Step #2: Sample applications
The following examples will require using the system console of the GPU host. Click on the System Console link in the left menu of this page to open a web-based SSH session.
Launch with the script you can find at /data/holoscan/assets/launch_run_container.sh with: bash /data/holoscan/assets/launch_run_container.sh
#!/bin/bash
set -x
sudo xhost +local:docker
IMAGE_NAME="nvcr.io/nvidia/clara-holoscan/clara_holoscan_sample_runtime"
IMAGE_TAG="v0.3.0"
ASSETS_PATH="/data/holoscan/assets"
TRT_ENGINES_PATH=${ASSETS_PATH}/trt_engines
sudo docker run -it --rm \
--runtime=nvidia -e NVIDIA_DRIVER_CAPABILITIES=graphics,video,compute,utility,display \
-e DISPLAY=${DISPLAY} -v /tmp/.X11-unix:/tmp/.X11-unix \
-v ${TRT_ENGINES_PATH}/endoscopy_Tesla-T4_c75_n40.engine:/workspace/test_data/endoscopy/model/tool_loc_convlstm_engines/Tesla-T4_c75_n40.engine \
-v ${TRT_ENGINES_PATH}/ultrasound_Tesla-T4_c75_n40.engine:/workspace/test_data/ultrasound/model/us_unet_256x256_nhwc_engines/Tesla-T4_c75_n40.engine \
-w /opt/holoscan_sdk/ \
${IMAGE_NAME}:${IMAGE_TAG}
We mount:
-v ${trt_engines_path}/endoscopy_Tesla-T4_c75_n40.engine:/workspace/test_data/endoscopy/model/tool_loc_convlstm_engines/Tesla-T4_c75_n40.engine \
-v ${trt_engines_path}/ultrasound_Tesla-T4_c75_n40.engine:/workspace/test_data/ultrasound/model/us_unet_256x256_nhwc_engines/Tesla-T4_c75_n40.engine \
…because it takes a very long time for the ONNX model to be converted to a TensorRT engine at runtime.
From within the container, we can do any of the following:
Running the endoscopy tool tracking app (C++ API) with a recorded video
cd /opt/holoscan_sdk \ && sed -i -e 's#^source:.*#source: replayer#' ./apps/endoscopy_tool_tracking/app_config.yaml \ && ./apps/endoscopy_tool_tracking/endoscopy_tool_tracking
Running the endoscopy tool tracking app (GXF API) with a recorded video
cd /opt/holoscan_sdk && ./apps/endoscopy_tool_tracking_gxf/tracking_replayer
Running the ultrasound spine scoliosis segmentation app (GXF API) with a recorded video
cd /opt/holoscan_sdk && ./apps/ultrasound_segmentation_gxf/segmentation_replayer
Tip: open this documentation in Firefox with address “localhost” in the Desktop VNC environment to copy/paste the commands to the terminal.
To find more details about the sample applications, please see: https://docs.nvidia.com/clara-holoscan/sdk-user-guide/clara_holoscan_applications.html