Did anyone find that the code cannot run directly?

[githubWangHC]

No description provided.

[shiguyuanjiu]

我感觉缺东西。q_value计算这里

[WangXuan2401]

So, Is the problem caused by the lack of real data?

[lemonskiller]

I changed
receiver_tensors = {
'user_id': tf.placeholder(tf.int64, [None, 1], name='user_id'),
'behavior_poi_id_list': tf.placeholder(tf.int64, [None, 10], name='behavior_poi_id_list'),
'ad_id_list': tf.placeholder(tf.int32, [None, 5], name='ad_id_list'),
'oi_id_list': tf.placeholder(tf.int32, [None, 5], name='oi_id_list'),
'context_id': tf.placeholder(tf.int32, [None, 1], name='context_id'),
'action': tf.placeholder(tf.int32, [None, 5], name='action')
}
from tf.float to tf.int32 and it works!