Skip to content

Latest commit

 

History

History
35 lines (24 loc) · 997 Bytes

README.md

File metadata and controls

35 lines (24 loc) · 997 Bytes

Segment Annotator

An annotation tool for creating labeled segmentation datasets.

Installation

Clone the repository locally:

git clone https://github.com/Steven-Holland/segment-annotator.git

Create the conda env using python>=3.10. Please follow the instructions here to install both PyTorch and TorchVision dependencies. Installing both PyTorch and TorchVision with CUDA support is strongly recommended.

conda create -n Annotator python=3.10
conda activate Annotator

Install the packages:

cd segment-annotator
pip install -r requirements.txt

Download SAM and FastSAM model checkpoints. Once they are downloaded, move them to ../src/assets/models.

Usage

Move into the src folder and run main.py

cd src
python main.py