os.makedirs('../results/{}/'.format(model_name))
with open('../results/{}/res.csv'.format(model_name),
'a',
encoding='utf-8') as resfile:
resfile.write('step,loss,rmse\n')
sess.run(tf.global_variables_initializer())
saver = tf.train.Saver()
bad_count = 0
init_rmse = 0.86
last_best_rmse = init_rmse
last_best_loss = 0
for epoch in range(hp.NUM_EPOCHS):
# load trainset
train_batches = data.load_data('train')
# training
for (batch_X_user,
batch_X_movie), batch_Y in tqdm(train_batches):
# unpack
batch_u_oids, batch_bu_seq = batch_X_user
batch_m_oids, batch_info, batch_actors, batch_des, batch_bm_seq = batch_X_movie
batch_r_seq = batch_Y
feed_dict = {
model.m_oids: batch_m_oids,
model.info: batch_info,
model.actors: batch_actors,
model.descriptions: batch_des,
model.u_oids: batch_u_oids,
model.r_seq: batch_r_seq,
model.dropout_keep_prob: hp.DROPOUT_KEEP_PROB
saver.restore(sess, fpath)
# Get the placeholders from the graph by name
m_oids = graph.get_tensor_by_name('movie_order_ids:0')
info = graph.get_tensor_by_name('info:0')
actors = graph.get_tensor_by_name('actors:0')
descriptions = graph.get_tensor_by_name('descriptions:0')
u_oids = graph.get_tensor_by_name('user_order_ids:0')
r_seq = graph.get_tensor_by_name('rating_sequence:0')
dropout_keep_prob = graph.get_tensor_by_name("dropout_keep_prob:0")
# Tensors we want to evaluate
mse_op = graph.get_tensor_by_name('mse/mse_op:0')
# load evalset
eval_iter = data.load_data('eval')
mse, count = 0.0, 0
for (sub_X_user, sub_X_movie), sub_Y in tqdm(eval_iter):
# unpack
sub_u_oids, sub_bu_seq = sub_X_user
sub_m_oids, sub_info, sub_actors, sub_des, sub_bm_seq = sub_X_movie
sub_r_seq = sub_Y
dev_feed_dict = {
m_oids: sub_m_oids,
info: sub_info,
actors: sub_actors,
descriptions: sub_des,
u_oids: sub_u_oids,
r_seq: sub_r_seq,
dropout_keep_prob: hp.DROPOUT_KEEP_PROB}