search

zhanwenchen / relaug

MIT License
0 stars 0 forks source link

Try deploying predcls baseline with lr=0.16, num_gpus to 2, and per-g… #116

Closed zhanwenchen closed 1 year ago

zhanwenchen commented 1 year ago

…pu batch size to 8.