Getting Started

After receiving your license file and gaining access to our container registry, you can follow the steps below to get Redact-Edge up and running.

Step 1: Pull the Docker Image

Once you've been granted access to the container registry, you will need to pull the Docker image that corresponds to the targeted version of Redact-Edge.

Log in to our container registry. We provide you with a key for this purpose. Use the following command to log in:
```
docker login docker.brighter.ai -u enterprise -p <key>
```
Use the following command to pull the image:
```
docker pull docker.brighter.ai/redact-edge:<redact_version>
```
Replace <redact_version> with the specific version of the image you are targeting.

Step 2: Run the Docker Container

After the Docker image is pulled, you can run the container using the following command:

    docker run -it -d \
    --gpus all \
    --user "$(id -u):$(id -g)" \
    -e CUDA_VISIBLE_DEVICES=0 \
    -v <input_folder_path>:/input \
    -v <output_folder_path>:/output \
    -v <license_file_path>:/license.json \
    docker.brighter.ai/redact-edge:<redact_version>

Here’s a description of the input fields in the command:

<input_folder_path>: The local directory on your host system that will be mounted to the /input directory inside the container. This is where the container will access the input data for processing.
<output_folder_path>: The local directory on your host system that will be mounted to the /output directory inside the container. This is where the container will store the processed output data.
<license_file_path>: The location of the license file you received from brighterAI. It should be mounted to /root/license.json inside the container to activate Redact-Edge.
<rtsp_output_stream_host_port>: The port on the host machine where the RTSP stream will be exposed.
<rtsp_output_stream_container_port>: The port inside the container used to run the anonymization pipeline.
This port must be exposed by the container so that the host machine (and potentially external clients) can access the anonymized RTSP stream.
<redact_version>: The version of the Redact-Edge docker image.

This will start the container and launch Redact-Edge. Since DeepStream requires optimized models, the container will begin optimizing the models upon startup.

Troubleshooting

If the container does not start and the logs do not include the message Optimization for model finished., this likely indicates an issue with the TensorRT (TRT) model optimization. Follow these steps to resolve the issue:

Check GPU Accessibility:

Ensure the GPU is accessible within the container by verifying that the NVIDIA runtime is set. You can do this either by configuring it as the default runtime in daemon.json or by adding --runtime=nvidia and --gpus all to the docker run command.