Skip to content

Latest commit

 

History

History
68 lines (57 loc) · 2.63 KB

Advanced.md

File metadata and controls

68 lines (57 loc) · 2.63 KB

3D U-Net FP32 inference - Advanced Instructions

This document has advanced instructions for running 3D U-Net FP32 inference, which provides more control over the individual parameters that are used. For more information on using /benchmarks/launch_benchmark.py, see the launch benchmark documentation.

Prior to using these instructions, please follow the setup instructions from the model's README and/or the AI Kit documentation to get your environment setup (if running on bare metal) and download the dataset, pretrained model, etc. If you are using AI Kit, please exclude the --docker-image flag from the commands below, since you will be running the the TensorFlow conda environment instead of docker.

Any of the launch_benchmark.py commands below can be run on bare metal by removing the --docker-image arg. Ensure that you have all of the required prerequisites installed in your environment before running without the docker container.

If you are new to docker and are running into issues with the container, see this document for troubleshooting tips.

Once your environment is setup, navigate to the benchmarks directory of the model zoo and set environment variables pointing to the directory for the dataset, pretrained model, and an output directory where log files will be written.

# cd to the benchmarks directory in the model zoo
cd benchmarks

export DATASET_DIR=<path to the dataset>
export OUTPUT_DIR=<directory where log files will be written>
export PRETRAINED_MODEL=<path to the pretrained model>

3D U-Net FP32 inference can be run with a batch size of 1:

python launch_benchmark.py \
  --precision fp32 \
  --model-name 3d_unet \
  --mode inference \
  --framework tensorflow \
  --docker-image intel/intel-optimized-tensorflow:1.15.2 \
  --in-graph $PRETRAINED_MODEL \
  --data-location ${DATASET_DIR} \
  --batch-size 1 \
  --socket-id 0 \
  --output-dir ${OUTPUT_DIR}

Below is an example tail of the log file:

Loading pre-trained model
Time spent per BATCH: ... ms
Total samples/sec: ... samples/s
Ran inference with batch size 1
Log location outside container: {--output-dir value}/benchmark_3d_unet_inference_fp32_20190116_234659.log