search

sacmehta / delight

DeLighT: Very Deep and Light-Weight Transformers
MIT License
467 stars 53 forks source link

[PTB] Reproducibility #15

Open DavidHerel opened 1 year ago

DavidHerel commented 1 year ago

Hi,

could you please provide script and hyper-parameters on how to train model on PTB to obtain same results as in your paper?

Thank you