Public/automatic
1
0
Fork 0
mirror of https://github.com/vladmandic/automatic synced 2026-09-20 01:31:13 +02:00
Code Issues Packages Projects Releases Wiki Activity

SDNQ Atten: add quantize_fp32

Browse Source
This commit is contained in:
Dity0
2026-08-24 19:34:54 +03:00
parent 3c2236c4ed
commit f6777b168d
3 changed files with 3 additions and 1 deletions
Download Patch File Download Diff File Expand all files Collapse all files
extensions-builtin/sdnq
+1 -1
Submodule extensions-builtin/sdnq updated: 3986a375fc...b99d9af158
modules/attention.py
+1
View File
@@ -38,6 +38,7 @@ def set_sdnq_attention():
smooth_k=shared.opts.sdnq_attention_smooth_k,
use_hadamard=shared.opts.sdnq_attention_use_hadamard,
hadamard_group_size=shared.opts.sdnq_attention_hadamard_group_size,
quantize_fp32=shared.opts.sdnq_attention_quantize_fp32,
use_fp16_accum=shared.opts.sdnq_attention_use_fp16_accum,
)
else:
modules/ui_definitions.py
+1
View File
@@ -258,6 +258,7 @@ def create_settings(cmd_opts):
"sdnq_attention_smooth_k": OptionInfo(True, "SDNQ Attention use Smooth K", gr.Checkbox),
"sdnq_attention_use_hadamard": OptionInfo(False, "SDNQ Attention use Hadamard", gr.Checkbox),
"sdnq_attention_use_fp16_accum": OptionInfo(False, "SDNQ Attention use FP16 Accumulation", gr.Checkbox),
"sdnq_attention_quantize_fp32": OptionInfo(True, "SDNQ Attention Quantize FP32", gr.Checkbox),
"sdnq_attention_matmul_type": OptionInfo("enabled", "SDNQ Attention MatMul type", gr.Radio, {"choices": sdnq_matmul_modes}),
"sdnq_attention_pv_matmul_type": OptionInfo("disabled", "SDNQ Attention PV MatMul type", gr.Radio, {"choices": sdnq_matmul_modes}),
"sdnq_attention_hadamard_group_size": OptionInfo(256, "SDNQ Attention Hadamard Group Size", gr.Slider, {"minimum": 4, "maximum": 1024, "step": 1}),