correct regression in loading of PaperCut and VoxelArt models (#1730)

This corrects a regression in loading of these models due to
a change of the embedding_manager parameter `num_vectors_per_token`

Fixes #1718

Author: lstein

configs/stable-diffusion/v1-inference.yaml

@@ -32,7 +32,7 @@ model:
         placeholder_strings: ["*"]
         initializer_words: ['sculpture']
         per_image_tokens: false
-        num_vectors_per_token: 8
+        num_vectors_per_token: 1
         progressive_words: False
 
     unet_config: