revise kernel

and add unused files

modules/processing.py

@@ -890,7 +890,7 @@ def process_images_inner(p: StableDiffusionProcessing) -> Processed:
             p.seeds = p.all_seeds[n * p.batch_size:(n + 1) * p.batch_size]
             p.subseeds = p.all_subseeds[n * p.batch_size:(n + 1) * p.batch_size]
 
-            latent_channels = getattr(shared.sd_model, 'latent_channels', opt_C)
+            latent_channels = shared.sd_model.forge_objects.vae.latent_channels
             p.rng = rng.ImageRNG((latent_channels, p.height // opt_f, p.width // opt_f), p.seeds, subseeds=p.subseeds, subseed_strength=p.subseed_strength, seed_resize_from_h=p.seed_resize_from_h, seed_resize_from_w=p.seed_resize_from_w)
 
             if p.scripts is not None:

modules/prompt_parser.py

@@ -318,17 +318,19 @@ def reconstruct_cond_batch(c: list[list[ScheduledPromptConditioning]], current_s
 
 
 def stack_conds(tensors):
-    # if prompts have wildly different lengths above the limit we'll get tensors of different shapes
-    # and won't be able to torch.stack them. So this fixes that.
-    token_count = max([x.shape[0] for x in tensors])
-    for i in range(len(tensors)):
-        if tensors[i].shape[0] != token_count:
-            last_vector = tensors[i][-1:]
-            last_vector_repeated = last_vector.repeat([token_count - tensors[i].shape[0], 1])
-            tensors[i] = torch.vstack([tensors[i], last_vector_repeated])
-
-    return torch.stack(tensors)
-
+    try:
+        result = torch.stack(tensors)
+    except:
+        # if prompts have wildly different lengths above the limit we'll get tensors of different shapes
+        # and won't be able to torch.stack them. So this fixes that.
+        token_count = max([x.shape[0] for x in tensors])
+        for i in range(len(tensors)):
+            if tensors[i].shape[0] != token_count:
+                last_vector = tensors[i][-1:]
+                last_vector_repeated = last_vector.repeat([token_count - tensors[i].shape[0], 1])
+                tensors[i] = torch.vstack([tensors[i], last_vector_repeated])
+        result = torch.stack(tensors)
+    return result
 
 
 def reconstruct_multicond_batch(c: MulticondLearnedConditioning, current_step):

modules/sd_vae_approx.py

@@ -58,7 +58,7 @@ def model():
             model_path = os.path.join(paths.models_path, "VAE-approx", model_name)
             download_model(model_path, 'https://github.com/AUTOMATIC1111/stable-diffusion-webui/releases/download/v1.0.0-pre/' + model_name)
 
-        loaded_model = VAEApprox(latent_channels=shared.sd_model.latent_channels)
+        loaded_model = VAEApprox(latent_channels=shared.sd_model.forge_objects.vae.latent_channels)
         loaded_model.load_state_dict(torch.load(model_path, map_location='cpu' if devices.device.type != 'cuda' else None))
         loaded_model.eval()
         loaded_model.to(devices.device, devices.dtype)