(news_categories)}

# get category probability
def get_category_proba( text ):
    max_length = 250
    # prepare token sequence
    inputs = tokenizer([text], padding=True, truncation=True, max_length=max_length, return_tensors="pt")
    # perform inference
    outputs = model(**inputs)
    # get output probabilities by doing softmax
    probs = outputs[0].softmax(1)

    # executing argmax function to get the candidate label index
    label_index = probs.argmax(dim=1)[0].tolist() # convert tensor to int
    # get the label name        
    label = idx2cate[ label_index ]

    # get the label probability
    proba = round(float(probs.tolist()[0][label_index]),2)

    response = {'label': label, 'proba': proba}

    return response
   
get_category_proba('俄羅斯2月24日入侵烏克蘭至今不到3個月，芬蘭已準備好扭轉奉行了75年的軍事不結盟政策，申請加入北約。芬蘭總理馬林昨天表示，「希望我們下星期能與瑞典一起提出申請」。')
{'label': '國際', 'proba': 0.99}