add paper

app.py

@@ -183,7 +183,7 @@ downscaled_outputs = default_outputs
 example_items = downscaled_images[:3] + downscaled_outputs[:3]
 
 
-def run_alignedthreemodelattnnodes(images, model, batch_size=1):
+def run_alignedthreemodelattnnodes(images, model, batch_size=16):
     
     use_cuda = torch.cuda.is_available() 
     device = torch.device("cuda" if use_cuda else "cpu")
@@ -535,17 +535,17 @@ def run_fn(
     }
     # print(kwargs)
     num_images = len(images)
-    if num_images > 100:
+    if num_images >= 100:
         return super_duper_long_run(model, images, **kwargs)
     if 'diffusion' in model_name.lower():
         return super_duper_long_run(model, images, **kwargs)
     if recursion:
         return longer_run(model, images, **kwargs)
-    if num_images > 50:
+    if num_images >= 50:
         return longer_run(model, images, **kwargs)
     if old_school_ncut:
         return longer_run(model, images, **kwargs)
-    if num_images > 10:
+    if num_images >= 10:
         return long_run(model, images, **kwargs)
     if embedding_method == "UMAP":
         if perplexity >= 250 or num_sample_tsne >= 500:
@@ -1020,6 +1020,8 @@ with demo:
         gr.Markdown('---')
         gr.Markdown('**Features are aligned across models and layers.** A linear alignment transform is trained for each model/layer, learning signal comes from 1) fMRI brain activation and 2) segmentation preserving eigen-constraints.')
         gr.Markdown('NCUT is computed on the concatenated graph of all models, layers, and images. Color is **aligned** across all models and layers.')
+        gr.Markdown('')
+        gr.Markdown("To see a good pattern, you will need to load 100 images. Running out of HuggingFace GPU Quota? Try [Demo](https://ncut-pytorch.readthedocs.io/en/latest/demo/) hosted at UPenn")
         gr.Markdown('---')
         with gr.Row():
             with gr.Column(scale=5, min_width=200):