Mirrors
/
stable-diffusion-webui
mirror of https://github.com/AUTOMATIC1111/stable-diffusion-webui.git
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity

test implementation based on kohaku diag-oft implementation

This commit is contained in:
v0xie 2023-11-01 22:34:27 -07:00
parent 6523edb8a4
commit a2fad6ee05
1 changed files with 37 additions and 20 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

57
extensions-builtin/Lora/network_oft.py
View File

@ -1,5 +1,6 @@
import torch import torch
import network import network
from einops import rearrange
class ModuleTypeOFT(network.ModuleType): class ModuleTypeOFT(network.ModuleType):
@ -30,35 +31,51 @@ class NetworkModuleOFT(network.NetworkModule):
self.org_module: list[torch.Module] = [self.sd_module] self.org_module: list[torch.Module] = [self.sd_module]
def merge_weight(self, R_weight, org_weight): # def merge_weight(self, R_weight, org_weight):
R_weight = R_weight.to(org_weight.device, dtype=org_weight.dtype) # R_weight = R_weight.to(org_weight.device, dtype=org_weight.dtype)
if org_weight.dim() == 4: # if org_weight.dim() == 4:
weight = torch.einsum("oihw, op -> pihw", org_weight, R_weight) # weight = torch.einsum("oihw, op -> pihw", org_weight, R_weight)
else: # else:
weight = torch.einsum("oi, op -> pi", org_weight, R_weight) # weight = torch.einsum("oi, op -> pi", org_weight, R_weight)
return weight # weight = torch.einsum(
# "k n m, k n ... -> k m ...",
# self.oft_diag * scale + torch.eye(self.block_size, device=device),
# org_weight
# )
# return weight
def get_weight(self, oft_blocks, multiplier=None): def get_weight(self, oft_blocks, multiplier=None):
constraint = self.constraint.to(oft_blocks.device, dtype=oft_blocks.dtype) # constraint = self.constraint.to(oft_blocks.device, dtype=oft_blocks.dtype)
block_Q = oft_blocks - oft_blocks.transpose(1, 2) # block_Q = oft_blocks - oft_blocks.transpose(1, 2)
norm_Q = torch.norm(block_Q.flatten()) # norm_Q = torch.norm(block_Q.flatten())
new_norm_Q = torch.clamp(norm_Q, max=constraint) # new_norm_Q = torch.clamp(norm_Q, max=constraint)
block_Q = block_Q * ((new_norm_Q + 1e-8) / (norm_Q + 1e-8)) # block_Q = block_Q * ((new_norm_Q + 1e-8) / (norm_Q + 1e-8))
m_I = torch.eye(self.block_size, device=oft_blocks.device).unsqueeze(0).repeat(self.num_blocks, 1, 1) # m_I = torch.eye(self.block_size, device=oft_blocks.device).unsqueeze(0).repeat(self.num_blocks, 1, 1)
block_R = torch.matmul(m_I + block_Q, (m_I - block_Q).inverse()) # block_R = torch.matmul(m_I + block_Q, (m_I - block_Q).inverse())
block_R_weighted = multiplier * block_R + (1 - multiplier) * m_I # block_R_weighted = multiplier * block_R + (1 - multiplier) * m_I
R = torch.block_diag(*block_R_weighted) # R = torch.block_diag(*block_R_weighted)
#return R
return self.oft_blocks
return R
def calc_updown(self, orig_weight): def calc_updown(self, orig_weight):
multiplier = self.multiplier() * self.calc_scale() multiplier = self.multiplier() * self.calc_scale()
R = self.get_weight(self.oft_blocks, multiplier) #R = self.get_weight(self.oft_blocks, multiplier)
merged_weight = self.merge_weight(R, orig_weight) R = self.oft_blocks.to(orig_weight.device, dtype=orig_weight.dtype)
#merged_weight = self.merge_weight(R, orig_weight)
updown = merged_weight.to(orig_weight.device, dtype=orig_weight.dtype) - orig_weight orig_weight = rearrange(orig_weight, '(k n) ... -> k n ...', k=self.num_blocks, n=self.block_size)
weight = torch.einsum(
'k n m, k n ... -> k m ...',
R * multiplier + torch.eye(self.block_size, device=orig_weight.device),
orig_weight
)
weight = rearrange(weight, 'k m ... -> (k m) ...')
#updown = merged_weight.to(orig_weight.device, dtype=orig_weight.dtype) - orig_weight
updown = weight.to(orig_weight.device, dtype=orig_weight.dtype) - orig_weight
output_shape = orig_weight.shape output_shape = orig_weight.shape
orig_weight = orig_weight orig_weight = orig_weight