Archives: Tweets
-
@moeller_ml @geomblog I did mean the weight init,…
@moeller_ml @geomblog I did mean the weight init, because clumpy parameter initialization might take longer to break symmetry?
-
@geomblog @moeller_ml … Wonder if hyperuniform (…
@geomblog @moeller_ml … Wonder if hyperuniform (vs pseudorandom) initializations in parameter search might encourage different solutions
-
RT @seattlish: @DJChairmanMoe @Mariners UM ACTUALL…
@DJChairmanMoe @Mariners UM ACTUALLY IT IS A TRILBY.
-
Naw, he—? He said—? ?? ? https://t.co/qp06LGA7WP
Naw, he—? He said—? ?? ? https://t.co/qp06LGA7WP
-
RT @reedfrich: News orgs used to have this ability…
News orgs used to have this ability, they were called ‘foreign correspondents.’ https://t.co/Bj6iCuujOO
-
Yo, @mcclure111 for next revision of SI standards,…
Yo, @mcclure111 for next revision of SI standards, thinking some real @longnow ahead https://t.co/O4p6TVBWuk
-
It’s bad when @cstross’ futures start looking like…
It’s bad when @cstross’ futures start looking like the happy outcomes. https://t.co/B9ij2TFlow
-
@megtra NEVER ENOUGH
@megtra NEVER ENOUGH