README

# AMD64 with Nvidia GPUs
## Base container
- Base: `nvcr.io/nvidia/cuda:11.4.3-cudnn8-devel-ubuntu20.04`
- Specs:
    - gRPC 1.60.2
    - OpenCV 4.8.1 (with CUDA Support)
        - Check and add GPU Compute Capability (e.g., 3090 uses 8.6) if necessary.
    - TensorRT 8.4.3.1
    - CUDA 11.4
    - CuDNN 8.4.1
    - Boost 1.84
    - Torch 1.11
- Build command:
    ```
    docker build -t pipeplusplus:base-amd64 -f ./DOCKERFILE-base-amd64 .
    ```

# Jetson devices
- Base: `nvcr.io/nvidia/l4t-ml:${JETPACK_VERSION}`
    * r35.2.1-py3 for the orins
    * r35.1.0-py3 for the xaviers
- Build command
    ```
    docker build --build-arg JETPACK_VERSION="r35.1.0-py3" -t pipeline-scheduler:r35.1.0-py3 -f DOCKERFILE-jetson .
    ```