kentonl/e2e-coref

how to understand the head_word_emb in the initial file ,glove_50_300_2.txt,how to get the file?

Opened this issue · 3 comments

Vipning commented
how to understand the head_word_emb in the initial file ,glove_50_300_2.txt,how to get the file?
Pab0 commented

As mentioned in the README in section Getting Started:

Download pretrained models at https://drive.google.com/file/d/1fkifqZzdzsOEo0DXMzCFjiNXqsKG_cHi

The link above points to e2e-coref.tgz, which contains the file glove_50_300_2.txt.

glove_300d_2w is just the name given to the .txt file within the experiments.conf configuration:

e2e-coref/experiments.conf

Lines 10 to 11 in 10c6fb6

glove_300d_2w {
path = glove_50_300_2.txt

  • 👍3
Vipning commented

is this file pretrained by using Glove word embeddings with a window size 2?

xiulinyang commented

is this file pretrained by using Glove word embeddings with a window size 2?

Hey, did you figure it out? I'm curious about it too. :)