experts weren't forwarded into constructer (wasted a few days of training garbage)

This commit is contained in:
mrq 2023-12-23 16:08:17 -06:00
parent 0db3203b21
commit e513d2ef19

View File

@ -18,6 +18,7 @@ def get_model(cfg):
d_model=cfg.dim,
n_heads=cfg.heads,
n_layers=cfg.layers,
n_experts=cfg.experts,
config = cfg,
)