entirely black image generated

[DennisPeeters]

Windows 10 installation
RTX 3060 laptop driver 531.79
Tried different models

All images come out black, not sure how to find a solution here

[herohung093]

I encounter the same issue, I ran fooocus on MacOS

[mlison]

same issue, also on mac

[marianoarga]

Same issue on Mac, i58400 RX6600XT

[so-joinplank]

same issue on mac pro m1

[mashb1t]

The issue has been addressed. Please make sure to use the latest version of Fooocus.
You can find more information about this here:

• [Bug]: Rendering Black images with sdxl AUTOMATIC1111/stable-diffusion-webui#12019
• [Issue]: SDXL 1.0 Producing Black Images vladmandic/sdnext#1858
• https://huggingface.co/stabilityai/stable-diffusion-2-1/discussions/9

[marianoarga]

The issue persist, using latest main branch code

[mashb1t]

Can you please share

• the model you were using
• the arguments for entry_with_update.py

with us? This allows us to debug if it's an fp16 sai issue or something with Fooocus.

[marianoarga]

Sure thing:

main branch, up to date
MacOs Monterrey, 32GB RAM, i58400 RX6600 XT 8GB

• juggernautXL_version6Rundiffusion.safetensors
• --preset realistic

[marianoarga]

Being run with --preset realistic changes the model to the "realisticStockPhoto_v10", so I tested both

[marianoarga]

The log output so far, already turned black the image (was initially entropy), it will remain black until the end of the process:

entry_with_update.py --preset realistic
Already up-to-date
Update succeeded.
[System ARGV] ['entry_with_update.py', '--preset', 'realistic']
Loaded preset: /Users/mariano/development/AI/Fooocus/presets/realistic.json
Failed to load config key: {"path_checkpoints": "/Users/mariano/development/Fooocus/models/checkpoints"} is invalid or does not exist; will use {"path_checkpoints": "../models/checkpoints/"} instead.
Failed to load config key: {"path_loras": "/Users/mariano/development/Fooocus/models/loras"} is invalid or does not exist; will use {"path_loras": "../models/loras/"} instead.
Failed to load config key: {"path_embeddings": "/Users/mariano/development/Fooocus/models/embeddings"} is invalid or does not exist; will use {"path_embeddings": "../models/embeddings/"} instead.
Failed to load config key: {"path_vae_approx": "/Users/mariano/development/Fooocus/models/vae_approx"} is invalid or does not exist; will use {"path_vae_approx": "../models/vae_approx/"} instead.
Failed to load config key: {"path_upscale_models": "/Users/mariano/development/Fooocus/models/upscale_models"} is invalid or does not exist; will use {"path_upscale_models": "../models/upscale_models/"} instead.
Failed to load config key: {"path_inpaint": "/Users/mariano/development/Fooocus/models/inpaint"} is invalid or does not exist; will use {"path_inpaint": "../models/inpaint/"} instead.
Failed to load config key: {"path_controlnet": "/Users/mariano/development/Fooocus/models/controlnet"} is invalid or does not exist; will use {"path_controlnet": "../models/controlnet/"} instead.
Failed to load config key: {"path_clip_vision": "/Users/mariano/development/Fooocus/models/clip_vision"} is invalid or does not exist; will use {"path_clip_vision": "../models/clip_vision/"} instead.
Failed to load config key: {"path_fooocus_expansion": "/Users/mariano/development/Fooocus/models/prompt_expansion/fooocus_expansion"} is invalid or does not exist; will use {"path_fooocus_expansion": "../models/prompt_expansion/fooocus_expansion"} instead.
Failed to load config key: {"path_outputs": "/Users/mariano/development/Fooocus/outputs"} is invalid or does not exist; will use {"path_outputs": "../outputs/"} instead.
Python 3.10.13 (main, Nov 1 2023, 16:44:37) [Clang 14.0.0 (clang-1400.0.29.202)]
Fooocus version: 2.1.860
Running on local URL: http://127.0.0.1:7865

To create a public link, set share=True in launch().
Total VRAM 32768 MB, total RAM 32768 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.
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.logit_scale', 'cond_stage_model.clip_l.text_projection', 'cond_stage_model.clip_g.transformer.text_model.embeddings.position_ids'}
Base model loaded: /Users/mariano/development/AI/Fooocus/models/checkpoints/realisticStockPhoto_v10.safetensors
Request to load LoRAs [['SDXL_FILM_PHOTOGRAPHY_STYLE_BetaV0.4.safetensors', 0.25], ['None', 1.0], ['None', 1.0], ['None', 1.0], ['None', 1.0]] for model [/Users/mariano/development/AI/Fooocus/models/checkpoints/realisticStockPhoto_v10.safetensors].
Loaded LoRA [/Users/mariano/development/AI/Fooocus/models/loras/SDXL_FILM_PHOTOGRAPHY_STYLE_BetaV0.4.safetensors] for UNet [/Users/mariano/development/AI/Fooocus/models/checkpoints/realisticStockPhoto_v10.safetensors] with 788 keys at weight 0.25.
Loaded LoRA [/Users/mariano/development/AI/Fooocus/models/loras/SDXL_FILM_PHOTOGRAPHY_STYLE_BetaV0.4.safetensors] for CLIP [/Users/mariano/development/AI/Fooocus/models/checkpoints/realisticStockPhoto_v10.safetensors] with 264 keys at weight 0.25.
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
[Fooocus Model Management] Moving model(s) has taken 4.05 seconds
App started successful. Use the app with http://127.0.0.1:7865/ or 127.0.0.1:7865
[Parameters] Adaptive CFG = 7
[Parameters] Sharpness = 2
[Parameters] ADM Scale = 1.5 : 0.8 : 0.3
[Parameters] CFG = 3.0
[Parameters] Seed = 6473121032613553998
[Parameters] Sampler = dpmpp_2m_sde_gpu - karras
[Parameters] Steps = 30 - 15
[Fooocus] Initializing ...
[Fooocus] Loading models ...
Refiner unloaded.
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.logit_scale', 'cond_stage_model.clip_l.text_projection', 'cond_stage_model.clip_g.transformer.text_model.embeddings.position_ids'}
Base model loaded: /Users/mariano/development/AI/Fooocus/models/checkpoints/juggernautXL_version6Rundiffusion.safetensors
Request to load LoRAs [['SDXL_FILM_PHOTOGRAPHY_STYLE_BetaV0.4.safetensors', 0.25], ['None', 1.0], ['None', 1.0], ['None', 1.0], ['None', 1.0]] for model [/Users/mariano/development/AI/Fooocus/models/checkpoints/juggernautXL_version6Rundiffusion.safetensors].
Loaded LoRA [/Users/mariano/development/AI/Fooocus/models/loras/SDXL_FILM_PHOTOGRAPHY_STYLE_BetaV0.4.safetensors] for UNet [/Users/mariano/development/AI/Fooocus/models/checkpoints/juggernautXL_version6Rundiffusion.safetensors] with 788 keys at weight 0.25.
Loaded LoRA [/Users/mariano/development/AI/Fooocus/models/loras/SDXL_FILM_PHOTOGRAPHY_STYLE_BetaV0.4.safetensors] for CLIP [/Users/mariano/development/AI/Fooocus/models/checkpoints/juggernautXL_version6Rundiffusion.safetensors] with 264 keys at weight 0.25.
Requested to load SDXLClipModel
Loading 1 new model
[Fooocus Model Management] Moving model(s) has taken 3.04 seconds
[Fooocus] Processing prompts ...
[Fooocus] Preparing Fooocus text #1 ...
[Prompt Expansion] a cute dog, vivid colors, elegant, highly detailed, sharp focus, intricate, innocent, fine aesthetic, colorful, magical, mystical, winning, deep background, professional, cinematic, ambient, artistic, sublime, extremely inspirational, composed, beautiful, dramatic, thought, epic, stunning, light, pristine, magic, pure, full, strong, creative, loving, amazing
[Fooocus] Encoding positive #1 ...
[Fooocus] Encoding negative #1 ...
[Parameters] Denoising Strength = 1.0
[Parameters] Initial Latent shape: Image Space (1152, 896)
Preparation time: 53.99 seconds
[Sampler] refiner_swap_method = joint
[Sampler] sigma_min = 0.0291671771556139, sigma_max = 14.614643096923828
Requested to load SDXL
Loading 1 new model
[Fooocus Model Management] Moving model(s) has taken 12.86 seconds
/Users/mariano/development/AI/Fooocus/ldm_patched/k_diffusion/sampling.py:699: UserWarning: MPS: nonzero op is supported natively starting from macOS 13.0. Falling back on CPU. This may have performance implications. (Triggered internally at /Users/runner/work/pytorch/pytorch/pytorch/aten/src/ATen/native/mps/operations/Indexing.mm:283.)
sigma_min, sigma_max = sigmas[sigmas > 0].min(), sigmas.max()
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)
0%| | 0/30 [00:00<?, ?it/s]/usr/local/lib/python3.10/site-packages/torch/nn/functional.py:3983: UserWarning: MPS: 'nearest' mode upsampling is supported natively starting from macOS 13.0. Falling back on CPU. This may have performance implications. (Triggered internally at /Users/runner/work/pytorch/pytorch/pytorch/aten/src/ATen/native/mps/operations/UpSample.mm:255.)
return torch._C._nn.upsample_nearest2d(input, output_size, scale_factors)
/Users/mariano/development/AI/Fooocus/modules/anisotropic.py:132: UserWarning: The operator 'aten::std_mean.correction' is not currently supported on the MPS backend and will fall back to run on the CPU. This may have performance implications. (Triggered internally at /Users/runner/work/pytorch/pytorch/pytorch/aten/src/ATen/mps/MPSFallback.mm:13.)
s, m = torch.std_mean(g, dim=(1, 2, 3), keepdim=True)
27%|████████████████████████████████████████▊ | 8/30 [02:26<06:00, 16.40s/it]/Users/mariano/development/AI/Fooocus/modules/core.py:260: RuntimeWarning: invalid value encountered in cast
x_sample = x_sample.cpu().numpy().clip(0, 255).astype(np.uint8)
43%|█████████████████████████████████████████████████████████████████▊

[mashb1t]

same issue on mac pro m1

@so-joinplank @marianoarga i just tested on a M1 Macbook Pro, works without issues in conda environment as described in https://github.com/lllyasviel/Fooocus?tab=readme-ov-file#mac. Please make sure you have followed all steps of the instruction and please confirm this also happens on a fresh installation.

You may use --disable-offload-from-vram to boost speed though between generations.

[marianoarga]

The issue persist, may be related to my i5 mac, thank you for your time, will test on other devices

[kamil6x]

@mashb1t
I reinstalled with the instructions from here: https://github.com/lllyasviel/Fooocus?tab=readme-ov-file#mac
and still getting a "grey" image now. The GPU is doing some kind of processing but not getting any image out. FYI: DiffusionBee works with my setup, but cannot load models into it.


[mashb1t]

You might try to use --always-cpu and check if inference in general works, but most likely the combination of MacOS with Radeon Graphics doesn't work and Fooocus doesn't offer official support for this.

Mac is not intensively tested. Below is an unofficial guideline for using Mac. You can discuss problems #129.

Maybe there already is a solution for you in #129, please try your luck there and feel free to reference this issue there. I'm so sorry!

[kamil6x]

Thanks for the response. Switched to something else as macOS is not really supported.

[Bonghui]

The issue has been addressed. Please make sure to use the latest version of Fooocus. You can find more information about this here:

• [Bug]: Rendering Black images with sdxl AUTOMATIC1111/stable-diffusion-webui#12019
• [Issue]: SDXL 1.0 Producing Black Images vladmandic/automatic#1858
• https://huggingface.co/stabilityai/stable-diffusion-2-1/discussions/9

How do I check the fooocus version and update?

[mashb1t]

@Bonghui Fooocus tries to auto-update on startup. You can find the version either in the console log on startup, in the browser tab title or in in the file fooocus_version.txt, where the current version 2.4.1 should be shown.