51 lines
1.5 KiB
Python
51 lines
1.5 KiB
Python
|
import pandas
|
||
|
|
import os
|
||
|
|
from sacred import Experiment
|
||
|
|
from sacred.observers import MongoObserver
|
||
|
|
from sacred.observers import FileStorageObserver
|
||
|
|
from keras.applications.densenet import layers
|
||
|
|
import tensorflow
|
||
|
|
import random
|
||
|
|
|
||
|
|
exint = Experiment('z-s434743-training')
|
||
|
|
|
||
|
|
exint.observers.append(MongoObserver(url='mongodb://admin:IUM_2021@172.17.0.1:27017',
|
||
|
|
db_name='sacred'))
|
||
|
|
exint.observers.append(FileStorageObserver('z-s434743-training/master'))
|
||
|
|
|
||
|
|
@exint.config
|
||
|
|
def my_config():
|
||
|
|
EPOCHS = int(os.environ['EPOCHS'])
|
||
|
|
|
||
|
|
|
||
|
|
@exint.main
|
||
|
|
def my_main(EPOCHS, _run):
|
||
|
|
_run.info["epochs"] = EPOCHS,
|
||
|
|
normalize = layers.Normalization()
|
||
|
|
train_data_x = pandas.read_csv('./X_train.csv')
|
||
|
|
_run.info["data"] = train_data_x
|
||
|
|
movies_all = train_data_x.copy()
|
||
|
|
movies_predict = train_data_x.pop('audienceScore')
|
||
|
|
normalize.adapt(movies_all)
|
||
|
|
norm_movies_model = tensorflow.keras.Sequential([
|
||
|
|
normalize,
|
||
|
|
layers.Dense(64),
|
||
|
|
layers.Dense(1)
|
||
|
|
])
|
||
|
|
norm_movies_model.compile(
|
||
|
|
loss=tensorflow.keras.losses.MeanSquaredError(),
|
||
|
|
optimizer=tensorflow.keras.optimizers.Adam())
|
||
|
|
|
||
|
|
norm_movies_model.fit(movies_all, movies_predict, epochs=EPOCHS)
|
||
|
|
counter = 0
|
||
|
|
while counter < 20:
|
||
|
|
counter+=1
|
||
|
|
value = counter * random.randint(5, 5000)
|
||
|
|
_run.log_scalar("training.accuracy", value * 2)
|
||
|
|
norm_movies_model.save('test')
|
||
|
|
|
||
|
|
exint.run()
|
||
|
|
|
||
|
|
exint.add_artifact('test/saved_model.pb')
|
||
|
|
exint.add_source_file('./ium_z434743/train.py')
|