yilifzf / racp

Recurrent Attention over Contextualized Page sequence for Click-Through Rate Prediction in Taobao Personalized Search Ranking

Introduction

Pipeline:

1. Prepare Data
2. Train Model

Running

We test our code on Python 3.7.5 and PyTorch 1.7.1.

1. Prepare Data

mkdir ./Data/raw/
mkdir ./Data/new/
mkdir ./Data/data/
cd ./Data/raw/
kaggle competitions download -c avito-context-ad-clicks
py7zr -x *.7z
cd ../../
python ./DataPreprocess/first_step.py
python ./DataPreprocess/second_step.py

When you see the files below, you can do the next work.

./Data/data/Train_data.csv
./Data/data/Val_data.csv
./Data/data/Test_data.csv
./Data/data/recent_history.pickle

2. Train Model

Firstly, set the sys_path in ./Runs/project_path.ini as the absoluate path of current project.

Seconly, run the command python ./Runs/run.py --model_name racp.