From 692a0b24224b389d9fbd6c1130c1b33a88d28376 Mon Sep 17 00:00:00 2001
From: layerdiffusion <19834515+lllyasviel@users.noreply.github.com>
Date: Wed, 7 Aug 2024 19:01:19 -0700
Subject: [PATCH] remove confusing codes that actually do nothing

---
 modules/sd_samplers_kdiffusion.py |  4 ----
 modules/sd_samplers_timesteps.py  | 13 -------------
 2 files changed, 17 deletions(-)

diff --git a/modules/sd_samplers_kdiffusion.py b/modules/sd_samplers_kdiffusion.py
index 5d304a28..dedd6a4a 100644
--- a/modules/sd_samplers_kdiffusion.py
+++ b/modules/sd_samplers_kdiffusion.py
@@ -133,8 +133,6 @@ class KDiffusionSampler(sd_samplers_common.Sampler):
         unet_patcher = self.model_wrap.inner_model.forge_objects.unet
         sampling_prepare(self.model_wrap.inner_model.forge_objects.unet, x=x)
 
-        self.model_wrap.predictor.to(x.device)
-
         steps, t_enc = sd_samplers_common.setup_img2img_steps(p, steps)
 
         sigmas = self.get_sigmas(p, steps).to(x.device)
@@ -194,8 +192,6 @@ class KDiffusionSampler(sd_samplers_common.Sampler):
         unet_patcher = self.model_wrap.inner_model.forge_objects.unet
         sampling_prepare(self.model_wrap.inner_model.forge_objects.unet, x=x)
 
-        self.model_wrap.predictor.to(x.device)
-
         steps = steps or p.steps
 
         sigmas = self.get_sigmas(p, steps).to(x.device)
diff --git a/modules/sd_samplers_timesteps.py b/modules/sd_samplers_timesteps.py
index 3e410f4d..514e1d63 100644
--- a/modules/sd_samplers_timesteps.py
+++ b/modules/sd_samplers_timesteps.py
@@ -38,21 +38,8 @@ class CFGDenoiserTimesteps(CFGDenoiser):
 
     def __init__(self, sampler):
         super().__init__(sampler)
-
-        self.alphas = 1.0 / (shared.sd_model.forge_objects.unet.model.predictor.sigmas ** 2.0 + 1.0)
         self.classic_ddim_eps_estimation = True
 
-    def get_pred_x0(self, x_in, x_out, sigma):
-        ts = sigma.to(dtype=int)
-        self.alphas = self.alphas.to(ts.device)
-
-        a_t = self.alphas[ts][:, None, None, None]
-        sqrt_one_minus_at = (1 - a_t).sqrt()
-
-        pred_x0 = (x_in - sqrt_one_minus_at * x_out) / a_t.sqrt()
-
-        return pred_x0
-
     @property
     def inner_model(self):
         if self.model_wrap is None: