Need help with Bytedance_ICME_challenge?
Click the “chat” button below for chat support from the developer who created it, or find similar developers for support.

About the developer

challenge-ICME2019-Bytedance
227 Stars 84 Forks 2 Commits 7 Opened issues

Services available

!
?

Need anything else?

Contributors list

REPO DESCRIPTION

Our FM implementation is based on tensorflow 1.12.0.
You can run our reference training code on-the-fly using the following command:

#--------------------------how to train----------------------------#
./train.sh      \
                

CODE STRUCTURE

 
#--------------------------run script------------------------------#
train.sh  

#----------------------------train---------------------------------#
train.py  

#------------------------common operation--------------------------#
common/  
        model_args.py  

#--------convert input text data into tensorflow batch need--------#
data_io/  
       data_parser.py  

#-------------prepare model and build up main framework------------#
models/  
       model.py  

#---------------common algorithm and models for recom--------------#
model_zoo/  
       fm.py  

#-----------------utils for str or data processing-----------------#
utils/  
       utils.py
 

ALGORITHM: FACTORIZATION MACHINE

image

BASELINE

Our baseline results with 5 features (userid, usercity, itemid,authorid,item_city):

  • TRACK2 LIKE TASK:
    auc: 86.5% 
    #------------------------params-------------------------#
    embedding_size = 40
    optimizer = adam  
    lr = 0.0005
    
  • TRACK FINISH TASK:
    auc: 69.8% 
    #------------------------params-------------------------#
    embedding_size = 40
    optimizer = adam   
    lr = 0.0001
    

We use cookies. If you continue to browse the site, you agree to the use of cookies. For more information on our use of cookies please see our Privacy Policy.