GrahLnn/delta_context2

0

stars

59

commits

Python

primary language

Jun 9, 2026

updated

README

δ context 2

A tool to add any translation subtitles to any video. Using Whisper, translate agent translation. I strongly recommend using gemini-1.5-pro as the translation model; the results are truly amazing (chinese i tested).

How to use

install

# If you are not using the uv package manager, you can simply copy the command that starts with pip.
# install pytorch
uv pip install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/cu118

# install delta context 2
uv pip install git+https://github.com/GrahLnn/delta_context2.git

useage

Prepare your .env file in your project. Fill in according to the .env.example example.

Then

from delta_context2 import VideoProcessor

source_lang, target_lang, country = "English", "Chinese", "China"
ytb_url = "https://www.youtube.com/watch?v=[id]"

video_processor = VideoProcessor(source_lang, target_lang, country)
video_processor.process(ytb_url)
# video in data/videos

todo

  • suport chinese subtitle.
  • other language subtitle. (maybe i won't do)
  • support for YouTube videos download.
  • support other platform download. (e.g. bilibili, x)
  • support local video laod.
  • let model verify the facts in the transcription.
  • generate illustrated articles correspondingly.
  • retrieve existing definitions of standardized translation.
  • suport more asr model (e.g. SenseVoice)

Contributors

GrahLnn

59 commits

GrahLnn/delta_context2

0

stars

59

commits

Python

primary language

Jun 9, 2026

updated

README

δ context 2

A tool to add any translation subtitles to any video. Using Whisper, translate agent translation. I strongly recommend using gemini-1.5-pro as the translation model; the results are truly amazing (chinese i tested).

How to use

install

# If you are not using the uv package manager, you can simply copy the command that starts with pip.
# install pytorch
uv pip install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/cu118

# install delta context 2
uv pip install git+https://github.com/GrahLnn/delta_context2.git

useage

Prepare your .env file in your project. Fill in according to the .env.example example.

Then

from delta_context2 import VideoProcessor

source_lang, target_lang, country = "English", "Chinese", "China"
ytb_url = "https://www.youtube.com/watch?v=[id]"

video_processor = VideoProcessor(source_lang, target_lang, country)
video_processor.process(ytb_url)
# video in data/videos

todo

  • suport chinese subtitle.
  • other language subtitle. (maybe i won't do)
  • support for YouTube videos download.
  • support other platform download. (e.g. bilibili, x)
  • support local video laod.
  • let model verify the facts in the transcription.
  • generate illustrated articles correspondingly.
  • retrieve existing definitions of standardized translation.
  • suport more asr model (e.g. SenseVoice)

Contributors

GrahLnn

59 commits

Languages

Python

100.0%