search

enricivi / adversarial_training_methods

Implementation of the methods proposed in **Adversarial Training Methods for Semi-Supervised Text Classification** on IMDB dataset (without pre-training)
MIT License
40 stars 11 forks source link

About vectors.txt #8

Closed Snowfy1031 closed 4 years ago

Snowfy1031 commented 4 years ago

Hello,nice to meet u, I tried to run your program, but I can't find ''/GloVe-1.2/vectors.txt'' in "pretrained_glove_embedding_script.ipynb" , I don't know what it is. Could u please help me? Thank you very much for your help.

enricivi commented 4 years ago

Hello Snowfy1031,

GloVe (https://nlp.stanford.edu/projects/glove/) is an unsupervised algorithm used to find numerical representation (real vector) of words.

If I am not wrong, in the original paper the authors have defined this embedding training the network on a different (but related) task. In this project we have decided to use as word embedding the ones given by GloVe in order to reduce the overall time needed to train the model (removing the pretraining).

We have not worked on this project since a couple of years thus I do not exactly remember every phase. Anyway, I think you have to use GloVe on your train+val set in order to compute the embedding for each word and then put this vectors into the embedding layer of our network (pay attention to the dimension of vector/embedding layer).

Kind regards,

On Sat, 4 Apr 2020 at 03:18, Snowfy1031 notifications@github.com wrote:

Hello,nice to meet u, I tried to run your program, but I can't find ''/GloVe-1.2/vectors.txt'' in "pretrained_glove_embedding_script.ipynb" , I don't know what it is. Could u please help me? Thank you very much for your help.

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/enry12/adversarial_training_methods/issues/8, or unsubscribe https://github.com/notifications/unsubscribe-auth/AFOZ3VIFVQG4K7UJVWQJR2DRK2DGLANCNFSM4L4W7TWA .

--

Enrico Civitelli

Snowfy1031 commented 4 years ago

Thank you so much for your help. Although I am a beginner, I will try to use Glove and solve it. If there is any other problem, I may ask you for help again.

I am truly grateful for your kindness.