How to setup airflow

Set airflow directory

export AIRFLOW_HOME="/home/avnish/census_consumer_project/census_consumer_complaint/airflow"

To install airflow

pip install apache-airflow

To configure databse

airflow db init

To create login user for airflow

airflow users create  -e avnish@ineuron.ai -f Avnish -l Yadav -p admin -r Admin  -u admin

To start scheduler

airflow scheduler

To launch airflow server

airflow webserver -p <port_number>

pip install pandas-tfrecords

pip install \
  --upgrade --ignore-installed \
  python-snappy==0.5.1 \
  --global-option=build_ext \
  --global-option="-I/usr/local/include" \
  --global-option="-L/usr/local/lib"

pip install twine python setup.py sdist bdist_wheel twine upload --repository-url https://test.pypi.org/legacy/ dist/* twine upload dist/*

To deploy your model

pip install tensorflow-serving-api

to inspect model

saved_model_cli show --dir <dir_path>

Above command will return tag set

saved_model_cli show --dir <dir_path> --tag_set <tag_name>

Above command will show available model signatures

Next: with tag_set and signature_def info, we can inspect model input and output

saved_model_cli show --dir <dir_path> --tag_set <tag_name> --signature_def <SignatureDef Key>

To inspect all signature without tag_set and signature_def saved_model_cli show --dir <dir_path> --all

Testing the model

Test model prediction using saved_model_cli with sample input data

--input_examples: input data formatted as a tf.Example data structure

other param

--outdir: by default output will be written in terminal

--overwrite: to write into a file

tf_debug: run in debug mode

To expose your model as an API using docker image tensorflow/serving

docker pull tensorflow/serving

Single model configuration

sudo docker run -p 8500:8500 \
-p 8501:8501\
--volumn <model_dir>:<target_dir>\
-e MODEL_NAME=<model_name>\
-e model_base_path=<target_dir>\
-t tensorflow/serving:latest

sudo docker run -p 8500:8500 -p 8501:8501 \
-v  /home/avnish/census_consumer_project/census_consumer_complaint/census_consumer_complaint_data/saved_models:/avnish/my_model \
-e MODEL_NAME=my_model \
-e MODEL_BASE_PATH=/avnish \
-t tensorflow/serving:latest

To inspect docker container directory

docker exec -it <conatiner_name> bash