esenthil2018

# List all saved artifacts
lineapy.catalog()

esenthil2018

pd.DataFrame(test_results, index=['Mean absolute error [MPG]']).T

esenthil2018

#DNN Model
def build_and_compile_model(norm):
  model = keras.Sequential([
      norm,
      layers.Dense(64, activation='relu'),
      layers.Dense(64, activation='relu'),
      layers.Dense(1)
  ])

  model.compile(loss='mean_absolute_error',

esenthil2018

#Linear regression Model
linear_model = tf.keras.Sequential([
    normalizer,
    layers.Dense(units=1)
])

linear_model.predict(train_features1[:10])

linear_model.layers[1].kernel

esenthil2018

model_artifact = lineapy.save(dnn_model, 'dnn_model')

#Assets written to: ram://044e58dc-eeb9-42b8-ab3a-b99c11c66927/assets   

esenthil2018

#The below is the code generated by get_code() method.
import pickle

import lineapy
import tensorflow as tf
from lineapy.utils.utils import prettify
from tensorflow import keras
from tensorflow.keras import layers

normalizer = tf.keras.layers.Normalization(axis=-1)

esenthil2018

train_art = lineapy.get("train_data")
train_art
#LineaArtifact(name='train_data', _version=0)
y_art = lineapy.get("train_labels")
y_art
#LineaArtifact(name='train_labels', _version=0)
model_art = lineapy.get("dnn_model")
model_art
#LineaArtifact(name='dnn_model', _version=0)

esenthil2018

import os
directory = lineapy.to_pipeline(
    [train_art.name,y_art.name, model_art.name], 
    framework = 'AIRFLOW',
    pipeline_name = "dnn_pipeline",
    dependencies = {'dnn_pipeline_dnn_model':{'dnn_pipeline_train_data','dnn_pipeline_y'}},
    output_dir = os.environ.get("AIRFLOW_HOME","~/airflow")+"/dags")

esenthil2018

# Store the variable as an artifact
train_artifact = lineapy.save(train_features, "train_data")

# Check object type
print(type(train_artifact))

# Store the variable as an artifact
train_labels = lineapy.save(train_labels, "train_labels")

# Check object type

esenthil2018

dataset = raw_dataset.copy()
dataset.tail()

dataset.isna().sum()
#drop na
dataset = dataset.dropna()

dataset['Origin'] = dataset['Origin'].map({1: 'USA', 2: 'Europe', 3: 'Japan'})
#one hot encoding
dataset = pd.get_dummies(dataset, columns=['Origin'], prefix='', prefix_sep='')