INNER CODE UNIT · Python

inference

graviraja/100-Days-of-NLP · applications/classification/ner_tagging/app/app.py:75

def inference(sentence):
    if isinstance(sentence, str):
        tokens = [words_vocab[words_vocab.START]] + sentence.split() + [words_vocab[words_vocab.END]]
    else:
        tokens = sentence
    
    chars = [['<START']] + [['<START>'] + [ch for ch in word] + ['<END>'] for word in tokens[1:-1]] + [['<END>']]

    char_seq = []
    for word in chars:
        word_len = len(word)
        # truncate the word if it is greater than max_word_len
        if word_len > MAX_WORD_LEN:
            word = word[:MAX_WORD_LEN]
        # pad the word if it less
        else:
            pad_length = MAX_WORD_LEN - word_len
            word = word + [chars_vocab.PAD] * pad_length

View source record →

📰 Research Paper
Loading…
⏳ Fetching content…