mlflow

MLproject

@@ -0,0 +1,12 @@
+name: tutorial
+
+docker_env:
+  image: karopa/ium:27
+
+entry_points:
+  main:
+    parameters:
+      epochs: {type: float, default: 30}
+    command: "python3 neural_network.py {epochs}"
+  test:
+    command: "python3 evaluate_network.py"

evaluate_network.py

@@ -4,47 +4,50 @@ from sklearn.metrics import mean_squared_error
 from tensorflow import keras
 import matplotlib.pyplot as plt
 
-model = keras.models.load_model('model')
-data = pd.read_csv("data_dev", sep=',', error_bad_lines=False,
-                   skip_blank_lines=True, nrows=527, names=["video_id", "last_trending_date",
-                                                            "publish_date", "publish_hour", "category_id",
-                                                            "channel_title", "views", "likes", "dislikes",
-                                                            "comment_count"]).dropna()
-X_test = data.loc[:, data.columns == "views"].astype(int)
-y_test = data.loc[:, data.columns == "likes"].astype(int)
 
-min_val_sub = np.min(X_test)
-max_val_sub = np.max(X_test)
-X_test = (X_test - min_val_sub) / (max_val_sub - min_val_sub)
-print(min_val_sub)
-print(max_val_sub)
+def evaluate_model():
+    model = keras.models.load_model('model')
+    data = pd.read_csv("data_dev", sep=',', error_bad_lines=False,
+                       skip_blank_lines=True, nrows=527, names=["video_id", "last_trending_date",
+                                                                "publish_date", "publish_hour", "category_id",
+                                                                "channel_title", "views", "likes", "dislikes",
+                                                                "comment_count"]).dropna()
+    X_test = data.loc[:, data.columns == "views"].astype(int)
+    y_test = data.loc[:, data.columns == "likes"].astype(int)
 
-min_val_like = np.min(y_test)
-max_val_like = np.max(y_test)
-print(min_val_like)
-print(max_val_like)
+    min_val_sub = np.min(X_test)
+    max_val_sub = np.max(X_test)
+    X_test = (X_test - min_val_sub) / (max_val_sub - min_val_sub)
+    print(min_val_sub)
+    print(max_val_sub)
 
-prediction = model.predict(X_test)
+    min_val_like = np.min(y_test)
+    max_val_like = np.max(y_test)
+    print(min_val_like)
+    print(max_val_like)
 
-prediction_denormalized = []
-for pred in prediction:
-    denorm = pred[0] * (max_val_like[0] - min_val_like[0]) + min_val_like[0]
-    prediction_denormalized.append(denorm)
+    prediction = model.predict(X_test)
 
-f = open("predictions.txt", "w")
-for (pred, test) in zip(prediction_denormalized, y_test.values):
-    f.write("predicted: %s expected: %s\n" % (str(pred), str(test[0])))
+    prediction_denormalized = []
+    for pred in prediction:
+        denorm = pred[0] * (max_val_like[0] - min_val_like[0]) + min_val_like[0]
+        prediction_denormalized.append(denorm)
 
-error = mean_squared_error(y_test, prediction_denormalized)
-print(error)
+    f = open("predictions.txt", "w")
+    for (pred, test) in zip(prediction_denormalized, y_test.values):
+        f.write("predicted: %s expected: %s\n" % (str(pred), str(test[0])))
 
-with open("rmse.txt", "a") as file:
-    file.write(str(error) + "\n")
+    error = mean_squared_error(y_test, prediction_denormalized)
+    print(error)
 
-with open("rmse.txt", "r") as file:
-    lines = file.readlines()
-    plt.plot(range(len(lines)), [line[:-2] for line in lines])
-    plt.tight_layout()
-    plt.ylabel('RMSE')
-    plt.xlabel('evaluation no')
-    plt.savefig('evaluation.png')
+    with open("rmse.txt", "a") as file:
+        file.write(str(error) + "\n")
+
+    with open("rmse.txt", "r") as file:
+        lines = file.readlines()
+        plt.plot(range(len(lines)), [line[:-2] for line in lines])
+        plt.tight_layout()
+        plt.ylabel('RMSE')
+        plt.xlabel('evaluation no')
+        plt.savefig('evaluation.png')
+    return error

neural_network.py

@@ -1,9 +1,24 @@
+import warnings
+
 import pandas as pd
 import numpy as np
-from sklearn.metrics import mean_squared_error
 
 from tensorflow import keras
 import sys
+import mlflow.sklearn
+
+import logging
+
+from evaluate_network import evaluate_model
+
+logging.basicConfig(level=logging.WARN)
+logger = logging.getLogger(__name__)
+
+mlflow.set_tracking_uri("http://localhost:5000")
+mlflow.set_experiment("s434765")
+
+warnings.filterwarnings("ignore")
+np.random.seed(40)
 
 def normalize_data(data):
     return (data - np.min(data)) / (np.max(data) - np.min(data))
@@ -29,17 +44,25 @@ y = (y - min_val_like) / (max_val_like - min_val_like)
 print(min_val_like)
 print(max_val_like)
 
+with mlflow.start_run() as run:
+    print("MLflow run experiment_id: {0}".format(run.info.experiment_id))
+    print("MLflow run artifact_uri: {0}".format(run.info.artifact_uri))
 
-model = keras.Sequential([
-    keras.layers.Dense(512,input_dim = X.shape[1], activation='relu'),
-    keras.layers.Dense(256, activation='relu'),
-    keras.layers.Dense(256, activation='relu'),
-    keras.layers.Dense(128, activation='relu'),
-    keras.layers.Dense(1,activation='linear'),
-])
+    mlflow.keras.autolog()
+    mlflow.log_param("epochs", int(sys.argv[1]))
+    model = keras.Sequential([
+        keras.layers.Dense(512,input_dim = X.shape[1], activation='relu'),
+        keras.layers.Dense(256, activation='relu'),
+        keras.layers.Dense(256, activation='relu'),
+        keras.layers.Dense(128, activation='relu'),
+        keras.layers.Dense(1,activation='linear'),
+    ])
 
-model.compile(loss='mean_absolute_error', optimizer="Adam", metrics=['mean_absolute_error'])
+    model.compile(loss='mean_absolute_error', optimizer="Adam", metrics=['mean_absolute_error'])
 
-model.fit(X, y, epochs=int(sys.argv[1]), validation_split = 0.3)
+    model.fit(X, y, epochs=int(sys.argv[1]), validation_split = 0.3)
 
-model.save('model')
+    model.save('model')
+
+    error = evaluate_model()
+    mlflow.log_metric("rmse", error)