Environment Configuration

deployment configuration env

The application behavior is controlled by environment variables, managed via a .env file in the project root.

Variable Reference

Variable	Description	Default / Example
ONNX_CHECKPOINT_FILENAME	Filename of the .onnx model to load from models/.	last-checkpoint-signs_502.pth.onnx
DOMAIN_NAME	Allowed origin for CORS and WebSocket connections.	http://localhost:8000
LOCAL_DEV	Flag to indicate local development environment. Affects data paths.	1
USE_CPU	Force PyTorch/ONNX to use CPU instead of CUDA.	1
OMP_NUM_THREADS	Number of threads for OpenMP (Linear Algebra).	1 (Container Default)

Setup

1. Copy the example file:

   cp .env.example .env

2. Edit .env to match your local setup.

Constants (src/core/constants.py)

Some configurations are hardcoded constants but can be influenced by environment flags:

• MODELS_DIR: Directory for model checkpoints.
• DATA_DIR: Root data directory.

Related Documentation

• constants.py Source Code
• Docker Setup