Getting Started

git clone https://github.com/mandarwagh9/openeyes.git
cd openeyes
pip install -r requirements.txt

ls /dev/video*
python3 -c "import cv2; cap = cv2.VideoCapture(0); print('OK' if cap.isOpened() else 'Error'); cap.release()"

ls -la /dev/video*
gst-launch-1.0 nvarguscamerasrc sensor-id=0 ! fakesink

# With debug window
python src/main.py --debug

# Headless
python src/main.py

# With ROS2
python src/main.py --ros2

# With world model (predictive tracking)
python src/main.py --world-model lewm --follow

# With industry template
python src/main.py --template warehouse --debug

# Turbo mode for max FPS
python src/main.py --turbo --world-model lewm

# Process video files
python src/main.py --video input.mp4 --output output.mp4

sudo bash scripts/jetson_perf.sh
# Sets MAXN SUPER mode + jetson_clocks

# LeWM world model (15M params, 100-200 Hz planning)
python src/main.py --world-model lewm --follow

# With predictive safety evaluation
python src/main.py --world-model lewm --safety-predict

# Industry templates (warehouse, manufacturing-qa, agriculture, retail)
python src/main.py --template warehouse
python src/main.py --template manufacturing-qa

# Depth model selection (midas-small, da3-small, da3-base)
python src/main.py --depth-model da3-small

# Enable LIDAR processing
python src/main.py --lidar --lidar-topic /scan

# Enable RealSense D455
python src/main.py --realsense

# Multi-camera mode
python src/main.py --multi-camera

# INT8 quantization (~2x faster)
python src/main.py --int8

# INT8 + DLA (~3x faster)
python src/main.py --int8 --dla

# Maximum FPS
python src/main.py --no-face --no-gesture --no-pose --no-depth --int8 --dla

# Enable safety controller
python src/main.py --safety --max-velocity 0.5 --min-distance 0.5

# Enable health monitoring
python src/main.py --health-monitor

# Full production mode
python src/main.py --safety --health-monitor --ota-update