github/ai-toolkit
1
0
Fork 0
Code Issues Pull Requests Actions 1 Packages Projects Releases Wiki Activity

remove unicode (#955)

Browse Source
This commit is contained in:
PlagueKind
2026-07-17 00:20:16 +02:00
committed by GitHub
parent 0552d85aa7
commit 3c75735ba2
1 changed files with 3 additions and 3 deletions
Download Patch File Download Diff File Expand all files Collapse all files

6
toolkit/models/lokr.py
View File

@@ -26,7 +26,7 @@ def factorization(dimension: int, factor: int = -1) -> tuple[int, int]:
In LoRA with Kroneckor Product, first value is a value for weight scale.
secon value is a value for weight.
Becuase of non-commutative property, A⊗B ≠ B⊗A. Meaning of two matrices is slightly different.
Becuase of non-commutative property, A(kron)B != B(kron)A. Meaning of two matrices is slightly different.
examples)
factor
@@ -151,7 +151,7 @@ class LokrModule(ToolkitModuleMixin, nn.Module):
torch.empty(shape[0][1], lora_dim))
self.lokr_w2_b = nn.Parameter(torch.empty(
lora_dim, shape[1][1]*shape[2]*shape[3]))
# w1 ⊗ (w2_a x w2_b) = (a, b)⊗((c, dim)x(dim, d*k1*k2)) = (a, b)⊗(c, d*k1*k2) = (ac, bd*k1*k2)
# w1 (kron) (w2_a x w2_b) = (a, b)(kron)((c, dim)x(dim, d*k1*k2)) = (a, b)(kron)(c, d*k1*k2) = (ac, bd*k1*k2)
self.op = F.conv2d
self.extra_args = {
@@ -191,7 +191,7 @@ class LokrModule(ToolkitModuleMixin, nn.Module):
torch.empty(shape[0][1], lora_dim))
self.lokr_w2_b = nn.Parameter(
torch.empty(lora_dim, shape[1][1]))
# w1 ⊗ (w2_a x w2_b) = (a, b)⊗((c, dim)x(dim, d)) = (a, b)⊗(c, d) = (ac, bd)
# w1 (kron) (w2_a x w2_b) = (a, b)(kron)((c, dim)x(dim, d)) = (a, b)(kron)(c, d) = (ac, bd)
else:
self.use_w2 = True
self.lokr_w2 = nn.Parameter(