diff --git a/backend/sampling/condition.py b/backend/sampling/condition.py
index ff6eb7a2..b594d1b8 100644
--- a/backend/sampling/condition.py
+++ b/backend/sampling/condition.py
@@ -89,6 +89,9 @@ class ConditionConstant(Condition):
 
 
 def compile_conditions(cond):
+    if cond is None:
+        return None
+
     if isinstance(cond, torch.Tensor):
         result = dict(
             cross_attn=cond,
diff --git a/modules/processing.py b/modules/processing.py
index d35a8303..85f7737b 100644
--- a/modules/processing.py
+++ b/modules/processing.py
@@ -483,7 +483,12 @@ class StableDiffusionProcessing:
         self.step_multiplier = total_steps // self.steps
         self.firstpass_steps = total_steps
 
-        self.uc = self.get_conds_with_caching(prompt_parser.get_learned_conditioning, negative_prompts, total_steps, [self.cached_uc], self.extra_network_data)
+        if self.cfg_scale == 1:
+            self.uc = None
+            print('Skipping unconditional conditioning when CFG = 1')
+        else:
+            self.uc = self.get_conds_with_caching(prompt_parser.get_learned_conditioning, negative_prompts, total_steps, [self.cached_uc], self.extra_network_data)
+
         self.c = self.get_conds_with_caching(prompt_parser.get_multicond_learned_conditioning, prompts, total_steps, [self.cached_c], self.extra_network_data)
 
     def get_conds(self):
diff --git a/modules/sd_samplers_cfg_denoiser.py b/modules/sd_samplers_cfg_denoiser.py
index b329a6c9..d7146600 100644
--- a/modules/sd_samplers_cfg_denoiser.py
+++ b/modules/sd_samplers_cfg_denoiser.py
@@ -173,7 +173,7 @@ class CFGDenoiser(torch.nn.Module):
             uncond = self.sampler.sampler_extra_args['uncond']
 
         cond_composition, cond = prompt_parser.reconstruct_multicond_batch(cond, self.step)
-        uncond = prompt_parser.reconstruct_cond_batch(uncond, self.step)
+        uncond = prompt_parser.reconstruct_cond_batch(uncond, self.step) if uncond is not None else None
 
         if self.mask is not None:
             noisy_initial_latent = self.init_latent + sigma[:, None, None, None] * torch.randn_like(self.init_latent).to(self.init_latent)