rl-on-gcp output directory is not being created when using --job-dir

[RyanRizzo96]

I am trying to follow the rl-on-gcp. I recreated the file structure on my local machine and proceed to train locally.

The train model locally commands in the tutorial are as follows:

> OUTPUT_DIR=rl_model 
> JOBNAME=rl_train_$(date -u +%y%m%d_%H%M%S)
> REGION=us-central1 
> PACKAGE_PATH=$PWD/rl_model_code/trainer 
> export PYTHONPATH=${PYTHONPATH}:${PWD}/rl_model_code 
> rm $OUTPUT_DIR
> 
> gcloud ml-engine local train\
>     --package-path=$PACKAGE_PATH\
>     --module-name=trainer.task\
>     --\
>     --outdir=$OUTPUT_DIR

The actual instructions given on the tutorial seem to be outdated, for example --outdir should be --job-dir, and gcloud ml-engine should be gcloud ai-platform.

I run the following working commands which worked on another tutorial in the official google docs:

gcloud ai-platform local train   --package-path trainer   --module-name trainer.task   --job-dir rl-model

#Result:

I am getting no errors however the output directory is not being created. Here is the console output:

/anaconda3/envs/.../.../python3.6/site-packages/h5py/__init__.py:36: FutureWarning: Conversion of the second argument of issubdtype from `float` to `np.floating` is deprecated. In future, it will be treated as `np.float64 == np.dtype(float).type`.
  from ._conv import register_converters as _register_converters
Usage:
    trainer.task --outdir=<name>
                 [--eval]
                 [--n_games_per_update=<N>]
                 [--n_hidden=<N>]
                 [--discount_rate=<rate>]
                 [--learning_rate=<rate>]
Options:
    -h, --help  Show this screen and exit.
    --eval  If in eval, make a gif.
    --outdir=<name> Location to save model (or of saved model if eval).
    --n_games_per_update=<N> Number of games to play.  [default: 10]
    --n_hidden=<N>  # of hidden units [default: 10]
    --discount_rate=<rate>  Reward discount rate. [default: 0.95]
    --learning_rate=<rate>  Learning rate. [default: 0.01]

Any suggestions as to why the job directory is not being created? How can I resolve this? Thanks in advance.