[BUG] sonar is broken :(

[Rogal80]

Hi, awesome script :) but it stopped working after las commit from Auto , I have this issue

Traceback (most recent call last):
File "I:\Github\stable-diffusion-webui\modules\call_queue.py", line 45, in f
res = list(func(*args, **kwargs))
File "I:\Github\stable-diffusion-webui\modules\call_queue.py", line 28, in f
res = func(*args, **kwargs)
File "I:\Github\stable-diffusion-webui\modules\txt2img.py", line 46, in txt2img
processed = modules.scripts.scripts_txt2img.run(p, *args)
File "I:\Github\stable-diffusion-webui\modules\scripts.py", line 328, in run
processed = script.run(p, *script_args)
File "I:\Github\stable-diffusion-webui\extensions\stable-diffusion-webui-sonar\scripts\sonar.py", line 951, in run
proc = process_images(p)
File "I:\Github\stable-diffusion-webui\extensions\stable-diffusion-webui-sonar\scripts\sonar.py", line 78, in process_images
res = process_images_inner(p)
File "I:\Github\stable-diffusion-webui\extensions\stable-diffusion-webui-sonar\scripts\sonar.py", line 180, in process_images_inner
samples_ddim = sample_func(p, conditioning=c, unconditional_conditioning=uc, seeds=seeds, subseeds=subseeds, subseed_strength=p.subseed_strength, prompts=prompts)
File "I:\Github\stable-diffusion-webui\extensions\stable-diffusion-webui-sonar\scripts\sonar.py", line 271, in StableDiffusionProcessingTxt2Img_sample
x = create_random_tensors([opt_C, self.firstphase_height // opt_f, self.firstphase_width // opt_f], seeds=seeds, subseeds=subseeds, subseed_strength=self.subseed_strength, seed_resize_from_h=self.seed_resize_from_h, seed_resize_from_w=self.seed_resize_from_w, p=self)
AttributeError: 'StableDiffusionProcessingTxt2Img' object has no attribute 'firstphase_height'

Please can you fix this ?

[Kahsolt]

Thanks for informing~ Now it should be fixed~ 😃

[Rogal80]

Hi, thanks for fix but still probably we will need to have another fix :( , right now, there is a big discussion about new highres fix, everybody are confused : AUTOMATIC1111/stable-diffusion-webui#6219

[Kahsolt]

Wow, it's really a long and interesting discussion. 😃
Since I rarely used hires.fix before, I'm looking into and testing it to get some understanding right now.

[Kahsolt]

Oh bad news, sonar grad is henceforth dead.
I updated readme for explaination, in short, this commit AUTOMATIC1111/stable-diffusion-webui@bd68e35 breaks gradient calculation. Currently there is no work around.

[Rogal80]

yea, I notice another issue but its only appear with latent upscalers :

Traceback (most recent call last):
File "I:\Github\new_webui\stable-diffusion-webui\modules\call_queue.py", line 45, in f
res = list(func(*args, **kwargs))
File "I:\Github\new_webui\stable-diffusion-webui\modules\call_queue.py", line 28, in f
res = func(*args, **kwargs)
File "I:\Github\new_webui\stable-diffusion-webui\modules\txt2img.py", line 46, in txt2img
processed = modules.scripts.scripts_txt2img.run(p, *args)
File "I:\Github\new_webui\stable-diffusion-webui\modules\scripts.py", line 328, in run
processed = script.run(p, *script_args)
File "I:\Github\new_webui\stable-diffusion-webui\extensions\stable-diffusion-webui-sonar\scripts\sonar.py", line 963, in run
proc = process_images(p)
File "I:\Github\new_webui\stable-diffusion-webui\extensions\stable-diffusion-webui-sonar\scripts\sonar.py", line 80, in process_images
res = process_images_inner(p)
File "I:\Github\new_webui\stable-diffusion-webui\extensions\stable-diffusion-webui-sonar\scripts\sonar.py", line 187, in process_images_inner
samples_ddim = sample_func(p, conditioning=c, unconditional_conditioning=uc, seeds=seeds, subseeds=subseeds, subseed_strength=p.subseed_strength, prompts=prompts)
File "I:\Github\new_webui\stable-diffusion-webui\extensions\stable-diffusion-webui-sonar\scripts\sonar.py", line 300, in StableDiffusionProcessingTxt2Img_sample
samples = torch.nn.functional.interpolate(samples, size=(target_height // opt_f, target_width // opt_f), mode=latent_scale_mode)
File "I:\Github\new_webui\stable-diffusion-webui\venv\lib\site-packages\torch\nn\functional.py", line 3961, in interpolate
raise NotImplementedError(
NotImplementedError: Input Error: Only 3D, 4D and 5D input Tensors supported (got 4D) for the modes: nearest | linear | bilinear | bicubic | trilinear | area | nearest-exact (got {'mode': 'bilinear', 'antialias': False})

[Kahsolt]

Not sonar's fault, this is due to a code conflict by maybe two commiters working on hires.fix Upscaler option.
In the latest commit, this code line in modules/processing.py:

samples = torch.nn.functional.interpolate(samples, size=(target_height // opt_f, target_width // opt_f), mode=latent_scale_mode)

has been changed to

samples = torch.nn.functional.interpolate(samples, size=(target_height // opt_f, target_width // opt_f), mode=latent_scale_mode["mode"], antialias=latent_scale_mode["antialias"])

which fixed this bug.

[Rogal80]

thx :)