prediction-flow

prediction-flow copied to clipboard

**Describe the bug** A clear and concise description of what the bug is. **To Reproduce** Steps to reproduce the behavior: 1. Go to '...' 2. Click on '....' 3. Scroll...

jzm-123

Hi， 谢谢大佬的code。想问一下在DIN/DIEN里是否会在未来加入DICE激活函数来取代torch自身提供的prelu? 目前发现的小问题，在gpu运行下(torch 1.9.0)： 在DIEN.py里因为pytorch版本更新带来的一些new behaviour，需要加入 keys_length=keys_length.to(torch.device("cpu")) 作为torch里的改动要求（需要非cuda的cpu long type）。之后在interest.py里加入 keys_length=keys_length.to(torch.device("cuda" if torch.cuda.is_available() else "cpu")) 来改回为GPU。 非常感谢！

YingHH1

How to deal with the multiple lables?

1

I'm frustrated when I realize the ESMM model based on prediction-flow, it has two labels which are click and post-click-conversation lables. It seems like `create_dataloader_fn` should be rewriten?

jellchou

attendance one key and many values

3

pairs=[{'ad': 'q_topic_1', 'pos_hist': 'm_interested_topics'}, {'ad': 'q_topic_2', 'pos_hist': 'm_interested_topics'}, {'ad': 'q_topic_3', 'pos_hist': 'm_interested_topics'}, {'ad': 'q_topic_4', 'pos_hist': 'm_interested_topics'}, {'ad': 'q_topic_5', 'pos_hist': 'm_interested_topics'} ],

liuchenailq