Fix issue with picking layers for quantization, adjust layers fo better quantization of cogview4

This commit is contained in:
Jaret Burkett
2025-03-05 13:44:40 -07:00
parent aa44828c0c
commit 4fe33f51c1
3 changed files with 78 additions and 4 deletions

View File

@@ -64,7 +64,8 @@ from toolkit.paths import ORIG_CONFIGS_ROOT, DIFFUSERS_CONFIGS_ROOT
from huggingface_hub import hf_hub_download
from toolkit.models.flux import add_model_gpu_splitter_to_flux, bypass_flux_guidance, restore_flux_guidance
from optimum.quanto import freeze, qfloat8, quantize, QTensor, qint4
from optimum.quanto import freeze, qfloat8, QTensor, qint4
from toolkit.util.quantize import quantize
from toolkit.accelerator import get_accelerator, unwrap_model
from typing import TYPE_CHECKING
from toolkit.print import print_acc