GPT2_Chinese_Demo

Required packages

requests
lxml
bs4
tqdm

torch
transformers (3.5.1)
wandb

flask

Dataset: 三國演義

python crawler.py

Data will be saved to ./data

We used first chapter as evaluation data

CUDA_VISIBLE_DEVICES=0 ./train.sh

Model will be saved to ./output

Name		Name	Last commit message	Last commit date
Latest commit History 6 Commits
templates		templates
.gitignore		.gitignore
README.md		README.md
crawler.py		crawler.py
demo.py		demo.py
run_language_modeling.py		run_language_modeling.py
train.sh		train.sh