train_model.py

from sklearn.metrics import classification_report
from sklearn.model_selection import GridSearchCV
from sklearn.linear_model import LogisticRegression
import argparse
import pickle
import h5py

ap = argparse.ArgumentParser()
ap.add_argument("-d","--db",required=True,help="path to input dataset")
ap.add_argument("-m","--model",required=True)
ap.add_argument("-j","--jobs",type =int,default=-1)
args = vars(ap.parse_args())

db = h5py.File(args["db"],"r")
i= int(db['labels'].shape[0]*0.75)

print("[INFO] tuning hyperparameters")
params ={"C":[0.1,1.0,10.0,100.0]}
model= GridSearchCV(LogisticRegression(),params,cv=2,n_jobs=args["jobs"])
model.fit(db["features"][:i],db["labels"][:i])

print("[INFO] best hyperparameter: {}".format(model.best_params_))

print("[INFO] evaluating")

preds = model.predict(db["features"][i:])
print(classification_report(db["labels"][i:],preds))

#serialize the model to disk 

print("[INFO] saving model..")

f = open(args["model"],"wb")
f.write(pickle.dumps(model.best_estimator_))
f.close()

db.close()