Metadata-Version: 2.1
Name: ytsum
Version: 0.1.5
Summary: Summarize YouTube video instantly with the power of distill-whisper and Mixtral-8B
Home-page: https://github.com/Siris2314/ytsum
Author: Arihant Tripathi
Author-email: tarihant2001@gmail.com
Classifier: Programming Language :: Python :: 3
Classifier: License :: OSI Approved :: MIT License
Classifier: Operating System :: OS Independent
Requires-Python: >=3.7
Description-Content-Type: text/markdown
Requires-Dist: torch ==2.1.1+cu118
Requires-Dist: requests ==2.32.3
Requires-Dist: pytube ==15.0.0
Requires-Dist: transformers ==4.36.2
Requires-Dist: langchain ==0.2.3
Requires-Dist: langchain-community ==0.2.4
Requires-Dist: langchain-together ==0.1.3

# ytsum

## Summarize YouTube videos instantly with the power of distil-whisper and Mixtral-8x7B

ytsum is a Python package that allows you to summarize YouTube videos by leveraging the power of distil-whisper for transcription and Mixtral-8x7B for generating detailed answers based on the video content.

## Setup

To use ytsum, you need to get an API key from Together AI. 

You can sign up and obtain your API key here: [Together AI](https://www.together.ai/).

## Installation

```bash
pip install ytsum
```
```python
from ytsum import answer_youtube_question, set_together_api_key

# Set your Together AI API key
set_together_api_key("your_together_ai_api_key")

# Example usage
youtube_url = "https://www.youtube.com/watch?v=example"
query = "What is the main topic of this video?"

result = answer_youtube_question(youtube_url, query)

print(result)
```
## Models used

Distil-large-v3: https://huggingface.co/distil-whisper/distil-large-v3

Mixtral-8x7B-SFT: https://huggingface.co/NousResearch/Nous-Hermes-2-Mixtral-8x7B-SFT
