FaSTA*: Fast-Slow Toolpath Agent with Subroutine Mining for Efficient Multi-turn Image Editing

📌 This repository is under construction. Some subtasks/tools are not fully supported yet.

---

Installation

1. Clone the Repository

git clone https://github.com/rishieraj/FaSTAR.git
cd FaSTAR  

2. Install Dependencies

Create a conda environment with Python 3.10 as per the following commands. Then install the dependencies from the requirements.txt file.

conda create -n fastar python=3.10
conda activate fastar
pip install -r requirements.txt  

3. Download Pre-trained Checkpoints

The required pre-trained model checkpoints must be downloaded from Google Drive and placed in the checkpoints/ folder. Please use this link to download the checkpoints.

---

Usage

Note: The API keys for OpenAI and StabilityAI need to be set in the high_level.py file before executing. To execute FaSTA*, run:

python high_level.py --image path/to/image.png --prompt "Edit this image" --output output.json --output_image final.png --alpha 0  

Example:

python high_level.py --image inputs/sample.jpg --prompt "Replace the cat with a dog and expand the image" --output Tree.json --output_image final_output.png --alpha 0

• --image: Path to input image.
• --prompt: Instruction for editing.
• --output: Path to save generated subtask tree.
• --output_image: Path to save the final output.
• --alpha: Cost-quality trade-off parameter.

---