improve code quality

modules/sd_samplers.py

@@ -38,9 +38,7 @@
     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 @@ def sample(self, p, x, conditioning, unconditional_conditioning, steps=None):
 
         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