allow apex norm to work
This commit is contained in:
parent
7391389276
commit
cdf9fdbc2d
@ -169,10 +169,16 @@ class AsymmetricAttention(nn.Module):
|
||||
pass
|
||||
else:
|
||||
from .layers import RMSNorm
|
||||
self.q_norm_x = RMSNorm(self.head_dim, device=device)
|
||||
self.k_norm_x = RMSNorm(self.head_dim, device=device)
|
||||
self.q_norm_y = RMSNorm(self.head_dim, device=device)
|
||||
self.k_norm_y = RMSNorm(self.head_dim, device=device)
|
||||
if rms_norm_func == "apex":
|
||||
self.q_norm_x = RMSNorm(self.head_dim)
|
||||
self.k_norm_x = RMSNorm(self.head_dim)
|
||||
self.q_norm_y = RMSNorm(self.head_dim)
|
||||
self.k_norm_y = RMSNorm(self.head_dim)
|
||||
else:
|
||||
self.q_norm_x = RMSNorm(self.head_dim, device=device)
|
||||
self.k_norm_x = RMSNorm(self.head_dim, device=device)
|
||||
self.q_norm_y = RMSNorm(self.head_dim, device=device)
|
||||
self.k_norm_y = RMSNorm(self.head_dim, device=device)
|
||||
|
||||
# Output layers. y features go back down from dim_x -> dim_y.
|
||||
self.proj_x = nn.Linear(dim_x, dim_x, bias=out_bias, device=device)
|
||||
|
||||
Loading…
x
Reference in New Issue
Block a user