fix pipe detection when processing prompts

javascript/black-teal.css

@@ -27,7 +27,7 @@
   --button-secondary-text-color: white;
   --button-secondary-background-fill: linear-gradient(to bottom right, var(--neutral-400), var(--neutral-700));
   --button-secondary-background-fill-hover: linear-gradient(to bottom right, var(--neutral-700), var(--neutral-400));
-  --block-title-text-color: var(--neutral-900);
+  --block-title-text-color: var(--neutral-300);
   --radius-sm: 2px;
   --radius-lg: 4px;
   --spacing-md: 4px;
@@ -134,7 +134,7 @@ svg.feather.feather-image, .feather .feather-image { display: none }
   --background-fill-secondary: none;
   --border-color-accent: var(--background-color);
   --border-color-primary: var(--background-color);
-  --link-text-color-active: var(--secondary-500);
+  --link-text-color-active: var(--primary-500);
   --link-text-color: var(--secondary-500);
   --link-text-color-hover: var(--secondary-400);
   --link-text-color-visited: var(--secondary-600);

modules/prompt_parser_diffusers.py

@@ -118,15 +118,15 @@ def compel_encode_prompt(
     is_refiner: bool = None,
     clip_skip: typing.Optional[int] = None,
 ):
-    if shared.sd_model_type not in {"sd", "sdxl"}:
-        shared.log.warning(f"Prompt parser: Compel not supported: {type(pipeline).__name__}")
+    if 'StableDiffusion' not in pipeline.__class__.__name__:
+        shared.log.warning(f"Prompt parser: Compel not supported: {pipeline.__class__.__name__}")
         return (None, None, None, None)
 
-    if not is_refiner and shared.sd_model_type == "sdxl":
+    if 'XL' in pipeline.__class__.__name__ and not is_refiner:
         embedding_type = ReturnedEmbeddingsType.PENULTIMATE_HIDDEN_STATES_NON_NORMALIZED
         if clip_skip is not None and clip_skip > 1:
             shared.log.warning(f"Prompt parser SDXL unsupported: clip_skip={clip_skip}")
-    elif is_refiner and shared.sd_refiner_type == "sdxl":
+    elif 'XL' in pipeline.__class__.__name__ and is_refiner:
         embedding_type = ReturnedEmbeddingsType.PENULTIMATE_HIDDEN_STATES_NON_NORMALIZED
         if clip_skip is not None and clip_skip > 1:
             shared.log.warning(f"Prompt parser SDXL unsupported: clip_skip={clip_skip}")

modules/sd_samplers_common.py

@@ -30,11 +30,11 @@ def single_sample_to_image(sample, approximation=None):
     elif approximation == 1:
         x_sample = sd_vae_approx.model()(sample.to(devices.device, devices.dtype).unsqueeze(0))[0].detach() * 0.5 + 0.5
         if shared.sd_model_type == "sdxl":
-            x_sample = x_sample[[2,1,0],:,:] #BGR to RGB
+            x_sample = x_sample[[2,1,0],:,:] # BGR to RGB
     elif approximation == 2:
         x_sample = sd_vae_approx.cheap_approximation(sample) * 0.5 + 0.5
         if shared.sd_model_type == "sdxl":
-            x_sample = x_sample[[2,1,0],:,:] #BGR to RGB
+            x_sample = x_sample[[2,1,0],:,:] # BGR to RGB
     elif approximation == 3:
         # x_sample = sample * 1.5
         # x_sample = sd_vae_taesd.model()(x_sample.to(devices.device, devices.dtype).unsqueeze(0))[0].detach()

modules/sd_samplers_compvis.py

@@ -121,10 +121,7 @@ class VanillaStableDiffusionSampler:
         return x, ts, cond, unconditional_conditioning
 
     def update_step(self, last_latent):
-        if self.mask is not None:
-            self.last_latent = self.init_latent * self.mask + self.nmask * last_latent
-        else:
-            self.last_latent = last_latent
+        self.last_latent = self.init_latent * self.mask + self.nmask * last_latent if self.mask is not None else last_latent
         sd_samplers_common.store_latent(self.last_latent)
         self.step += 1
         state.sampling_step = self.step

modules/sd_vae_approx.py

@@ -36,6 +36,7 @@ def model():
     global sd_vae_approx_model # pylint: disable=global-statement
 
     if sd_vae_approx_model is None:
+        from modules.shared import log
         model_path = os.path.join(paths.models_path, "VAE-approx", "model.pt")
         sd_vae_approx_model = VAEApprox()
         if not os.path.exists(model_path):
@@ -43,13 +44,13 @@ def model():
         sd_vae_approx_model.load_state_dict(torch.load(model_path, map_location='cpu' if devices.device.type != 'cuda' else None))
         sd_vae_approx_model.eval()
         sd_vae_approx_model.to(devices.device, devices.dtype)
+        log.info(f"Loaded VAE-approx model: {model_path}")
 
     return sd_vae_approx_model
 
 
 def cheap_approximation(sample):
     # https://discuss.huggingface.co/t/decoding-latents-to-rgb-without-upscaling/23204/2
-
     coefs = torch.tensor([
         [0.298, 0.207, 0.208],
         [0.187, 0.286, 0.173],

modules/taesd/sd_vae_taesd.py

@@ -28,10 +28,12 @@ def model(model_class = 'sd', model_type = 'decoder'):
         model_path = os.path.join(paths_internal.models_path, "TAESD", f"tae{model_class}_{model_type}.pth")
         download_model(model_path)
         if os.path.exists(model_path):
+            from modules.shared import log
             taesd_models[f'{model_class}-{model_type}'] = TAESD(decoder_path=model_path, encoder_path=None) if model_type == 'decoder' else TAESD(encoder_path=model_path, decoder_path=None)
             vae = taesd_models[f'{model_class}-{model_type}']
             vae.eval()
             vae.to(devices.device, devices.dtype_vae)
+            log.info(f"Loaded VAE-approx model: {model_path}")
         else:
             raise FileNotFoundError('TAESD model not found')
     if vae is None: