Mirrored_Repos/InvokeAI
1
0
Fork 0
mirror of https://github.com/invoke-ai/InvokeAI synced 2024-08-30 20:32:17 +00:00
Code Issues Packages Projects Releases Wiki Activity

Make quantized loading fast.

Browse Source
This commit is contained in:
Ryan Dick 2024-08-09 16:39:43 +00:00 committed by Brandon
parent 152da482cd
commit 8b1cef978c
1 changed files with 8 additions and 9 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

17
invokeai/backend/requantize.py
View File

@ -1,14 +1,13 @@
from typing import Any, Dict
import torch
from optimum.quanto.nn import QModuleMixin
from optimum.quanto.quantize import _quantize_submodule, freeze
from optimum.quanto.quantize import _quantize_submodule
def custom_freeze(model: torch.nn.Module):
for name, m in model.named_modules():
if isinstance(m, QModuleMixin):
m.freeze()
# def custom_freeze(model: torch.nn.Module):
# for name, m in model.named_modules():
# if isinstance(m, QModuleMixin):
# m.weight =
# m.freeze()
def requantize(
@ -47,8 +46,8 @@ def requantize(
for name, param in m.named_buffers(recurse=False):
setattr(m, name, move_tensor(param, "cpu"))
# Freeze model and move to target device
freeze(model)
model.to(device)
# freeze(model)
# model.to(device)
# Load the quantized model weights
model.load_state_dict(state_dict, strict=False)