Mirror/stable-diffusion-webui
1
0
Fork 0
mirror of https://github.com/AUTOMATIC1111/stable-diffusion-webui.git synced 2024-06-07 21:20:49 +00:00
Code Issues Packages Projects Releases Wiki Activity

Fix dtype casting for OFT module

Browse Source
This commit is contained in:
Kohaku-Blueleaf 2024-01-05 16:31:48 +08:00
parent a06dab8d7a
commit f8f38c7c28
1 changed files with 3 additions and 3 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

6
extensions-builtin/Lora/network_oft.py
View File

@ -56,7 +56,7 @@ class NetworkModuleOFT(network.NetworkModule):
self.block_size, self.num_blocks = factorization(self.out_dim, self.dim)
def calc_updown(self, orig_weight):
oft_blocks = self.oft_blocks.to(orig_weight.device, dtype=orig_weight.dtype)
oft_blocks = self.oft_blocks.to(orig_weight.device)
eye = torch.eye(self.block_size, device=self.oft_blocks.device)
if self.is_kohya:
@ -66,7 +66,7 @@ class NetworkModuleOFT(network.NetworkModule):
block_Q = block_Q * ((new_norm_Q + 1e-8) / (norm_Q + 1e-8))
oft_blocks = torch.matmul(eye + block_Q, (eye - block_Q).float().inverse())
R = oft_blocks.to(orig_weight.device, dtype=orig_weight.dtype)
R = oft_blocks.to(orig_weight.device)
# This errors out for MultiheadAttention, might need to be handled up-stream
merged_weight = rearrange(orig_weight, '(k n) ... -> k n ...', k=self.num_blocks, n=self.block_size)
@ -77,6 +77,6 @@ class NetworkModuleOFT(network.NetworkModule):
)
merged_weight = rearrange(merged_weight, 'k m ... -> (k m) ...')
updown = merged_weight.to(orig_weight.device, dtype=orig_weight.dtype) - orig_weight
updown = merged_weight.to(orig_weight.device) - orig_weight.to(merged_weight.dtype)
output_shape = orig_weight.shape
return self.finalize_updown(updown, orig_weight, output_shape)