https://github.com/jeon-jinhyeok/llmdiary
Emotion Analysis Diary Using LLM
https://github.com/jeon-jinhyeok/llmdiary
diary django kobert llm postgresql
Last synced: 28 days ago
JSON representation
Emotion Analysis Diary Using LLM
- Host: GitHub
- URL: https://github.com/jeon-jinhyeok/llmdiary
- Owner: Jeon-Jinhyeok
- Created: 2024-11-13T06:06:17.000Z (7 months ago)
- Default Branch: master
- Last Pushed: 2025-02-14T15:10:24.000Z (4 months ago)
- Last Synced: 2025-05-07T21:48:43.197Z (29 days ago)
- Topics: diary, django, kobert, llm, postgresql
- Language: HTML
- Homepage:
- Size: 770 KB
- Stars: 1
- Watchers: 1
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- Support: support/__init__.py
Awesome Lists containing this project
README
# π LLM Diary: κ°μ λΆμ κΈ°λ° λ€μ΄μ΄λ¦¬ νλ«νΌ
![Apache 2.0][license-shield]



![TailwindCSS][tailwind-shield]
[](https://huggingface.co/xxhyeok/koBERT-emotion)

![]()
## π Table of Contents
- [π About the Project](#-about-the-project)
- [π Features](#-features)
- [π» Technologies](#-technologies)
- [π Getting Started](#-getting-started)
- [π Usage](#-usage)
- [π§ Contact](#-contact)
- [π License](#-license)## π About the Project
κ±°λ μΈμ΄ λͺ¨λΈ(LLM)μ μ¬μ©νμ¬ μ¬μ©μμ μΌκΈ°λ₯Ό λ°νμΌλ‘ κ°μ μ λΆμνκ³ , μ¬μ©μμκ² κ°μ ν΅κ³ λ° λ§μΆ€ν μλ΄μ μ μνλ μμ€ν .## π Features
### π·οΈROLE
- **User**
- **Counselor**
- **Admin**
- **UserSupporter**### π Emotion Analsis
- **Model**: [koBERT](https://huggingface.co/monologg/kobert).
- **Fine-tuning Dataset**: [κ°μ±λν λ§λμΉ](https://aihub.or.kr/aihubdata/data/view.do?currMenu=115&topMenu=100&aihubDataSe=data&dataSetSn=86)
- **Emotion** : λΆλ Έ, μ¬ν, λΆμ, μμ², λΉν©, κΈ°μ¨### π₯ Counsel
- User can request a consultation, view connselor info
- Counselor can schedule, conduct, suggest consultations.## π» Technologies
- [Python](https://www.python.org) 3.10.0
- [Django](https://www.djangoproject.com/) 5.1.3
- [tailwindcss](https://tailwindcss.com/) 3.4.14
- [transformers](https://huggingface.co/docs/transformers/ko/index) 4.47.0
## π Getting Started
### 0οΈβ£ Pre-requisite
The easiest way to start the LLMDiary is through **docker compose**. Before running this with the following commands, make sure that [Docker](https://www.docker.com/) and [Docker Compose](https://docs.docker.com/compose/install/) are installed on your machine:### 1οΈβ£ Clone the Repository
```bash
git clone https://github.com/Jeon-Jinhyeok/LLMDiary.git
cd LLMDiary
```
### 2οΈβ£ Download the Model (Hugging Face)
```bash
sudo apt install git-lfs
git lfs clone https://huggingface.co/xxhyeok/koBERT-emotion
```
> π‘ **μ°Έκ³ :** Hugging Face λͺ¨λΈ μ μ₯μμ μ κ·Όνλ €λ©΄ SSHλ₯Ό μ΄μ©ν λ‘κ·ΈμΈ μΈμ¦μ΄ νμν μ μμ΅λλ€.
> [Hugging Face SSH Key μ€μ κ°μ΄λ](https://huggingface.co/docs/hub/security-tokens) μ°Έκ³
### 3οΈβ£ Configure Environment Variables
- setup.envνμΌ μμ
```ini
POSTGRES_DB=emotion_diary_db
POSTGRES_USER=admin
POSTGRES_PASSWORD=your-password
POSTGRES_HOST=db
POSTGRES_PORT=5432
DJANGO_SECRET_KEY=your-secret-key
DEBUG=False
ALLOWED_HOSTS=*
```
- setup.env νμΌ μ μ©
```bash
source ./setup.env
```### 4οΈβ£ Run with Docker
```bash
docker compose up
```
### 5οΈβ£ Access the Web Interface
```plaintext
http://localhost:8000
```## π Usage
### Login Page
---
### Diary List with Representative emotion
---
### Diary Details
- Diary Details with Result of Emotion Analysis and Recommended Action
- You can edit or delete the written diary.
---### Home
- You can check latest Diaries in Home
- You can check emotional statistics for one week.## π§ Contact
- [Jinhyeok, JEON](https://github.com/Jeon-Jinhyeok) - [email protected]
- [Donggeon, KIM](https://github.com/ehdrjs6831) -## π License
Apache 2.0---
[license-shield]: https://img.shields.io/badge/license-Apache%202.0-blue?style=round
[repository-size-shield]: https://img.shields.io/github/repo-size/dev-ujin/readme-template?labelColor=D8D8D8&color=BE81F7
[django-shield]: "https://img.shields.io/badge/django-092E20?style=flat-square&logo=django&logoColor=white"
[tailwind-shield]: https://img.shields.io/badge/TailwindCSS-38B2AC?style=flat-square&logo=tailwind-css&logoColor=white