Training cars on Google colab

colab_tpot_totalcheck.py

#from google.colab import files;uploaded = files.upload()
#!pip install TPOT
from tpot import TPOTRegressor
from sklearn.model_selection import train_test_split
df=pd.read_csv('32928_chile_autos.csv')
df=df[(df.Año>1980)&(df.Kilometraje>1000)&(df.Precio<100000000)&(df.Precio>1000000)]

data_columns = df.columns[2:-1]   # add some more!
target_columns = df.columns[-1]   # use logs!
X,y = df[data_columns], df[target_columns]
tt = train_test_split(X.values, y.values,train_size=0.75, test_size=0.25)
Xtrain,Xtest,ytrain,ytest=tt

tp = TPOTRegressor(generations, population_size, verbosity=verbosity,
                scoring='negative_absolute_mean_error',
                config_dict='TPOT MDR',
                scoring='balanced_accuracy')

tp.fit(Xtrain, ytrain)
y_fit = tp.predict(X_test)
tp.export(os.path.join(MODEL_FOLDER, 'tpot_cars_pipeline.py'))
print(tp.score(Xtest, ytest))