e-caste / MLAI-Project-SIMMC-Subtask1-Optimization

A PyTorch deep learning model for multi-class multi-label user intention prediction. Slides and paper at link below

Home Page:https://cloud.caste.dev/s/RfSDiJBW6anEwLk

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

MLAI-Project-SIMMC-Subtask1-Optimization

This code has been tested with Python 3.7 and 3.8.
All info can be found in main.ipynb, and the code can be directly run from the Jupyter notebook both locally and on Google Colab.

The dataset is in the data/simmc_fashion directory, and has been already preprocessed with the following settings:

  • tokenizer: BertTokenizer
  • history: 0, 6, 12
  • NUM_ATTRIBUTES = 14 + 1

Base model architecture:
network

About

A PyTorch deep learning model for multi-class multi-label user intention prediction. Slides and paper at link below

https://cloud.caste.dev/s/RfSDiJBW6anEwLk


Languages

Language:Jupyter Notebook 78.6%Language:Python 19.3%Language:Shell 2.1%