WinterShiver

I have successfully loaded the pretrained weights from e2e-coref.tgz and get 72.39 on the test set. According to your NAACL18 paper [paper](https://arxiv.org/abs/1804.05392), this corresponds to "e2e + ELMo + hyperparameter...

The implementation of pre-training head, including 3 pre-training task heads, and can load `geolayoutlm_large_pretrain.pt` which contains weights of `ptm_head.predictions` and `ptm_head.triple_geometric_head`. Will this part of code be released?