Merge pull request #10 from TLX-CTR-Algorithm/liuys

Liuys

DNN/flags.py

@@ -12,6 +12,7 @@ def parse_args(check=True):
  parser.add_argument('--embed_dim', type=int, default=config.embed_dim)
  parser.add_argument('--learning_rate', type=float, default=config.learning_rate)
  parser.add_argument('--oridata_dim', type=int, default=config.oridata_dim)
+ parser.add_argument('--valid_switch', type=int, default=config.valid_switch)
  # 路径和文件配置
  parser.add_argument('--encod_train_path', type=str, default=config.encod_train_path)
  parser.add_argument('--encod_vaild_path', type=str, default=config.encod_vaild_path)

DNN/train.py

@@ -102,21 +102,42 @@ def train_model(batch_size=FLAGS.batch_size):
 
  logging.info('----------------------valid-----------------------')
  #使用验证数据，验证模型性能
- valid_batches = utils.genbatch(valid_inputs, valid_labels, batch_size=FLAGS.batch_size)
- for step in range(len(valid_inputs) // batch_size):
- batch_valid_inputs,batch_valid_lables = next(valid_batches)
- valid_continous_inputs = batch_valid_inputs[:, 0:FLAGS.encod_cat_index_begin]
- valid_categorial_inputs = batch_valid_inputs[:,FLAGS.encod_cat_index_begin:FLAGS.encod_cat_index_end]
- feed_dict = { dnnmodel.categorial_inputs:valid_categorial_inputs, dnnmodel.continous_inputs:valid_continous_inputs, dnnmodel.label:batch_valid_lables, dnnmodel.keep_prob:FLAGS.keep_prob }
- #with tf.Session() as sess:
+ if FLAGS.valid_switch == 0:
+ valid_continous_inputs = valid_inputs[:, 0:FLAGS.encod_cat_index_begin]
+ valid_categorial_inputs = valid_inputs[:, FLAGS.encod_cat_index_begin:FLAGS.encod_cat_index_end]
+ feed_dict = {dnnmodel.categorial_inputs: valid_categorial_inputs,
+ dnnmodel.continous_inputs: valid_continous_inputs, dnnmodel.label: valid_labels,
+ dnnmodel.keep_prob: FLAGS.keep_prob}
  valid_global_step, logits, loss, accuracy, auc, end_points, labels = sess.run([dnnmodel.global_step, dnnmodel.logits, dnnmodel.loss, dnnmodel.accuracy, dnnmodel.auc, dnnmodel.end_points, dnnmodel.label], feed_dict=feed_dict)
- #np.savetxt('./log/logits.log', end_points['logits'])
- #np.savetxt('./log/pre.log', end_points['prediction'])
- #np.savetxt('./log/labels.log', labels)
- if step % FLAGS.logfrequency == 0:
- #每间隔指定的频率打印日志并存储checkpoint文件
- logging.info('valid: step [{0}] loss [{1}] auc [{2}] accuracy [{3}]'.format(global_step, loss, auc, accuracy))
- #saver.save(sess, os.path.join(FLAGS.model_ouput_dir, "model.ckpt"), global_step=global_step)
+ logging.info(
+ 'valid0: step [{0}] loss [{1}] auc [{2}] accuracy [{3}]'.format(global_step, loss, auc,
+ accuracy))
+ else:
+ valid_batches = utils.genbatch(valid_inputs, valid_labels, batch_size=FLAGS.batch_size)
+ loss_list = []
+ auc_list = []
+ accuracy_list = []
+ for step in range(len(valid_inputs) // batch_size):
+ batch_valid_inputs,batch_valid_lables = next(valid_batches)
+ valid_continous_inputs = batch_valid_inputs[:, 0:FLAGS.encod_cat_index_begin]
+ valid_categorial_inputs = batch_valid_inputs[:,FLAGS.encod_cat_index_begin:FLAGS.encod_cat_index_end]
+ feed_dict = { dnnmodel.categorial_inputs:valid_categorial_inputs, dnnmodel.continous_inputs:valid_continous_inputs, dnnmodel.label:batch_valid_lables, dnnmodel.keep_prob:FLAGS.keep_prob }
+ valid_global_step, logits, loss, accuracy, auc, end_points, labels = sess.run([dnnmodel.global_step, dnnmodel.logits, dnnmodel.loss, dnnmodel.accuracy, dnnmodel.auc, dnnmodel.end_points, dnnmodel.label], feed_dict=feed_dict)
+
+ loss_list.append(loss)
+ auc_list.append(auc)
+ accuracy_list.append(accuracy)
+ #np.savetxt('./log/logits.log', end_points['logits'])
+ #np.savetxt('./log/pre.log', end_points['prediction'])
+ #np.savetxt('./log/labels.log', labels)
+ #if step % FLAGS.logfrequency == 0:
+ #每间隔指定的频率打印日志并存储checkpoint文件
+ # logging.info('valid: step [{0}] loss [{1}] auc [{2}] accuracy [{3}]'.format(global_step, loss, auc, accuracy))
+ valid_loss = np.mean(loss_list)
+ valid_auc = np.mean(auc_list,0)
+ valid_accuracy = np.mean(accuracy_list)
+ logging.info( 'valid1: step [{0}] loss [{1}] auc [{2}] accuracy [{3}]'.format(global_step, valid_loss, valid_auc, valid_accuracy))
+
  epoch = (global_step * batch_size) // count_data
  logging.info('has completed epoch:{}'.format(epoch))
  if epoch >= FLAGS.Max_epoch or global_step >= FLAGS.Max_step:

config.py

@@ -31,6 +31,7 @@
 dnn_log_path = os.path.join(dnn_log_dir, dnn_log_file)
 encod_cat_index_begin = 4
 encod_cat_index_end = 30
+valid_switch = 1
 # 训练参数
 batch_size = 1000
 keep_prob = 0.8