Public/automatic
1
0
Fork 0
mirror of https://github.com/vladmandic/automatic synced 2026-09-10 23:08:43 +02:00
Code Issues Packages Projects Releases Wiki Activity

refactor(lora): remove the lora_apply_te setting

Browse Source 
Text encoder networks now apply unconditionally in the native path.

- remove the option, the per-request parameter and the control threading
- collapse activate_filtered into plain activate at all call sites
- drop the toggle from the prompt embed cache key
- register the retired key so existing configs load without warnings
This commit is contained in:
CalamitousFelicitousness
2026-07-18 04:16:29 +01:00
parent c42adcf680
commit 55c3eb1325
10 changed files with 8 additions and 24 deletions
Download Patch File Download Diff File Expand all files Collapse all files
modules/processing_diffusers.py
+2 -2
View File
@@ -149,7 +149,7 @@ def process_base(p: processing.StableDiffusionProcessing):
if 'detailer' in p.ops:
desc = 'Detail'
p.prompts, p.network_data = extra_networks.parse_prompts(p.prompts, p.network_data)
extra_networks.activate_filtered(p) # networks must patch weights before prompt encode so te loras affect embeds
extra_networks.activate(p) # networks must patch weights before prompt encode so te loras affect embeds
base_args = set_pipeline_args(
p=p,
model=shared.sd_model,
@@ -313,7 +313,7 @@ def process_hires(p: processing.StableDiffusionProcessing, output):
prompts, p.network_data = extra_networks.parse_prompts(prompts)
reset_prompts = True
if reset_prompts or ('base' in p.skip):
extra_networks.activate_filtered(p)
extra_networks.activate(p)
hires_args = set_pipeline_args(
p=p,