The graph representation data for the Chinese characters is not provided in this GitHub repo. You can download the "zinode" and "zishp" datasets here: https://tinyurl.com/StrokeOrderCharacterEmbeddings. Also at that Google Drive link, you will find glyce/glyce/my_export_models with the saved models that obtained the reported performances. The suggested download locations are the same as in the Google Drive link.
The task data (Resume NER, CTB5 POS, PKU CWS, BQ Sentence Pair, and Fudan Sentiment Analysis) datasets are not provided here. You can download them through Glyce's GitHub: https://github.com/ShannonAI/glyce/blob/master/docs/dataset_download.md. The suggested download location is adjacent to the outermost glyce folder.
The extra font data that Glyce requires to boost its performance is also not provided here. You can download them through Glyce's GitHub: https://github.com/ShannonAI/glyce/tree/master/glyce/fonts. The suggested download location is inside the innermost glyce folder.