aidetector-backend

Requirements

• Python 3.11.6

Installation

Follow these steps to set up and use the DMimageDetection project:

1. Clone the repository:

git clone https://github.com/GetHypercube/aidetector-backend.git

2. Create a Python virtual environment:

python -m venv .venv

3. Activate the Python environment:

• For Windows:

  .venv\Scripts\activate
  

• For OSX:

  source .venv/bin/activate
  

4. Upgrade pip:

python -m pip install --upgrade pip

5. Install the necessary requirements:

pip install -r requirements.txt

6. If you have a GPU, make sure Pytorch is compiled with GPU support. Here is the example installation command for Windows:

pip install torch --index-url https://download.pytorch.org/whl/cu124

Please go to the following link to find the detailed instructions.

7. Download the weights from this link and place them under /weights directory in your project folder.

8. You need to create a .env file with the following variables:

AWS_ACCESS_KEY=XXX
AWS_SECRET_KEY=XXX
OPENAI_API_KEY=XXX
MONGODB_URL=XXX
DEV_ENV=true

9. If you are going to use the API, please make sure to have the AWS CLI installed and properly configured.

Usage

To run the detector on an image:

python main.py --image_path test.jpg

To run the detector on a folder:

python main.py --image_folder test.jpg

To run the detector on a folder and a specific models:

python main.py --image_folder image.jpg --model dMDetectorResults

If you want to use the API, first run the server:

python api.py

To create the CSV file neccesary to calculate evaluations of the models:

python main.py --image_folder tests --true_label True

And to calculate evaluations, open evaluations.ipynb notebook.

In Windows, run the CURL command:

curl -X POST -F "file=@real_biggan.png" http://localhost:80/detect

Please see main.py for additional options

Docker usage

To build and run the container for development (CPU):

docker build -t aidetector-backend .
docker run --rm -p 8080:8080 aidetector-backend

To build and run the container for development (GPU):

docker build -t aidetector-backend-gpu -f Dockerfile.gpu .
docker run --rm --gpus all -p 8080:8080 aidetector-backend-gpu