Public/automatic
1
0
Fork 0
mirror of https://github.com/vladmandic/automatic synced 2026-09-18 16:54:33 +02:00
Code Issues Packages Projects Releases Wiki Activity

add quantization_config for post-load

Browse Source 
Signed-off-by: Vladimir Mandic <mandic00@live.com>
This commit is contained in:
Vladimir Mandic
2025-10-07 10:14:31 -04:00
parent 7fdc880a73
commit 0092a8b86b
1 changed files with 19 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files
modules/model_quant.py
+19
View File
@@ -530,6 +530,7 @@ def sdnq_quantize_model(model, op=None, sd_model=None, do_gc: bool = True, weigh
backup_embeddings = copy.deepcopy(model.get_input_embeddings())
t0 = time.time()
model = apply_sdnq_to_module(
model,
weights_dtype=weights_dtype,
@@ -548,6 +549,24 @@ def sdnq_quantize_model(model, op=None, sd_model=None, do_gc: bool = True, weigh
modules_dtype_dict=modules_dtype_dict.copy(),
op=op,
)
from modules.sdnq import SDNQConfig
model.quantization_config = SDNQConfig(
weights_dtype=weights_dtype,
group_size=shared.opts.sdnq_quantize_weights_group_size,
svd_rank=shared.opts.sdnq_svd_rank,
use_svd=shared.opts.sdnq_use_svd,
quant_conv=shared.opts.sdnq_quantize_conv_layers,
use_quantized_matmul=shared.opts.sdnq_use_quantized_matmul,
use_quantized_matmul_conv=shared.opts.sdnq_use_quantized_matmul_conv,
dequantize_fp32=shared.opts.sdnq_dequantize_fp32,
non_blocking=shared.opts.diffusers_offload_nonblocking,
quantization_device=quantization_device,
return_device=return_device,
modules_to_not_convert=modules_to_not_convert,
modules_dtype_dict=modules_dtype_dict.copy(),
)
t1 = time.time()
timer.load.add('sdnq', t1 - t0)
model.quantization_method = 'SDNQ'