tjm35 / asymmetric-tiling-sd-webui Goto Github PK

I got this when I tried loading this script:
Error loading script: asymmetric_tiling.py Traceback (most recent call last): File "E:\stable-diffusion-webui\modules\scripts.py", line 69, in load_scripts compiled = compile(text, path, 'exec') File "E:\stable-diffusion-webui\scripts\asymmetric_tiling.py", line 70 <title>asymmetric-tiling-sd-webui/asymmetric_tiling.py at main · tjm35/asymmetric-tiling-sd-webui</title> ^ SyntaxError: invalid character '·' (U+00B7)

Hi,

Do you think it would possible to add asymetric tiling such as:

• left / top
• right / bot
• left / bot
• right / top
  ?

Hi, I noticed when looking the auto1111 output that every frame was throwing a error with the asymmetric-tiling script installed.

error:

Traceback (most recent call last):
  File "E:\Stable-Diffusion\auto1111-new\modules\scripts.py", line 417, in process
    script.process(p, *script_args)
TypeError: Script.process() missing 5 required positional arguments: 'active', 'tileX', 'tileY', 'startStep', and 'stopStep'```

I'm not sure what the issue is, I talked to Deforum Dev's and they said to reach out to you guys.  Ideally it would be awesome to use this in deforum as well.

Thank you for all your hard work!

Hi! I'm using Unity to connect to the local Automatic1111 API to create dynamic textures on the fly. Is there a way to access the TileX/ TileY settings using the API? Thanks!

Comfyui is much faster than A1111. Really want to make pano with SDXL

To create a public link, set share=True in launch().
*** Failed reading extension data from Git repository (asymmetric-tiling-sd-webui)
Traceback (most recent call last):
File "K:\StableDiffusion\sdwebui\modules\extensions.py", line 63, in do_read_info_from_repo
commit = repo.head.commit
File "K:\StableDiffusion\sdwebui\py310\lib\site-packages\git\refs\symbolic.py", line 226, in _get_commit
obj = self._get_object()
File "K:\StableDiffusion\sdwebui\py310\lib\site-packages\git\refs\symbolic.py", line 219, in _get_object
return Object.new_from_sha(self.repo, hex_to_bin(self.dereference_recursive(self.repo, self.path)))
File "K:\StableDiffusion\sdwebui\py310\lib\site-packages\git\objects\base.py", line 94, in new_from_sha
oinfo = repo.odb.info(sha1)
File "K:\StableDiffusion\sdwebui\py310\lib\site-packages\git\db.py", line 40, in info
hexsha, typename, size = self._git.get_object_header(bin_to_hex(binsha))
File "K:\StableDiffusion\sdwebui\modules\gitpython_hack.py", line 18, in get_object_header
ret = subprocess.check_output(
File "subprocess.py", line 421, in check_output
File "subprocess.py", line 505, in run
File "subprocess.py", line 1154, in communicate
File "subprocess.py", line 1546, in _communicate
subprocess.TimeoutExpired: Command '['git', 'cat-file', '--batch-check']' timed out after 2 seconds

For some reason the Asymmetric Tiling extension is not listed among the extensions available for download and install on the Automatic1111 Webui Extension tab.

Also, if you try to copy-paste the extension's github address on the Install from URL tab, it tells you the extension folder already exists and doesn't do anything.

I used to use the script version but I accidentally deleted it and when I went to reinstall it I discovered there was an extension version - which is much more useful as it can be combined with other scripts now ! Anyways, had I known this was available since November, I would have installed it earlier. And now I am wondering if maybe many other people were not aware of its existence.

Thanks a lot for your Asymmetric Tiling tool - I use it very often and it does its job perfectly. All that was missing, in my uninformed opinion, was an extension version!

Enabling tiling has a tendency to cause images to become much more homogenous, especially if it is enabled for step 0.

Hi, thank you for the asymmetric tiling script, very good results !

I was wondering about what do other type of padding do ? those defined in https://github.com/pytorch/pytorch/blob/master/torch/nn/modules/conv.py
i.e. reflect and replicate ? Are they useless ?

Hello friend, the new version forge of sd-webui cannot use asymmetric controllers, can you upgrade it? Thanks a million

*** Error running postprocess: D:\forgeUI-N\webui\extensions\asymmetric-tiling-sd-webui\scripts\asymmetric_tiling.py
Traceback (most recent call last):
File "D:\forgeUI-N\webui\modules\scripts.py", line 838, in postprocess
script.postprocess(p, processed, *script_args)
File "D:\forgeUI-N\webui\extensions\asymmetric-tiling-sd-webui\scripts\asymmetric_tiling.py", line 56, in postprocess
self.__restoreConv2DMethods()
File "D:\forgeUI-N\webui\extensions\asymmetric-tiling-sd-webui\scripts\asymmetric_tiling.py", line 74, in __restoreConv2DMethods
for layer in modules.sd_hijack.model_hijack.layers:
TypeError: 'NoneType' object is not iterable

when i using diffusers, i replace vae and unet conv2d forward like __replacementConv2DConvForward, but generate also have seam.Is this right to do?

How to set up active, tileX, tileY in txt2img API?
I set these as:
"active": "True",
"tileX": "True",
"tileY": "False",
but not working!

Was anyone able to replicate this in a standalone script without using WebUI?

This extension is super useful, I was wondering if a bit more artistic control could be added to give variety ?

adding a padding / falloff value would be amazing

Hi, I use everyday your fabulous X-Tiling feature for 360-panoramas.
It would be great, if your Tiling would apply in the Highres-fix-Step as well.
Hope it can be done without too much effort

here you see a seam due to Highresfix on and X-Tiling does not apply anymore

As title; currently our approach to tiling is to circularize every 2D convolution in the model regardless of size or purpose, on the assumption that every 2D structure represents image data in a spatial layout.

This does generally seem to work fine, but it is possible that some distortion is being caused by circularizing convolutions that are operating on other kinds of data. Resolving this issue would require going through the SD model, understanding the purpose of each use of Conv2D, then modifying this script so that it only circularizes convolutions that genuinely represent image-space data.

Recently switched to using Forge instead of A111.Its giving me error script can't be run for some reason and its not tiling.Is it possible to use this with forge UI?

https://gyazo.com/3665ceaede366cb1f0205d76cf517c44