Create TensorFlow model

In [ ]:

    

!sudo chown -R jupyter:jupyter /home/jupyter/training-data-analyst


    

In [ ]:

    

# Ensure the right version of Tensorflow is installed.
!pip freeze | grep tensorflow==2.1


    

In [ ]:

    

# change these to try this notebook out
BUCKET = 'cloud-training-demos-ml'
PROJECT = 'cloud-training-demos'
REGION = 'us-central1'


    

In [ ]:

    

import os
os.environ['BUCKET'] = BUCKET
os.environ['PROJECT'] = PROJECT
os.environ['REGION'] = REGION


    

In [ ]:

    

%%bash
if ! gsutil ls | grep -q gs://${BUCKET}/; then
  gsutil mb -l ${REGION} gs://${BUCKET}
fi


    

In [ ]:

    

import shutil
import numpy as np
import tensorflow as tf


    

In [ ]:

    

# Determine CSV, label, and key columns
CSV_COLUMNS = 'weight_pounds,is_male,mother_age,plurality,gestation_weeks,key'.split(',')
LABEL_COLUMN = 'weight_pounds'
KEY_COLUMN = 'key'

# Set default values for each CSV column
DEFAULTS = [[0.0], ['null'], [0.0], ['null'], [0.0], ['nokey']]
TRAIN_STEPS = 1000


    

In [ ]:

    

# Create an input function reading a file using the Dataset API
# Then provide the results to the Estimator API
def read_dataset(filename_pattern, mode, batch_size = 512):
  def _input_fn():
    def decode_csv(line_of_text):
      # TODO #1: Use tf.decode_csv to parse the provided line
      # TODO #2: Make a Python dict.  The keys are the column names, the values are from the parsed data
      # TODO #3: Return a tuple of features, label where features is a Python dict and label a float
      return features, label
    
    # TODO #4: Use tf.gfile.Glob to create list of files that match pattern
    file_list = None

    # Create dataset from file list
    dataset = (tf.compat.v1.data.TextLineDataset(file_list)  # Read text file
                 .map(decode_csv))  # Transform each elem by applying decode_csv fn
    
    # TODO #5: In training mode, shuffle the dataset and repeat indefinitely
    #                (Look at the API for tf.data.dataset shuffle)
    #          The mode input variable will be tf.estimator.ModeKeys.TRAIN if in training mode
    #          Tell the dataset to provide data in batches of batch_size 

    
    # This will now return batches of features, label
    return dataset
  return _input_fn


    

In [ ]:

    

# Define feature columns


    

In [ ]:

    

# Create serving input function to be able to serve predictions later using provided inputs
def serving_input_fn():
    feature_placeholders = {
        'is_male': tf.compat.v1.placeholder(tf.string, [None]),
        'mother_age': tf.compat.v1.placeholder(tf.float32, [None]),
        'plurality': tf.compat.v1.placeholder(tf.string, [None]),
        'gestation_weeks': tf.compat.v1.placeholder(tf.float32, [None])
    }
    features = {
        key: tf.expand_dims(tensor, -1)
        for key, tensor in feature_placeholders.items()
    }
    return tf.compat.v1.estimator.export.ServingInputReceiver(features, feature_placeholders)


    

In [ ]:

    

# Create estimator to train and evaluate
def train_and_evaluate(output_dir):
  EVAL_INTERVAL = 300
  run_config = tf.estimator.RunConfig(save_checkpoints_secs = EVAL_INTERVAL,
                                      keep_checkpoint_max = 3)
  # TODO #1: Create your estimator
  estimator = None
  train_spec = tf.estimator.TrainSpec(
                       # TODO #2: Call read_dataset passing in the training CSV file and the appropriate mode
                       input_fn = None,
                       max_steps = TRAIN_STEPS)
  exporter = tf.estimator.LatestExporter('exporter', serving_input_fn)
  eval_spec = tf.estimator.EvalSpec(
                       # TODO #3: Call read_dataset passing in the evaluation CSV file and the appropriate mode
                       input_fn = None,
                       steps = None,
                       start_delay_secs = 60, # start evaluating after N seconds
                       throttle_secs = EVAL_INTERVAL,  # evaluate every N seconds
                       exporters = exporter)
  tf.estimator.train_and_evaluate(estimator, train_spec, eval_spec)


    

In [ ]:

    

# Run the model
shutil.rmtree('babyweight_trained', ignore_errors = True) # start fresh each time
tf.compat.v1.summary.FileWriterCache.clear()
train_and_evaluate('babyweight_trained')