michael/stable-diffusion-webui
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity

improve code quality

Browse source
This commit is contained in:
C43H66N12O12S2 2022-09-29 13:30:33 +03:00 committed by AUTOMATIC1111
parent b6f80bdcc2
commit 965dcf4469
1 changed files with 7 additions and 9 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

16
modules/sd_samplers.py
View file

@ -38,9 +38,7 @@ samplers = [
SamplerData('DDIM', lambda model: VanillaStableDiffusionSampler(ldm.models.diffusion.ddim.DDIMSampler, model), []),
SamplerData('PLMS', lambda model: VanillaStableDiffusionSampler(ldm.models.diffusion.plms.PLMSSampler, model), []),
]
samplers_for_img2img = [x for x in samplers if x.name != 'PLMS']
samplers_for_img2img.remove(samplers_for_img2img[6])
samplers_for_img2img.remove(samplers_for_img2img[6])
samplers_for_img2img = [x for x in samplers if x.name not in ['PLMS', 'DPM fast', 'DPM adaptive']]
sampler_extra_params = {
'sample_euler': ['s_churn', 's_tmin', 's_tmax', 's_noise'],
@ -314,12 +312,12 @@ class KDiffusionSampler:
extra_params_kwargs = self.initialize(p)
if 'sigma_min' in inspect.signature(self.func).parameters:
extra_params_kwargs['sigma_min'] = self.model_wrap.sigmas[0].item()
extra_params_kwargs['sigma_max'] = self.model_wrap.sigmas[-1].item()
if 'n' in inspect.signature(self.func).parameters:
samples = self.func(self.model_wrap_cfg, x, sigma_min=self.model_wrap.sigmas[0].item(), sigma_max=self.model_wrap.sigmas[-1].item(), n=steps, extra_args={'cond': conditioning, 'uncond': unconditional_conditioning, 'cond_scale': p.cfg_scale}, disable=False, callback=self.callback_state, **extra_params_kwargs)
return samples
samples = self.func(self.model_wrap_cfg, x, sigma_min=self.model_wrap.sigmas[0].item(), sigma_max=self.model_wrap.sigmas[-1].item(), extra_args={'cond': conditioning, 'uncond': unconditional_conditioning, 'cond_scale': p.cfg_scale}, disable=False, callback=self.callback_state, **extra_params_kwargs)
return samples
samples = self.func(self.model_wrap_cfg, x, sigmas, extra_args={'cond': conditioning, 'uncond': unconditional_conditioning, 'cond_scale': p.cfg_scale}, disable=False, callback=self.callback_state, **extra_params_kwargs)
extra_params_kwargs['n'] = steps
else:
extra_params_kwargs['sigmas'] = sigmas
samples = self.func(self.model_wrap_cfg, x, extra_args={'cond': conditioning, 'uncond': unconditional_conditioning, 'cond_scale': p.cfg_scale}, disable=False, callback=self.callback_state, **extra_params_kwargs)
return samples