Auto translation of new anime episodes based on Yui-MHCP001 LLM
Make sure you have Python >= 3.9 installed on your system
pip install yuisubIf you wanna use the a2t module, you need to install Whisper first
# pip install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/cu118
pip install openai-whisperyuisub can be used from the command line to generate bilingual ASS files. Here's how to use it:
yuisub -h # Displays help messageyuisub can also be used as a library
from yuisub import translate, bilingual, load
from yuisub.a2t import WhisperModel
# sub from audio
model = WhisperModel(name="medium", device="cuda")
sub = model.transcribe(audio="path/to/audio.mp3")
# sub from file
# sub = load("path/to/input.srt")
# generate bilingual subtitle
sub_zh = translate(
sub=sub,
model="gpt_model_name",
api_key="your_openai_api_key",
base_url="api_url",
bangumi_url="https://bangumi.tv/subject/424883/"
)
sub_bilingual = bilingual(
sub_origin=sub,
sub_zh=sub_zh
)
# save the ASS files
sub_zh.save("path/to/output.zh.ass")
sub_bilingual.save("path/to/output.bilingual.ass")This project is licensed under the BSD 3-Clause - see the LICENSE file for details.