search

alex-tifrea / poincare_glove

Implementation of the "Poincare Glove: Hyperbolic word embeddings" paper
GNU Lesser General Public License v2.1
83 stars 20 forks source link

Time taken to train word embeddings #15

Open ayushbits opened 3 years ago

ayushbits commented 3 years ago

Hi @alex-tifrea ,

I'm training word embeddings on the RCV1 corpus. I've generated the vocab (7MB) and co-occurence file (7.5GB) from the glove's code for RCV1 dataset. I run the following command to train ./run_glove.sh --train --root .. -coocc_file ../poincare_glove2/GloVe/cooccurrence.bin --vocab_file ../poincare_glove2/GloVe/vocab.txt --epochs 50 --workers 20 --restrict_vocab 200000 --lr 0.01 --poincare 1 --bias --size 100 --dist_func cosh-dist-sq

The scripts uses only single CPU core and do not dump anything in the logs/* file.

Could anyone provide any pointers to what I am doing wrong or something else is wrong?

Thanks

Clarioooo commented 2 weeks ago

Hi @ayushbits ,

I encountered the same problem as you, did you solve it?