2.3.0

[mashb1t]

TL;DR

• Add performance "lightning" (based on SDXL-Lightning 4 step LoRA)
• Add preset selection to UI, disable with argument --disable-preset-selection. Use --always-download-new-model to download missing models on preset switch.
• Add temp path cleanup on startup, configurable in config.txt
• Add support for wildcard subdirectories
• Add scrollable 2 column layout for styles for better structure
• Improve Colab resource needs for T4 instances (default), positively tested with all image prompt features
• Improve anime preset, now uses style Fooocus Semi Realistic instead of Fooocus Negative (less wet look images)

What's Changed

• fix: parse width and height as int when applying metadata by @mashb1t in fix: parse width and height as int when applying metadata #2452
• fix: do not attempt to remove non-existing image grid file by @mashb1t in fix: do not attempt to remove non-existing image grid file #2456
• feat: add troubleshooting guide to bug report template again by @mashb1t in feat: add troubleshooting guide to bug report template again #2489
• feat: use jpeg instead of jpg, use enums instead of strings by @mashb1t in feat: use jpeg instead of jpg, use enums instead of strings #2453
• feat: add performance lightning with 4 step LoRA by @mashb1t in feat: add performance sdxl lightning #2415
• fix: change synthetic refiner switch from 0.5 to 0.8 by @xhoxye in fix refiner switching issues #2165
• feat: add config for temp path and temp path cleanup on launch by @midareashi and @mashb1t in Feat: Add ability to change the Gradio temp path for saving images. Added ability to clear Gradio temp files and folders on launch #1992
• feat: scan wildcard subdirectories by @Cruxial0 and @mashb1t in feat: Scan wildcard subdirectories #2466
• fix: use correct method call for interrupt_current_processing by @mashb1t in fix: use correct method call for interrupt_current_processing #2506
• feat: read wildcards in order 通配符增强，切换顺序读取。 by @xhoxye and @mashb1t in Wildcard enhancement, toggle read in order, prompt queue. 通配符增强，切换顺序读取。 #1761
• feat: use scrollable 2 column layout for styles by @hswlab in Styles Grouping/Sorting #1770 #1883
• feat: allow to add disabled LoRAs in config by @mashb1t in feat: allow to add disabled LoRAs in config #2507
• fix: prioritize VRAM over RAM in Colab, preventing out of memory issues by @Peppe289 and @mashb1t in Colab: Balances the data distribution #1710
• fix: update xformers to 0.0.23 by @mashb1t in fix: update incompatible xformers version #2517
• feat: update xformers to 0.0.23 in Dockerfile by @josephrocca in Update xformers to 0.0.23 in Dockerfile #2519
• fix: parse seed as string to display correctly in metadata preview by @mashb1t in fix: parse seed as string to display correctly in metadata preview #2536
• feat: allow user add custom preset without block git pull by @Zxilly in feat: allow user add custom preset without block git pull #2520
• feat: add preset selection to Gradio UI (session based) by @mashb1t in Add preset selection to Gradio UI (session based) #1570
• fix: add error output for unsupported images by @shaye059 in fix: hang on unsupported image #2537
• feat: improve anime preset by adding style Fooocus Semi Realistic by @DavidDragonsage in Improved Negative Prompt Support for Anime, Paintings and Other Non-Photographic Media #2481 #2492
• feat: add backwards compatibility for presets without disable/enable LoRA boolean by @mashb1t

New Contributors

• @midareashi made their first contribution in Feat: Add ability to change the Gradio temp path for saving images. Added ability to clear Gradio temp files and folders on launch #1992
• @Cruxial0 made their first contribution in feat: Scan wildcard subdirectories #2466
• @hswlab made their first contribution in Styles Grouping/Sorting #1770 #1883
• @Peppe289 made their first contribution in Colab: Balances the data distribution #1710
• @josephrocca made their first contribution in Update xformers to 0.0.23 in Dockerfile #2519
• @Zxilly made their first contribution in feat: allow user add custom preset without block git pull #2520
• @shaye059 made their first contribution in fix: hang on unsupported image #2537
• @DavidDragonsage made their first contribution in Improved Negative Prompt Support for Anime, Paintings and Other Non-Photographic Media #2481 #2492

Full Changelog: 2.2.1...2.3.0

---

This discussion was created from the release 2.3.0.

[CadLoop]

As always infinite gratitude for this great work! :) Whenever I go, I start conversation "do you know this epic github repo?"

[Tra2x]

With reach release update I have some Issues after the automatic update. Any Ideas?
Most probably it has to do with my automatic replace of the config.txt (I use different ones for normal image creation when my kids are also with me and the 'private' images ;) )

Fooocus startup
traxx@traxx-ubuntu:~$ cat ./x_fooocus.sh
#!/bin/bash
cd Fooocus
rm config.txt
cp config_x.txt config.txt
source fooocus_env/bin/activate
export HSA_OVERRIDE_GFX_VERSION=10.3.0
export HCC_AMDGPU_TARGET=gfx1032
python entry_with_update.py --always-low-vram

Fooocus Log
Already up-to-date
Update succeeded.
[System ARGV] ['entry_with_update.py', '--always-low-vram']
Python 3.10.12 (main, Nov 20 2023, 15:14:05) [GCC 11.4.0]
Fooocus version: 2.3.0
Failed to load config key: {"default_aspect_ratio": "1152896"} is invalid; will use {"default_aspect_ratio": "10241024"} instead.
[Cleanup] Attempting to delete content of temp dir /tmp/fooocus
[Cleanup] Cleanup successful
Total VRAM 8176 MB, total RAM 48091 MB
Set vram state to: LOW_VRAM
Always offload VRAM
Device: cuda:0 AMD Radeon RX 6650 XT : native
VAE dtype: torch.float32
Using sub quadratic optimization for cross attention, if you have memory or speed issues try using: --attention-split
Exception in thread Thread-2 (worker):
Traceback (most recent call last):
File "/usr/lib/python3.10/threading.py", line 1016, in _bootstrap_inner
self.run()
File "/usr/lib/python3.10/threading.py", line 953, in run
self._target(*self._args, **self._kwargs)
File "/home/traxx/Fooocus/modules/async_worker.py", line 32, in worker
import modules.default_pipeline as pipeline
File "/home/traxx/Fooocus/modules/default_pipeline.py", line 257, in
Traceback (most recent call last):
File "/home/traxx/Fooocus/entry_with_update.py", line 46, in
from launch import *
File "/home/traxx/Fooocus/launch.py", line 136, in
from webui import *
File "/home/traxx/Fooocus/webui.py", line 364, in
for i, (enabled, filename, weight) in enumerate(modules.config.default_loras):
ValueError: not enough values to unpack (expected 3, got 2)
loras=get_enabled_loras(modules.config.default_loras)
File "/home/traxx/Fooocus/modules/util.py", line 366, in get_enabled_loras
return [[lora[1], lora[2]] for lora in loras if lora[0]]
File "/home/traxx/Fooocus/modules/util.py", line 366, in
return [[lora[1], lora[2]] for lora in loras if lora[0]]
IndexError: list index out of range

Thanks in advance

[cuffins]

It seems there is a general issue with the "config.txt". With every start Fooocus is writing a new "config_modification_tutorial.txt" and ignores the entries in "config.txt". Maybe this is the reason for your problem.

[Tra2x]

Yes, I used the original config.txt and replaced my values. Now it's working fine. Thanks

[mashb1t]

Can you please provide the previous config.txt leading to the error mentioned above?

[Tra2x]

config_x.txt
This is the working file. When I add "default_max_lora_number": 10, then I get the mentioned error.

Thanks for asking (y)

[mashb1t]

fixed in #2576

[bsdesigner]

What should I write in the config.txt file to have these checkboxes enabled by default?

[mashb1t]

You'll have to change the code in https://github.com/lllyasviel/Fooocus/blob/main/webui.py#L412 to True as this parameter isn't configurable.

[bsdesigner]

Thanks for the quick reply! But I'm a complete dummy about codes, so I'd better stay out of it ))

[bsdesigner]

Figured it out, did it, it worked, it's all on. Thanks again for the tip!

[mashb1t]

Just be aware of that the automated update might not work anymore when doing changes in your local environment

[generativekey]

I've used all the UIs you can imagine, but nothing compares to this one. The ease, consistency, speed, practicality, and comfort that this UI offers is out of this world. You're truly doing an amazing job. I believe that Fooocus is the best UI we have today.

Not to forget mentioning that the upscale and inpainting really work wonders! Very well done! I am very grateful, friends.

[Kira8822]

Will we have this version 2.3.0 to work offline locally for Windows?

[mashb1t]

@Kira8822 Fooocus always works offline if you already have downloaded all the necessary models. If thr background of your question is referring to a new zip file then the answer is no, this will only be created for python package updates, which wasn't the case in 2.3.0.

[Kira8822]

Excuse my bad English, I'm with the translator, yes that's what I was referring to a new zip file, because version 2.3.0 is really very good, especially because you prioritized the load it had in RAM over the GPU, instead in the zip file doesn't happen so it consumes a lot of ram and that's what is happening to me, even though I have 32 GB of ram it consumes a lot and that GPU prioritization section in the zip would be nice. I don't know if you could implement it if it's not too much trouble?

[mashb1t]

@Kira8822 The Fooocus version in the zip file will automatically update to the latest one on first start, no need to create a new zip file. If you need to have an "offline" zip file, feel free to zip your current folder after creating images, downloading all necessary models and loras etc.
RAM and GPU usage has been optimized in Colab, you can also use the same parameters locally by adding them to your run.bat.
--always-high-vram is the argument which was added in Colab, but there are more possible optimisations you can find in the mentioned PR above.

[Kira8822]

I understand, thank you...

[Cochonfou]

Hello, and thanks a lot for the update.
I am using Fooocus on a M3 Pro mac. The new "lightning" performance setting does not seem to work on my configuration, as it always results in black images or dark color blotches. Everything else works fine as far as I know.

Would you have an idea of what could be wrong ?

Thanks a lot.

{
"prompt": "steam train",
"negative_prompt": "",
"prompt_expansion": "steam train, detailed, elegant, highly color saturated, intricate, full perfect, sharp focus, beautiful, epic composition, cinematic, amazing light, dynamic background, advanced, atmosphere, lively, magical, very inspirational, professional, decorated, stunning, inspired, creative, positive, trendy, cute, adorable, pretty, pure, coherent, fine detail, polished, complex, enhanced",
"styles": "['Fooocus V2', 'Fooocus Enhance', 'Fooocus Sharp']",
"performance": "Lightning",
"resolution": "(1152, 896)",
"guidance_scale": 1.0,
"sharpness": 0.0,
"adm_guidance": "(1.0, 1.0, 0.0)",
"base_model": "juggernautXL_v8Rundiffusion.safetensors",
"refiner_model": "None",
"refiner_switch": 1.0,
"adaptive_cfg": 1.0,
"sampler": "euler",
"scheduler": "sgm_uniform",
"seed": "5000134209646226646",
"lora_combined_6": "sdxl_lightning_4step_lora.safetensors : 1.0",
"metadata_scheme": false,
"version": "Fooocus v2.3.0"
}

Already up-to-date
Update succeeded.
[System ARGV] ['entry_with_update.py']
Python 3.10.14 (main, Mar 21 2024, 11:21:31) [Clang 14.0.6 ]
Fooocus version: 2.3.0
[Cleanup] Attempting to delete content of temp dir /var/folders/m4/w360gldx1xx7nb2v895_533w0000gn/T/fooocus
[Cleanup] Cleanup successful
Total VRAM 36864 MB, total RAM 36864 MB
Set vram state to: SHARED
Always offload VRAM
Device: mps
VAE dtype: torch.float32
Using sub quadratic optimization for cross attention, if you have memory or speed issues try using: --attention-split
Refiner unloaded.
Running on local URL:  http://127.0.0.1:7865

To create a public link, set `share=True` in `launch()`.
model_type EPS
UNet ADM Dimension 2816
Using split attention in VAE
Working with z of shape (1, 4, 32, 32) = 4096 dimensions.
Using split attention in VAE
extra {'cond_stage_model.clip_l.text_projection', 'cond_stage_model.clip_l.logit_scale', 'cond_stage_model.clip_g.transformer.text_model.embeddings.position_ids'}
Base model loaded: /Users/Shared/Fooocus/models/checkpoints/juggernautXL_v8Rundiffusion.safetensors
Request to load LoRAs [['sd_xl_offset_example-lora_1.0.safetensors', 0.1], ['None', 1.0], ['None', 1.0], ['None', 1.0], ['None', 1.0]] for model [/Users/Shared/Fooocus/models/checkpoints/juggernautXL_v8Rundiffusion.safetensors].
Loaded LoRA [/Users/Shared/Fooocus/models/loras/sd_xl_offset_example-lora_1.0.safetensors] for UNet [/Users/Shared/Fooocus/models/checkpoints/juggernautXL_v8Rundiffusion.safetensors] with 788 keys at weight 0.1.
Fooocus V2 Expansion: Vocab with 642 words.
Fooocus Expansion engine loaded for cpu, use_fp16 = False.
Requested to load SDXLClipModel
Requested to load GPT2LMHeadModel
Loading 2 new models
Started worker with PID 37397
App started successful. Use the app with http://127.0.0.1:7865/ or 127.0.0.1:7865
Loaded preset: /Users/Shared/Fooocus/presets/lightning.json
Enter Lightning mode.
[Fooocus] Downloading Lightning components ...
[Parameters] Adaptive CFG = 1.0
[Parameters] Sharpness = 0.0
[Parameters] ControlNet Softness = 0.25
[Parameters] ADM Scale = 1.0 : 1.0 : 0.0
[Parameters] CFG = 1.0
[Parameters] Seed = 5000134209646226646
[Parameters] Sampler = euler - sgm_uniform
[Parameters] Steps = 4 - 4
[Fooocus] Initializing ...
[Fooocus] Loading models ...
Refiner unloaded.
Request to load LoRAs [['None', 1.0], ['None', 1.0], ['None', 1.0], ['None', 1.0], ['None', 1.0], ('sdxl_lightning_4step_lora.safetensors', 1.0)] for model [/Users/Shared/Fooocus/models/checkpoints/juggernautXL_v8Rundiffusion.safetensors].
Loaded LoRA [/Users/Shared/Fooocus/models/loras/sdxl_lightning_4step_lora.safetensors] for UNet [/Users/Shared/Fooocus/models/checkpoints/juggernautXL_v8Rundiffusion.safetensors] with 788 keys at weight 1.0.
Requested to load SDXLClipModel
Loading 1 new model
unload clone 1
[Fooocus] Processing prompts ...
[Fooocus] Preparing Fooocus text #1 ...
[Prompt Expansion] steam train, detailed, elegant, highly color saturated, intricate, full perfect, sharp focus, beautiful, epic composition, cinematic, amazing light, dynamic background, advanced, atmosphere, lively, magical, very inspirational, professional, decorated, stunning, inspired, creative, positive, trendy, cute, adorable, pretty, pure, coherent, fine detail, polished, complex, enhanced
[Fooocus] Encoding positive #1 ...
[Parameters] Denoising Strength = 1.0
[Parameters] Initial Latent shape: Image Space (896, 1152)
Preparation time: 2.18 seconds
[Sampler] refiner_swap_method = joint
[Sampler] sigma_min = 0.6951494812965393, sigma_max = 14.614640235900879
Requested to load SDXL
Loading 1 new model
[Fooocus Model Management] Moving model(s) has taken 9.85 seconds
huggingface/tokenizers: The current process just got forked, after parallelism has already been used. Disabling parallelism to avoid deadlocks...
To disable this warning, you can either:
	- Avoid using `tokenizers` before the fork if possible
	- Explicitly set the environment variable TOKENIZERS_PARALLELISM=(true | false)
100%|█████████████████████████████████████████████| 4/4 [00:11<00:00,  2.99s/it]
Requested to load AutoencoderKL
Loading 1 new model
[Fooocus Model Management] Moving model(s) has taken 1.13 seconds
Image generated with private log at: /Users/Shared/Fooocus/outputs/2024-03-23/log.html
Generating and saving time: 25.87 seconds
Total time: 28.08 seconds

[mashb1t]

=> #2580
Can't reproduce with my M1 Macbook Pro, works for me.

[Cochonfou]

Thanks I will continue in the bug ticket.