wip incorrect OFT implementation

This commit is contained in:
v0xie 2023-10-17 23:35:50 -07:00
parent 861cbd5636
commit ec718f76b5
2 changed files with 87 additions and 0 deletions

View File

@ -0,0 +1,82 @@
import torch
import network
class ModuleTypeOFT(network.ModuleType):
def create_module(self, net: network.Network, weights: network.NetworkWeights):
if all(x in weights.w for x in ["oft_blocks"]):
return NetworkModuleOFT(net, weights)
return None
# adapted from https://github.com/kohya-ss/sd-scripts/blob/main/networks/oft.py
class NetworkModuleOFT(network.NetworkModule):
def __init__(self, net: network.Network, weights: network.NetworkWeights):
super().__init__(net, weights)
self.oft_blocks = weights.w["oft_blocks"]
self.alpha = weights.w["alpha"]
self.dim = self.oft_blocks.shape[0]
self.num_blocks = self.dim
#if type(self.alpha) == torch.Tensor:
# self.alpha = self.alpha.detach().numpy()
if "Linear" in self.sd_module.__class__.__name__:
self.out_dim = self.sd_module.out_features
elif "Conv" in self.sd_module.__class__.__name__:
self.out_dim = self.sd_module.out_channels
self.constraint = self.alpha * self.out_dim
self.block_size = self.out_dim // self.num_blocks
self.oft_multiplier = self.multiplier()
# replace forward method of original linear rather than replacing the module
# self.org_forward = self.sd_module.forward
# self.sd_module.forward = self.forward
def get_weight(self):
block_Q = self.oft_blocks - self.oft_blocks.transpose(1, 2)
norm_Q = torch.norm(block_Q.flatten())
new_norm_Q = torch.clamp(norm_Q, max=self.constraint)
block_Q = block_Q * ((new_norm_Q + 1e-8) / (norm_Q + 1e-8))
I = torch.eye(self.block_size, device=self.oft_blocks.device).unsqueeze(0).repeat(self.num_blocks, 1, 1)
block_R = torch.matmul(I + block_Q, (I - block_Q).inverse())
block_R_weighted = self.oft_multiplier * block_R + (1 - self.oft_multiplier) * I
R = torch.block_diag(*block_R_weighted)
return R
def calc_updown(self, orig_weight):
oft_blocks = self.oft_blocks.to(orig_weight.device, dtype=orig_weight.dtype)
block_Q = oft_blocks - oft_blocks.transpose(1, 2)
norm_Q = torch.norm(block_Q.flatten())
new_norm_Q = torch.clamp(norm_Q, max=self.constraint)
block_Q = block_Q * ((new_norm_Q + 1e-8) / (norm_Q + 1e-8))
I = torch.eye(self.block_size, device=oft_blocks.device).unsqueeze(0).repeat(self.num_blocks, 1, 1)
block_R = torch.matmul(I + block_Q, (I - block_Q).inverse())
block_R_weighted = self.oft_multiplier * block_R + (1 - self.oft_multiplier) * I
R = torch.block_diag(*block_R_weighted)
#R = self.get_weight().to(orig_weight.device, dtype=orig_weight.dtype)
# W = R*W_0
updown = orig_weight + R
output_shape = [R.size(0), orig_weight.size(1)]
return self.finalize_updown(updown, orig_weight, output_shape)
# def forward(self, x, y=None):
# x = self.org_forward(x)
# if self.oft_multiplier == 0.0:
# return x
# R = self.get_weight().to(x.device, dtype=x.dtype)
# if x.dim() == 4:
# x = x.permute(0, 2, 3, 1)
# x = torch.matmul(x, R)
# x = x.permute(0, 3, 1, 2)
# else:
# x = torch.matmul(x, R)
# return x

View File

@ -11,6 +11,7 @@ import network_ia3
import network_lokr
import network_full
import network_norm
import network_oft
import torch
from typing import Union
@ -28,6 +29,7 @@ module_types = [
network_full.ModuleTypeFull(),
network_norm.ModuleTypeNorm(),
network_glora.ModuleTypeGLora(),
network_oft.ModuleTypeOFT(),
]
@ -183,6 +185,9 @@ def load_network(name, network_on_disk):
elif sd_module is None and "lora_te1_text_model" in key_network_without_network_parts:
key = key_network_without_network_parts.replace("lora_te1_text_model", "0_transformer_text_model")
sd_module = shared.sd_model.network_layer_mapping.get(key, None)
elif sd_module is None and "oft_unet" in key_network_without_network_parts:
key = key_network_without_network_parts.replace("oft_unet", "diffusion_model")
sd_module = shared.sd_model.network_layer_mapping.get(key, None)
# some SD1 Loras also have correct compvis keys
if sd_module is None: