ium_434765/neural_network.py

69 lines
2.0 KiB
Python
Raw Normal View History

2021-05-20 23:48:25 +02:00
import warnings
2021-04-24 21:18:57 +02:00
import pandas as pd
2021-04-24 22:23:04 +02:00
import numpy as np
from tensorflow import keras
2021-05-17 21:36:02 +02:00
import sys
2021-05-20 23:48:25 +02:00
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)
2021-05-17 19:24:30 +02:00
def normalize_data(data):
return (data - np.min(data)) / (np.max(data) - np.min(data))
2021-05-17 21:04:57 +02:00
data = pd.read_csv("data_train", sep=',', skip_blank_lines=True, nrows=1087, error_bad_lines=False,
names=["video_id", "last_trending_date", "publish_date", "publish_hour", "category_id",
"channel_title", "views", "likes", "dislikes", "comment_count"]).dropna()
X = data.loc[:,data.columns == "views"].astype(int)
y = data.loc[:,data.columns == "likes"].astype(int)
2021-04-24 22:23:04 +02:00
2021-05-17 19:24:30 +02:00
min_val_sub = np.min(X)
max_val_sub = np.max(X)
X = (X - min_val_sub) / (max_val_sub - min_val_sub)
print(min_val_sub)
print(max_val_sub)
2021-04-24 22:23:04 +02:00
2021-05-17 19:24:30 +02:00
min_val_like = np.min(y)
max_val_like = np.max(y)
y = (y - min_val_like) / (max_val_like - min_val_like)
2021-04-24 22:23:04 +02:00
2021-05-17 19:24:30 +02:00
print(min_val_like)
print(max_val_like)
2021-04-24 22:23:04 +02:00
2021-05-20 23:48:25 +02:00
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))
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'),
])
2021-04-24 22:23:04 +02:00
2021-05-20 23:48:25 +02:00
model.compile(loss='mean_absolute_error', optimizer="Adam", metrics=['mean_absolute_error'])
2021-04-24 22:23:04 +02:00
2021-05-20 23:48:25 +02:00
model.fit(X, y, epochs=int(sys.argv[1]), validation_split = 0.3)
2021-04-24 22:23:04 +02:00
2021-05-20 23:48:25 +02:00
model.save('model')
2021-04-24 22:23:04 +02:00
2021-05-20 23:48:25 +02:00
error = evaluate_model()
mlflow.log_metric("rmse", error)