flow2 / README.md
AndrianBalanescu
fix: require auth only when FLOW_API_KEY is explicitly set
434c049
|
Raw
History Blame Contribute Delete
1.43 kB

A newer version of the Gradio SDK is available: 6.25.0

Upgrade
metadata
title: AI Creative Studio
emoji: 🎡
colorFrom: blue
colorTo: purple
sdk: gradio
sdk_version: 5.28.0
app_file: app.py
pinned: false
hardware: zero-a10g

AI Creative Studio β€” abalanescu/flow

Multi-modal AI showcase leveraging ZeroGPU (40 min/day A100) + Inference API (~$2/mo credit).

πŸš€ Quick Start

# 1. Clone & set up
git clone https://huggingface.co/spaces/abalanescu/flow
cd flow
cp .env.example .env
export $(cat .env | grep -v '^#' | xargs)

# 2. Install deps
uv pip install -e ".[dev,telemetry]"

# 3. Run
python -m app

πŸ§ͺ Local Test (no GPU needed for inference API calls)

python -c "from app import demo; demo()"

πŸ“ Project Structure

hf/
β”œβ”€β”€ app.py              # Gradio app (5 tabs)
β”œβ”€β”€ pyproject.toml      # Dependencies
β”œβ”€β”€ .env.example        # Secrets template
β”œβ”€β”€ .gitignore
β”œβ”€β”€ freight/            # Freight SFT generator
β”‚   β”œβ”€β”€ generator.py
β”‚   β”œβ”€β”€ sample_data/
β”‚   └── train_qwen_lora.py
β”œβ”€β”€ scripts/            # Utility scripts
β”œβ”€β”€ docs/               # Documentation
└── tests/              # (TODO: add tests)

πŸ”‘ Two HF Resources

Resource What it runs How it's billed
ZeroGPU Local GGUF models via @spaces.GPU 40 min/day A100, free
Inference API Serverless hosted models ~$2/mo credit

πŸ“– License

MIT