Skip to content

cheezycoding/LinkedinThoughts

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

2 Commits
 
 
 
 
 
 
 
 

Repository files navigation

LinkedIn Thoughts 💭

AI-powered satirical LinkedIn post generator. Create r/LinkedInLunatics-worthy content with a custom-trained MLA (Multi-Head Latent Attention) transformer.

🚀 Live Demo

🏗️ Architecture

├── frontend/          # Next.js app (Vercel)
│   └── src/app/       # React components
└── backend/           # FastAPI + PyTorch (Cloud Run)
    ├── api.py         # REST API
    ├── model_mla.py   # LuhGPT model definition
    ├── mla_v2.py      # Multi-Head Latent Attention
    └── Dockerfile     # Container config

🧠 Model Details

  • Architecture: DeepSeek-style MLA Transformer
  • Parameters: ~114M
  • Training Data: Satirical LinkedIn posts
  • Features:
    • Multi-Head Latent Attention (KV compression)
    • Weight tying (embedding ↔ lm_head)
    • Pre-norm architecture
    • RoPE positional encoding

🛠️ Local Development

Backend

cd backend
pip install -r requirements.txt
# Download model weights to backend/linkedin_mla_best.pt
python api.py

Frontend

cd frontend
npm install
npm run dev

📝 API Usage

curl -X POST https://linkedin-lunatics-708213822442.asia-southeast1.run.app/generate \
  -H "Content-Type: application/json" \
  -d '{"prompt": "I just fired", "temperature": 0.3, "max_tokens": 100}'

⚠️ Disclaimer

This is satire. Please don't actually post AI-generated content on LinkedIn. Greg Hustleworth III is not a real person.

About

No description, website, or topics provided.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors