navjordj
Re:Posting a 5x5 crossword every day
on Wednesday September 04, 2024 @07:36AM (#41444457)
Could you share some details on what algorithm you use to generate the puzzles?
Re:33TB of text data for a 1T-parameter model
on Sunday April 16, 2023 @06:53AM (#35588901)
The original T5 models were trained with masked language modelling similar to BERT and afterwards also trained on a autoregressive task similar to GPT.
Re:33TB of text data for a 1T-parameter model
on Sunday April 16, 2023 @06:51AM (#35588893)
I would be very suprised if this hasn't already been implemented since GPT-3/3.5