prepare_data_train.sh

INPUT_DIR="./input"
if [ ! -d "$INPUT_DIR" ]; then
    mkdir -p $INPUT_DIR
fi

pushd "$INPUT_DIR"
echo "Downloading dataset (train + val + test) from Kaggle..."
kaggle competitions download -c lyft-motion-prediction-autonomous-vehicles
echo "Unpacking..."
unzip lyft-motion-prediction-autonomous-vehicles.zip
popd

if [ ! -d "$INPUT_DIR/scenes" ]; then
    mkdir -p "$INPUT_DIR/scenes"
fi
pushd "$INPUT_DIR/scenes"
echo "Downloading Full Training dataset (train_XXL) from Kaggle..."
kaggle datasets download philculliton/lyft-full-training-set
echo "Unpacking..."
unzip lyft-full-training-set.zip
mv train_full.zarr train_XXL.zarr
popd

pushd src/1st_level
echo "Prerendering val dataset..."
python prerender_raster.py --dset_name val --scene_step 16 --skip_frame_step 0 --initial_scenes 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 --num_jobs 16 --dir_name pre_render_h01248_XXL

echo "Prerendering train_XXL dataset..."
python prerender_raster.py --dset_name train_XXL --scene_step 32 --skip_frame_step 7 --initial_scenes 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 --num_jobs 32 --dir_name pre_render_h01248_XXL
popd