"model_name": [],
"left_out": [],
"top_from": [],
"measure": [],
"mean": [],
"sd": []
}
left_out = [1]
tops = [5]
# print(left_out)
for model in models_name:
model = model.replace("\n", "")
restored_model = keras.models.load_model(model)
for lo in left_out:
for top in tops:
dd = pussy.predict_users_all(users_test, seqs, restored_model, lo,
top, n_movies)
d = pussy.append_to_dict(d, dd, lo, top, model)
df = pd.DataFrame(d)
df.to_csv(path_or_buf="predict_test_{}.csv".format(file_models),
sep=',',
index=False)
log.write("model restored\n")
restored_model.summary(print_fn=lambda x: log.write(x + '\n'))
users_test = np.load(users_test_file)
seqs = pussy.make_vectors(users_test, ratings)
seqs = pussy.make_vectors(users_train, ratings)
print(users_train.shape)
d = {"left_out": [], "top_from": [], "measure": [], "mean": [], "sd": []}
print(left_out)
dd = pussy.predict_users_all(users_train[:1000], seqs, restored_model, 1, 10,
n_movies)
d = pussy.append_to_dict(d, dd, left_out, 10)
df = pd.DataFrame(d)
df.to_csv(path_or_buf="fdfsdfsdf18", sep=',', index=False)
sys.exit(1)
#
# train model
#
start_time = datetime.datetime.now()
log.write("start training at {}\n".format(start_time))