lIlBrother's picture
Update: README
86185c1
metadata
language:
  - ko
license: apache-2.0
library_name: kenlm
tags:
  - audio
  - automatic-speech-recognition
  - text2text-generation
datasets:
  - korean-wiki

ko-ctc-kenlm-spelling-only-wiki

Table of Contents

Model Details

  • Model Description
    • ์Œํ–ฅ ๋ชจ๋ธ์„ ์œ„ํ•œ N-gram Base์˜ LM์œผ๋กœ ์ž์†Œ๋ณ„ ๋‹จ์–ด๊ธฐ๋ฐ˜์œผ๋กœ ๋งŒ๋“ค์–ด์กŒ์œผ๋ฉฐ, KenLM์œผ๋กœ ํ•™์Šต๋˜์—ˆ์Šต๋‹ˆ๋‹ค. ํ•ด๋‹น ๋ชจ๋ธ์€ ko-spelling-wav2vec2-conformer-del-1s๊ณผ ์‚ฌ์šฉํ•˜์‹ญ์‹œ์˜ค.
    • HuggingFace Transformers Style๋กœ ๋ถˆ๋Ÿฌ์™€ ์‚ฌ์šฉํ•  ์ˆ˜ ์žˆ๋„๋ก ์ฒ˜๋ฆฌํ–ˆ์Šต๋‹ˆ๋‹ค.
    • pyctcdecode lib์„ ์ด์šฉํ•ด์„œ๋„ ๋ฐ”๋กœ ์‚ฌ์šฉ๊ฐ€๋Šฅํ•ฉ๋‹ˆ๋‹ค.
    • data๋Š” wiki korean์„ ์‚ฌ์šฉํ–ˆ์Šต๋‹ˆ๋‹ค.
      spelling vocab data์— ์—†๋Š” ๋ฌธ์žฅ์€ ์ „๋ถ€ ์ œ๊ฑฐํ•˜์—ฌ, ์˜คํžˆ๋ ค LM์œผ๋กœ Outlier๊ฐ€ ๋ฐœ์ƒํ•  ์†Œ์š”๋ฅผ ์ตœ์†Œํ™” ์‹œ์ผฐ์Šต๋‹ˆ๋‹ค.
      ํ•ด๋‹น ๋ชจ๋ธ์€ ์ฒ ์ž์ „์‚ฌ ๊ธฐ์ค€์˜ ๋ฐ์ดํ„ฐ๋กœ ํ•™์Šต๋œ ๋ชจ๋ธ์ž…๋‹ˆ๋‹ค. (์ˆซ์ž์™€ ์˜์–ด๋Š” ๊ฐ ํ‘œ๊ธฐ๋ฒ•์„ ๋”ฐ๋ฆ„)
  • Developed by: TADev (@lIlBrother)
  • Language(s): Korean
  • License: apache-2.0

How to Get Started With the Model

import librosa
from pyctcdecode import build_ctcdecoder
from transformers import (
    AutoConfig,
    AutoFeatureExtractor,
    AutoModelForCTC,
    AutoTokenizer,
    Wav2Vec2ProcessorWithLM,
)
from transformers.pipelines import AutomaticSpeechRecognitionPipeline

audio_path = ""

# ๋ชจ๋ธ๊ณผ ํ† ํฌ๋‚˜์ด์ €, ์˜ˆ์ธก์„ ์œ„ํ•œ ๊ฐ ๋ชจ๋“ˆ๋“ค์„ ๋ถˆ๋Ÿฌ์˜ต๋‹ˆ๋‹ค.
model = AutoModelForCTC.from_pretrained("42MARU/ko-spelling-wav2vec2-conformer-del-1s")
feature_extractor = AutoFeatureExtractor.from_pretrained("42MARU/ko-spelling-wav2vec2-conformer-del-1s")
tokenizer = AutoTokenizer.from_pretrained("42MARU/ko-spelling-wav2vec2-conformer-del-1s")
processor = Wav2Vec2ProcessorWithLM("42MARU/ko-ctc-kenlm-spelling-only-wiki")

# ์‹ค์ œ ์˜ˆ์ธก์„ ์œ„ํ•œ ํŒŒ์ดํ”„๋ผ์ธ์— ์ •์˜๋œ ๋ชจ๋“ˆ๋“ค์„ ์‚ฝ์ž….
asr_pipeline = AutomaticSpeechRecognitionPipeline(
    model=model,
    tokenizer=processor.tokenizer,
    feature_extractor=processor.feature_extractor,
    decoder=processor.decoder,
    device=-1,
)

# ์Œ์„ฑํŒŒ์ผ์„ ๋ถˆ๋Ÿฌ์˜ค๊ณ  beamsearch ํŒŒ๋ผ๋ฏธํ„ฐ๋ฅผ ํŠน์ •ํ•˜์—ฌ ์˜ˆ์ธก์„ ์ˆ˜ํ–‰ํ•ฉ๋‹ˆ๋‹ค.
raw_data, _ = librosa.load(audio_path, sr=16000)
kwargs = {"decoder_kwargs": {"beam_width": 100}}
pred = asr_pipeline(inputs=raw_data, **kwargs)["text"]
# ๋ชจ๋ธ์ด ์ž์†Œ ๋ถ„๋ฆฌ ์œ ๋‹ˆ์ฝ”๋“œ ํ…์ŠคํŠธ๋กœ ๋‚˜์˜ค๋ฏ€๋กœ, ์ผ๋ฐ˜ String์œผ๋กœ ๋ณ€ํ™˜ํ•ด์ค„ ํ•„์š”๊ฐ€ ์žˆ์Šต๋‹ˆ๋‹ค.
result = unicodedata.normalize("NFC", pred)
print(result)
# ์•ˆ๋…•ํ•˜์„ธ์š” 123 ํ…Œ์ŠคํŠธ์ž…๋‹ˆ๋‹ค.