tf_export.srun

#!/bin/bash
#SBATCH --job-name=tfexport
#SBATCH --output=slurm_logs/tf_export_%j.out
#SBATCH --error=slurm_logs/tf_export_%j.err
#SBATCH --ntasks=1
#SBATCH --cpus-per-task=8
#SBATCH --nodes=1-1
#SBATCH --gres=gpu:tesla:1
#SBATCH --partition=taylor,kamiak,cahnrs_gpu,free_gpu,vcea,cahnrs,cahnrs_bigmem
#SBATCH --time=0-00:30:00
#SBATCH --mem=20G

. config.py

#
# ---
#

TFArch=$1
if [[ -z $TFArch ]]; then
    echo "Specify which model to train."
    exit 1
else
    echo "TFArch: $TFArch"
fi

export OMP_NUM_THREADS=$SLURM_CPUS_PER_TASK
module load cuda/8.0.44 cudnn/6.0_cuda8.0 python3/3.5.0
data="$remotedir"

echo "Making sure TensorFlow installed: starting"
pip install --user tensorflow-gpu pillow lxml jupyter matplotlib
echo "Making sure TensorFlow installed: done"

# Get the last checkpoint prefix, e.g.
# path/to/tflogs/train/ssd_inception_v2/model.ckpt-200000
lastCheckpoint="$(ls "$data/$datasetFolder/$datasetTFtrainlogs/$TFArch/"model.ckpt-* | \
    sort | tail -n 1 | sed -r 's#\.[^\.]+$##g')"
outputDir="$data/$datasetFolder/${TFArch}.pb"
echo "Last checkpoint: $lastCheckpoint"
echo "Saving to: $outputDir"

echo "Exporting network: started"
cd models/research
export PYTHONPATH="$PYTHONPATH:$(pwd):$(pwd)/slim"
python3 object_detection/export_inference_graph.py \
    --input_type image_tensor \
    --pipeline_config_path="$data/$datasetFolder/tf_${TFArch}.config" \
    --trained_checkpoint_prefix="$lastCheckpoint" \
    --output_directory="$outputDir"
echo "Exporting network: done"