[Bug]: Loras embedding stoped working

Checklist

[X] The issue has not been resolved by following the troubleshooting guide
[ ] The issue exists on a clean installation of Fooocus
[X] The issue exists in the current version of Fooocus
[x] The issue has not been reported before recently
[ ] The issue has been reported before but has not been fixed yet

What happened?

Hello, i have seen recently that Loras embedding doesn't work like it did before. It doesn't work in direct prompt and neither on wildcards. This was the most useful feature of Fooocus for me, please take a look to the problem, it stopped working since version 2.5.5 i think. I'm sure i'm using the code correctly, it's <lora:LORA_NAME:STRENGTH> , right ?

Steps to reproduce the problem

1 - Generate any image using <lora:LORA_NAME:STRENGTH> (example: <lora:anime_character:1>)on prompt or call wildcard with the lora embedded, the "LORA_NAME.safetensors" is in models/loras/ folder 2 - The lora doesn't work 3 - Call the same lora in Fooocus dropdown list menu instead of embedding it 4 - The lora work

What should have happened?

The lora should work as it did in former versions

What browsers do you use to access Fooocus?

Mozilla Firefox

Where are you running Fooocus?

Locally

What operating system are you using?

Windows 10

Console logs

[Parameters] Adaptive CFG = 7
[Parameters] CLIP Skip = 2
[Parameters] Sharpness = 0
[Parameters] ControlNet Softness = 0.25
[Parameters] ADM Scale = 1.5 : 0.8 : 0.3
[Parameters] Seed = 1221783870755033148
[Parameters] CFG = 7.5
[Fooocus] Loading control models ...
[Parameters] Sampler = dpmpp_2m_sde_gpu - karras
[Parameters] Steps = 30 - 24
[Fooocus] Initializing ...
[Fooocus] Loading models ...
Refiner unloaded.
Request to load LoRAs [] for model [E:\Imágenes\Fooocus_win64_X-X-XXX\Fooocus\models\checkpoints\BASE\Anime\BASE-Hentai 05-autismmixSDXL_autismmixPony.safetensors].
Requested to load SDXLClipModel
Loading 1 new model
unload clone 1
[Fooocus] Processing prompts ...
[Wildcards] processing: __AutismGoodAnimeStyle__, __AutismErotic__
[Wildcards] (score_9, score_8_up, score_7_up), ultra detailed, realistic art, soft shading, __AutismErotic__
[Wildcards] (score_9, score_8_up, score_7_up), ultra detailed, realistic art, soft shading, rating_questionable, source_anime, __pose_erotic__, mid shot, __Angulos__, looking at viewer, (solo), pretty, perfect eyes, perfect hands, erotic, seductive, blush face
[Wildcards] processing: (score_9, score_8_up, score_7_up), ultra detailed, realistic art, soft shading, rating_questionable, source_anime, __pose_erotic__, mid shot, __Angulos__, looking at viewer, (solo), pretty, perfect eyes, perfect hands, erotic, seductive, blush face
[Wildcards] (score_9, score_8_up, score_7_up), ultra detailed, realistic art, soft shading, rating_questionable, source_anime, Legs outstretched arms stretched out, seductive pose, mid shot, __Angulos__, looking at viewer, (solo), pretty, perfect eyes, perfect hands, erotic, seductive, blush face
[Wildcards] (score_9, score_8_up, score_7_up), ultra detailed, realistic art, soft shading, rating_questionable, source_anime, Legs outstretched arms stretched out, seductive pose, mid shot, selfie, looking at viewer, (solo), pretty, perfect eyes, perfect hands, erotic, seductive, blush face
[Wildcards] processing: __01EroticStyleNegative__
[Wildcards] six fingers, extra fingers, bad hands, bad body, bad eyes, (nude breasts, nude vagina), translucent clothing above nipples, (thick lines, thick inking, very thick lines, very thick inking), low res, 3d, monochrome, sketch, painting, realistic, worst quality, (low quality:1.4), blurred, blur
[Wildcards] processing: __NicoleMainClothing__, covered nipples, (__background_crowded_place__), voluptuous body
[Wildcards] black ribbon, black shorts, tube top, black legwear, covered nipples, (__background_crowded_place__), voluptuous body
[Wildcards] black ribbon, black shorts, tube top, black legwear, covered nipples, (Film Festival Crowd: Movie buffs eager to watch red carpet glamour), voluptuous body
[Wildcards] processing: __Nicole__
[Wildcards] <lora:PONYXLNicole:0.8>, NicoleNSFW, green eyes, pink hair, long hair, two side up, large breasts, mole on breast, mole under eye, navel, hairclip
[Fooocus] Encoding positive #1 ...
[Fooocus] Encoding negative #1 ...
[Parameters] Denoising Strength = 1.0
[Parameters] Initial Latent shape: Image Space (1344, 768)
Preparation time: 8.26 seconds
Using karras scheduler.
[Fooocus] Preparing task 1/1 ...
[Sampler] refiner_swap_method = joint
[Sampler] sigma_min = 0.0291671771556139, sigma_max = 14.614643096923828
Requested to load SDXL
Loading 1 new model
loading in lowvram mode 1738.1789360046387
[Fooocus Model Management] Moving model(s) has taken 4.92 seconds
 33%|████████████████████████████████████████████                                                                                        | 10/30 [01:55<03:50, 11.54s/it]
User stopped
[Enhance] Skipping, preconditions aren't met
Processing time (total): 120.35 seconds
Total time: 128.69 seconds

Additional information

EDIT: It looks like the problem exists since more previous versions, maybe since v2.4.2 because i have been taking a look into my older image generations and seems that is since the Month of June when i started to see that Lora characters doesn't look 100% similar to the originals i wanted to see. Also the checkpoint i use, "AutismXL", has inside its training data a lot of fan-art characters that confused me all this time since today when i was trying to generate a new character that is not part of that checkpoint training and the alert started when i saw this last character doesn't look like the original even in a 30% of it's appearance. Also i must to add that some Loras are reinforced with descriptions of the characters and these descriptions if are enough accurate sometimes generate a similar clone of it even when the Lora embedding didn't work and this fact confused me even more until today. (sorry for my bad English, i hope to have explained enough well the problem)

lllyasviel / Fooocus