[1.1.449] Anyline Preprocessor & MistoLine SDXL model

[huchenlei]

Anyline Preprocessor

Anyline is a ControlNet line preprocessor that accurately extracts object edges, image details, and textual content from most images. Users can input any type of image to quickly obtain line drawings with clear edges, sufficient detail preservation, and high fidelity text, which are then used as input for conditional generation in Stable Diffusion.

https://github.com/TheMistoAI/ComfyUI-Anyline

Misto Line SDXL Model

MistoLine: A Versatile and Robust SDXL-ControlNet Model for Adaptable Line Art Conditioning.

MistoLine is an SDXL-ControlNet model that can adapt to any type of line art input, demonstrating high accuracy and excellent stability. It can generate high-quality images (with a short side greater than 1024px) based on user-provided line art of various types, including hand-drawn sketches, different ControlNet line preprocessors, and model-generated outlines. MistoLine eliminates the need to select different ControlNet models for different line preprocessors, as it exhibits strong generalization capabilities across diverse line art conditions.

We developed MistoLine by employing a novel line preprocessing algorithm Anyline and retraining the ControlNet model based on the Unet of stabilityai/ stable-diffusion-xl-base-1.0, along with innovations in large model training engineering. MistoLine showcases superior performance across different types of line art inputs, surpassing existing ControlNet models in terms of detail restoration, prompt alignment, and stability, particularly in more complex scenarios.

https://github.com/TheMistoAI/MistoLine

How to Use

Download model from https://huggingface.co/TheMistoAI/MistoLine/tree/main and rename the model to mistoLine_softedge_sdxl_fp16.safetensors.

Limitations

As you can see in the generation result, there can be some unwanted detailed being captured by anyline preprocessor, such as incorrectly captured hair detail.

To deal with this, it is recommended to end controlnet guidance early. Here is the result by setting ending control step to 0.5:

Lowering controlnet weight does not help much on smoothing hair details. Also it allows generation of unwanted shapes.

Setting Control Mode to My Prompt is more important does remove unwanted hair details, but it also allows generation of unwanted shapes.

Generation Params:

1girl,, absurdres, highres, (masterpiece:1.2), (best quality, highest quality),
Negative prompt: (lowres, low quality, worst quality:1.2), (text:1.2), watermark, (frame:1.2), deformed, ugly, deformed eyes, blur, out of focus, blurry, deformed cat, deformed, photo, anthropomorphic cat, monochrome, photo, pet collar, gun, weapon, blue, 3d, drones, drone, buildings in background, green
Steps: 30, Sampler: DPM++ 2M, Schedule type: Karras, CFG scale: 7, Seed: 2805618844, Size: 1024x1024, Model hash: aeb7e9e689, Model: juggernautXL_v8Rundiffusion, Clip skip: 2, ControlNet 0: "Module: softedge_anyline, Model: mistoLine_softedge_sdxl_fp16 [2583d73c], Weight: 1.0, Resize Mode: Crop and Resize, Processor Res: 1280, Threshold A: 2.0, Threshold B: 0.5, Guidance Start: 0.0, Guidance End: 1.0, Pixel Perfect: True, Control Mode: My prompt is more important", Version: v1.9.3-13-g8e355f08

[CCpt5]

Excellent! Just tested with something photorealistic and the results are way better than w/ typical Canny.

Stopping it at ~50% definitely seems to produce the best results. Hope people pick up on that.

[Slartibart23]

call me an idiot... but where do i found and the A1111/Forge softedge_anyline preprocessor?? :)

[huchenlei]

It is not supported in Forge yet. To use it, just update your sd-webui-controlnet extension to 1.1.449.

[Slartibart23]

ahh... thanx a lot :-)

[dan4ik94]

@Slartibart23 @huchenlei I'm already using mistoline with lineart & softedge in Forge, works fine. Would anyline be better?

[TimeSurging]

Ask a question: Can you install the Anyline preprocessor in Forge? thanks.

[don1138]

Is this a CUDA/PC-only feature?

I'm on an M1 Mac, and when I tried using the mistoline_rank256.safetensors model with softedge_anyline, I got this error:

Traceback (most recent call last):
  ...
  File "/Users/USER1/stable-diffusion-webui/venv/lib/python3.10/site-packages/torch/serialization.py", line 250, in validate_cuda_device
    raise RuntimeError('Attempting to deserialize object on a CUDA '
RuntimeError: Attempting to deserialize object on a CUDA device but torch.cuda.is_available() is False. If you are running on a CPU-only machine, please use torch.load with map_location=torch.device('cpu') to map your storages to the CPU.

[huchenlei]

Can you file an issue with full stacktrace? thanks!

From the current stacktrace, I am not sure whether it is the TEED model has issue loading or ControlNet model.

[don1138]

In further testing, I found that these preprocessors throw errors:

• softedge_anyline
• softedge_teed

while these work correctly, with no errors:

• softedge_pidinet
• softedge_pidisafe
• softedge_hedsafe
• softedge_hed

This may invalidate this as an issue, but In any case, here's the extended info as requested.

---

I'm running ControlNet v1.1.4491

A111:

• version: v1.9.3-2-gca57a324
• python: 3.10.14
• torch: 2.4.0.dev20240523
• xformers: N/A
• gradio: 3.41.2

---

Here's the output from testing it with Run Preprocessor (the explosion icon):

2024-05-30 10:39:21,080 - ControlNet - INFO - Preview Resolution = 1280
Traceback (most recent call last):
  File "/Users/Barracuda/stable-diffusion-webui/venv/lib/python3.10/site-packages/gradio/routes.py", line 488, in run_predict
    output = await app.get_blocks().process_api(
  File "/Users/Barracuda/stable-diffusion-webui/venv/lib/python3.10/site-packages/gradio/blocks.py", line 1431, in process_api
    result = await self.call_function(
  File "/Users/Barracuda/stable-diffusion-webui/venv/lib/python3.10/site-packages/gradio/blocks.py", line 1103, in call_function
    prediction = await anyio.to_thread.run_sync(
  File "/Users/Barracuda/stable-diffusion-webui/venv/lib/python3.10/site-packages/anyio/to_thread.py", line 33, in run_sync
    return await get_asynclib().run_sync_in_worker_thread(
  File "/Users/Barracuda/stable-diffusion-webui/venv/lib/python3.10/site-packages/anyio/_backends/_asyncio.py", line 877, in run_sync_in_worker_thread
    return await future
  File "/Users/Barracuda/stable-diffusion-webui/venv/lib/python3.10/site-packages/anyio/_backends/_asyncio.py", line 807, in run
    result = context.run(func, *args)
  File "/Users/Barracuda/stable-diffusion-webui/venv/lib/python3.10/site-packages/gradio/utils.py", line 707, in wrapper
    response = f(*args, **kwargs)
  File "/Users/Barracuda/stable-diffusion-webui/extensions/sd-controlnet/scripts/controlnet_ui/controlnet_ui_group.py", line 928, in run_annotator
    result = preprocessor.cached_call(
  File "/Users/Barracuda/stable-diffusion-webui/extensions/sd-controlnet/scripts/supported_preprocessor.py", line 196, in cached_call
    result = self._cached_call(input_image, *args, **kwargs)
  File "/Users/Barracuda/stable-diffusion-webui/extensions/sd-controlnet/scripts/utils.py", line 82, in decorated_func
    return cached_func(*args, **kwargs)
  File "/Users/Barracuda/stable-diffusion-webui/extensions/sd-controlnet/scripts/utils.py", line 66, in cached_func
    return func(*args, **kwargs)
  File "/Users/Barracuda/stable-diffusion-webui/extensions/sd-controlnet/scripts/supported_preprocessor.py", line 209, in _cached_call
    return self(*args, **kwargs)
  File "/Users/Barracuda/stable-diffusion-webui/extensions/sd-controlnet/scripts/preprocessor/teed.py", line 87, in __call__
    self.model = TEEDDetector(mteed=True)
  File "/Users/Barracuda/stable-diffusion-webui/extensions/sd-controlnet/annotator/teed/__init__.py", line 26, in __init__
    self.load_mteed_model()
  File "/Users/Barracuda/stable-diffusion-webui/extensions/sd-controlnet/annotator/teed/__init__.py", line 49, in load_mteed_model
    self.model.load_state_dict(unsafe_torch_load(model_path))
  File "/Users/Barracuda/stable-diffusion-webui/venv/lib/python3.10/site-packages/torch/serialization.py", line 1035, in load
    return _load(opened_zipfile,
  File "/Users/Barracuda/stable-diffusion-webui/venv/lib/python3.10/site-packages/torch/serialization.py", line 1456, in _load
    result = unpickler.load()
  File "/opt/homebrew/Cellar/python@3.10/3.10.14/Frameworks/Python.framework/Versions/3.10/lib/python3.10/pickle.py", line 1213, in load
    dispatch[key[0]](self)
  File "/opt/homebrew/Cellar/python@3.10/3.10.14/Frameworks/Python.framework/Versions/3.10/lib/python3.10/pickle.py", line 1254, in load_binpersid
    self.append(self.persistent_load(pid))
  File "/Users/Barracuda/stable-diffusion-webui/venv/lib/python3.10/site-packages/torch/serialization.py", line 1426, in persistent_load
    typed_storage = load_tensor(dtype, nbytes, key, _maybe_decode_ascii(location))
  File "/Users/Barracuda/stable-diffusion-webui/venv/lib/python3.10/site-packages/torch/serialization.py", line 1400, in load_tensor
    wrap_storage=restore_location(storage, location),
  File "/Users/Barracuda/stable-diffusion-webui/venv/lib/python3.10/site-packages/torch/serialization.py", line 394, in default_restore_location
    result = fn(storage, location)
  File "/Users/Barracuda/stable-diffusion-webui/venv/lib/python3.10/site-packages/torch/serialization.py", line 371, in _deserialize
    device = _validate_device(location, backend_name)
  File "/Users/Barracuda/stable-diffusion-webui/venv/lib/python3.10/site-packages/torch/serialization.py", line 344, in _validate_device
    raise RuntimeError(f'Attempting to deserialize object on a {backend_name.upper()} '
RuntimeError: Attempting to deserialize object on a CUDA device but torch.cuda.is_available() is False. If you are running on a CPU-only machine, please use torch.load with map_location=torch.device('cpu') to map your storages to the CPU.

---

And here's the output from Generate:

2024-05-30 10:40:24,907 - ControlNet - INFO - unit_separate = False, style_align = False
2024-05-30 10:40:25,157 - ControlNet - INFO - Loading model: mistoLine_softedge_sdxl_rank256 [0f320932]
2024-05-30 10:40:25,459 - ControlNet - INFO - Loaded state_dict from [/Users/Barracuda/stable-diffusion-webui/models/ControlNet/mistoLine_softedge_sdxl_rank256.safetensors]
2024-05-30 10:40:25,459 - ControlNet - INFO - Using ControlNet lora (SDXL)
2024-05-30 10:40:25,490 - ControlNet - INFO - ControlNet model mistoLine_softedge_sdxl_rank256 [0f320932](ControlModelType.ControlLoRA) loaded.
2024-05-30 10:40:25,505 - ControlNet - INFO - Using preprocessor: softedge_anyline
2024-05-30 10:40:25,505 - ControlNet - INFO - preprocessor resolution = 1016
*** Error running process: /Users/Barracuda/stable-diffusion-webui/extensions/sd-controlnet/scripts/controlnet.py
    Traceback (most recent call last):
      File "/Users/Barracuda/stable-diffusion-webui/modules/scripts.py", line 825, in process
        script.process(p, *script_args)
      File "/Users/Barracuda/stable-diffusion-webui/extensions/sd-controlnet/scripts/controlnet.py", line 1222, in process
        self.controlnet_hack(p)
      File "/Users/Barracuda/stable-diffusion-webui/extensions/sd-controlnet/scripts/controlnet.py", line 1207, in controlnet_hack
        self.controlnet_main_entry(p)
      File "/Users/Barracuda/stable-diffusion-webui/extensions/sd-controlnet/scripts/controlnet.py", line 941, in controlnet_main_entry
        controls, hr_controls, additional_maps = get_control(
      File "/Users/Barracuda/stable-diffusion-webui/extensions/sd-controlnet/scripts/controlnet.py", line 290, in get_control
        controls, hr_controls = list(zip(*[preprocess_input_image(img) for img in optional_tqdm(input_images)]))
      File "/Users/Barracuda/stable-diffusion-webui/extensions/sd-controlnet/scripts/controlnet.py", line 290, in <listcomp>
        controls, hr_controls = list(zip(*[preprocess_input_image(img) for img in optional_tqdm(input_images)]))
      File "/Users/Barracuda/stable-diffusion-webui/extensions/sd-controlnet/scripts/controlnet.py", line 242, in preprocess_input_image
        result = preprocessor.cached_call(
      File "/Users/Barracuda/stable-diffusion-webui/extensions/sd-controlnet/scripts/supported_preprocessor.py", line 196, in cached_call
        result = self._cached_call(input_image, *args, **kwargs)
      File "/Users/Barracuda/stable-diffusion-webui/extensions/sd-controlnet/scripts/utils.py", line 82, in decorated_func
        return cached_func(*args, **kwargs)
      File "/Users/Barracuda/stable-diffusion-webui/extensions/sd-controlnet/scripts/utils.py", line 66, in cached_func
        return func(*args, **kwargs)
      File "/Users/Barracuda/stable-diffusion-webui/extensions/sd-controlnet/scripts/supported_preprocessor.py", line 209, in _cached_call
        return self(*args, **kwargs)
      File "/Users/Barracuda/stable-diffusion-webui/extensions/sd-controlnet/scripts/preprocessor/teed.py", line 87, in __call__
        self.model = TEEDDetector(mteed=True)
      File "/Users/Barracuda/stable-diffusion-webui/extensions/sd-controlnet/annotator/teed/__init__.py", line 26, in __init__
        self.load_mteed_model()
      File "/Users/Barracuda/stable-diffusion-webui/extensions/sd-controlnet/annotator/teed/__init__.py", line 49, in load_mteed_model
        self.model.load_state_dict(unsafe_torch_load(model_path))
      File "/Users/Barracuda/stable-diffusion-webui/venv/lib/python3.10/site-packages/torch/serialization.py", line 1035, in load
        return _load(opened_zipfile,
      File "/Users/Barracuda/stable-diffusion-webui/venv/lib/python3.10/site-packages/torch/serialization.py", line 1456, in _load
        result = unpickler.load()
      File "/opt/homebrew/Cellar/python@3.10/3.10.14/Frameworks/Python.framework/Versions/3.10/lib/python3.10/pickle.py", line 1213, in load
        dispatch[key[0]](self)
      File "/opt/homebrew/Cellar/python@3.10/3.10.14/Frameworks/Python.framework/Versions/3.10/lib/python3.10/pickle.py", line 1254, in load_binpersid
        self.append(self.persistent_load(pid))
      File "/Users/Barracuda/stable-diffusion-webui/venv/lib/python3.10/site-packages/torch/serialization.py", line 1426, in persistent_load
        typed_storage = load_tensor(dtype, nbytes, key, _maybe_decode_ascii(location))
      File "/Users/Barracuda/stable-diffusion-webui/venv/lib/python3.10/site-packages/torch/serialization.py", line 1400, in load_tensor
        wrap_storage=restore_location(storage, location),
      File "/Users/Barracuda/stable-diffusion-webui/venv/lib/python3.10/site-packages/torch/serialization.py", line 394, in default_restore_location
        result = fn(storage, location)
      File "/Users/Barracuda/stable-diffusion-webui/venv/lib/python3.10/site-packages/torch/serialization.py", line 371, in _deserialize
        device = _validate_device(location, backend_name)
      File "/Users/Barracuda/stable-diffusion-webui/venv/lib/python3.10/site-packages/torch/serialization.py", line 344, in _validate_device
        raise RuntimeError(f'Attempting to deserialize object on a {backend_name.upper()} '
    RuntimeError: Attempting to deserialize object on a CUDA device but torch.cuda.is_available() is False. If you are running on a CPU-only machine, please use torch.load with map_location=torch.device('cpu') to map your storages to the CPU.

[changpeidong]

Why can I only generate 1:1 images in webui, the size is 1024x1024, and other ratios are not effective?

[dongshaojia]

Set Pixel Perfect to false

[zou-yf]

Could you please tell me why the image generated by my Anyline preprocessor looks like this?

[dongshaojia]

你需要下载对应的controlnet模型，他会自动从huggingface上下载，可能你访问不了huggingface，你可以在控制台看看那个模型的名称，从魔搭社区或者网上找可以访问的下载地址，下载到对应的文件夹 At 2024-07-18 16:01:31, "zou-yf" ***@***.***> wrote: Could you please tell me why the image generated by my Anyline preprocessor looks like this? 347763764-f7a73885-d180-44dc-9504-a1a7b0e7d3d3.png (view on web) — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you commented.Message ID: ***@***.***>

[zou-yf]

我下载了MTEED.pth模型，并放置在了sd-webui-controlnet\annotator\downloads\TEED 目录下，我在4080的机器上是可以正常处理的，可是我在a100的机器上无法正常生成线稿，这个图可能看的比较明显

[dongshaojia]

模型那点后边的那个刷新一下，在列表里选softedge的模型试试 发自我的iPhone

…

在 2024年7月19日，08:48，zou-yf ***@***.***> 写道：  我下载了MTEED.pth模型，并放置在了sd-webui-controlnet\annotator\downloads\TEED 目录下，我在4080的机器上是可以正常处理的，可是我在a100的机器上无法正常生成线稿，这个图可能看的比较明显，我不知道为什么 — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you commented.

[zou-yf]

还是一样的效果，不管我是否加载模型

[Mochalif085]

can i use this for sd1.5?. is this only for sdxl?

[dan4ik94]

only sdxl.