Open wangliangguo opened 8 years ago
The pretrained model consists of the zip file downloads referenced on the readme. These files are simply dictionaries mapping each word to a word vector. Does that make sense?
@Russell91 thanks for your reply. but i want to load these pretrained vector to the model and then sent some context of a word to this model to get a predict. How can i do that? (load the pretrained vector to the model)
Next word prediction is a separate task, that you may indeed use GloVe vectors for, but which actually requires additional machinery. I would suggest you look up other guides on the internet for how to do this. You'll find strategies that do use GloVe vectors as features for a machine learning model.
I have some unknow words(unk) in my data, i want to send the context of unk to the pretrained model to get the prediction of embedding for the unk words. But i can't found pretrained model here, is there any such resources?