Skip to content

Latest commit

 

History

History
45 lines (35 loc) · 1.08 KB

File metadata and controls

45 lines (35 loc) · 1.08 KB

Starter Code for Deep Learning Coding Project 4

Acknowledgement

We thank Shusheng Xu and Yunfei Li for their initial contribution and Wei Fu with document formatting and package version testing.

Dataset

Download the datasets from https://cloud.tsinghua.edu.cn/d/1b60ac7277994f60b2e2/ and organize the folders as follows:

dl_course_hw4
├── Datasets
    └── CCPC
    ├── CLS
    └── couplet
├── generation
├── classification
└── README.md

Generation

cd generation
# Seq2Seq generation
python train.py  --seq2seq --model-type transformer [optional arguments]
python train.py  --seq2seq --model-type lstm [optional arguments]
# Unconditional generaton
python train.py --model-type transformer [optional arguments]
python train.py --model-type lstm [optional arguments]
# evaluate
python evaluation.py

Classification

cd classification
# training
python train.py [optional arguments]
# evaluate
python evaluation.py