https://github.com/odynvolk/jensen
A LLaMA 2 powered Telegram chatbot.
https://github.com/odynvolk/jensen
chatbot llama2 llamacpp llm telegram
Last synced: about 1 year ago
JSON representation
A LLaMA 2 powered Telegram chatbot.
- Host: GitHub
- URL: https://github.com/odynvolk/jensen
- Owner: odynvolk
- License: mit
- Created: 2023-08-10T18:42:45.000Z (almost 3 years ago)
- Default Branch: main
- Last Pushed: 2024-02-11T13:07:49.000Z (over 2 years ago)
- Last Synced: 2024-10-18T22:08:35.882Z (over 1 year ago)
- Topics: chatbot, llama2, llamacpp, llm, telegram
- Language: Python
- Homepage:
- Size: 18.6 KB
- Stars: 7
- Watchers: 3
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# jensen
A LLM powered Telegram bot.
## Prerequisites:
### Telegram bot
First you need to create a Telegram bot to interact with. See instructions [here](https://core.telegram.org/bots).
### LLM model
You need a LLM in GGUF format. You can find quite a few through [TheBloke on HuggingFace](https://huggingface.co/TheBloke)
who has done an enormous service to the community by converting different models to GGUF and quantized them. Pick one that suits your
needs and hardware requirements.
### Python
You need Python 3 on your machine.
- Micromamba (optional)
For handling the packages needed for different enivronments. Easy to install with asdf.
```bash
$ asdf install
$ micromamba create -n jensen python=3.13
$ micromamba activate jensen
```
- LLMs, Telegram etc
For using LLM models, Telegram API etc
```bash
$ pip install -r requirements.txt
```
## Configuration
Create a `.env` file with the following properties.
```
# LLM
MODEL_PATH= (mandatory)
MAX_TOKENS=
N_CTX=
N_GPU_LAYERS
N_THREADS=
USE_MLOCK=
# Telegram
API_KEY= (mandatory)
POLL_INTERVAL=
```
## Usage
Start the application:
```bash
python ./jensen/app.py
```