i-jayus / recsystem-pytorch

推荐系统论文算法实现，包括序列推荐，多任务学习，元学习等。 Recommendation system papers implementations, including sequence recommendation, multi-task learning, meta-learning, etc.

Hi there 👋

I am Jayus. A master student in Chinese University of Hongkong, Shenzhen, major in data science. I try to do some interesting things with you！

• 🔭 I’m currently working on paper for recommender systems implement with Pytorch 😊
• Please give me any advice！

😆 Hey! Welcome to my profile

🍉 Take a look at my stats 🌱

recsystem-pytorch's Issues

I want to ask a question. In ETA model, is long_term id shared embedding with short_term id? Thank you!

class ETA(nn.Module):
....
def forward(self,user,item,long_term,short_term):
"""
:param user: user id
:param item: item id
:param long_term: long-term behavior sequence
:param short_term: short-term behavior sequence
"""
user = torch.flatten(self.user_embedding(user))
item = torch.flatten(self.item_embedding(item))
hashed_item = self.hashing(item)
long_item = []
for i in range(len(long_term)):
long_item.append(torch.flatten(self.item_embedding(long_term[i])))

    short_item = []
    for i in range(len(short_term)):
        short_item.append(torch.flatten(self.item_embedding(short_term[i])))

...

Comments for this project

You can add any comments here.