Ecosyste.ms: Awesome
An open API service indexing awesome lists of open source software.
https://github.com/jalvarezz13/summarai
SummarAI utilizes PyMovie and Whisper to transcribe videos, enabling you to ask questions about the content using Llama2 and Llama-index for insightful interaction.
https://github.com/jalvarezz13/summarai
llama-index llama2 pymovie whisper
Last synced: 1 day ago
JSON representation
SummarAI utilizes PyMovie and Whisper to transcribe videos, enabling you to ask questions about the content using Llama2 and Llama-index for insightful interaction.
- Host: GitHub
- URL: https://github.com/jalvarezz13/summarai
- Owner: jalvarezz13
- License: mit
- Created: 2023-10-13T16:44:54.000Z (about 1 year ago)
- Default Branch: main
- Last Pushed: 2023-10-13T20:31:03.000Z (about 1 year ago)
- Last Synced: 2024-10-18T22:01:20.348Z (2 months ago)
- Topics: llama-index, llama2, pymovie, whisper
- Language: Python
- Homepage:
- Size: 62 MB
- Stars: 0
- Watchers: 1
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
🦙 SummarAI 🤖
### Table of Contents
- [📢 What is _SummarAI_?](#-what-is-summarAI)
- [⚙️ Execution](#%EF%B8%8F-execution)
- [⭐ Current Features](#-current-features)
- [⚒️ Troubleshooting](#-troubleshooting)
## 📢 What is _SummarAI_?
**SummarAI** employs the combination of _PyMovie_ and _Whisper_ to provide video transcription services, allowing users to ask questions about the video's content with the help of _Llama2_ and _LlamaIndex_ for more engaging and insightful interactions.
**SummarAI** is divided into three different files:
- `1_extract_audio.py`: This file utilizes the _PyMovie_ library to extract audio from the specified video in a matter of seconds.
- `2_transcribe_audio.py`: In this case, the "speech recognition" model from _OpenAI_, known as _Whisper_, is used to transcribe the audio into a text file.
- `3_llm_over_data`: Finally, this file makes use of _LlamaIndex_, which, based on the _Llama2-13B_ model, provides context to the data from the previously prepared video and responds accurately within that context.
## ⚙️ Execution
To run the **SummarAI**, you can use `pipenv`. First, ensure you have `pipenv` installed. Then, navigate to the root directory of the project, where you can install the dependencies and activate the virtual environment using the following commands:
```bash
pipenv install
pipenv shell
```
Afterwards, simply run the file you want using one of these commands:
```bash
python 1_extract_audio.py
python 2_transcribe_audio.py
python 3_llm_over_data.py
```
## ⭐ Current features
✅ Extract audio from videos
✅ Transcribe audio using Whisper
✅ Resolve your inquiries with the assistance of Llama2
## ⚒️ Troubleshooting
If you encounter issues while installing the `llama-cpp-python` library, you may need to run the following commands:
```bash
set CMAKE_ARGS=-DLLAMA_CUBLAS=on
set FORCE_CMAKE=1
```
If the error persists, please check your C compiler (_gcc_) or, if you are using Windows, update/install the _"Desktop development with C++"_ feature by utilizing the [Microsoft C++ Build Tools](https://visualstudio.microsoft.com/es/visual-cpp-build-tools/)