re added projection layer to suit the x embedder input

app.py

@@ -1,8 +1,8 @@
-import spaces
 import gradio as gr
 import numpy as np
 import random
 import torch
+import torch.nn as nn
 from PIL import Image
 from torchvision import transforms
 from diffusers import DiffusionPipeline
@@ -19,6 +19,9 @@ pipe.enable_model_cpu_offload()
 pipe.vae.enable_slicing()
 pipe.vae.enable_tiling()
 
+# Add a projection layer to match x_embedder input
+projection = nn.Linear(16, 64).to(device).to(dtype)
+
 def preprocess_image(image, image_size):
     preprocess = transforms.Compose([
         transforms.Resize((image_size, image_size), interpolation=transforms.InterpolationMode.LANCZOS),
@@ -40,6 +43,10 @@ def process_latents(latents, height, width):
     latents = latents.permute(0, 2, 3, 1).reshape(1, -1, latents.shape[1])
     print(f"Reshaped latent shape: {latents.shape}")
     
+    # Project latents from 16 to 64 dimensions
+    latents = projection(latents)
+    print(f"Projected latent shape: {latents.shape}")
+    
     return latents
 
 @spaces.GPU()