Keep only 3 models to avoid RAM problems

app.py

@@ -5,7 +5,7 @@ import pandas as pd, numpy as np
 from transformers import CLIPProcessor, CLIPModel
 from st_clickable_images import clickable_images
 
-MODEL_NAMES = ["base-patch32", "base-patch16", "large-patch14", "large-patch14-336"]
+MODEL_NAMES = ["base-patch32", "base-patch16", "large-patch14-336"] #, "large-patch14", "large-patch14-336"] 
 
 
 @st.cache(allow_output_mutation=True)
@@ -165,10 +165,10 @@ def main():
     corpus = st.radio("", ["Unsplash", "Movies"])
 
     models_dict = {
-        "ViT-B/32 (quickest)": "base-patch32",
-        "ViT-B/16 (quick)": "base-patch16",
-        "ViT-L/14 (slow)": "large-patch14",
-        "ViT-L/14@336px (slowest)": "large-patch14-336",
+        "ViT-B/32 (small)": "base-patch32",
+        "ViT-B/16 (medium)": "base-patch16",
+        #"ViT-L/14 (slow)": "large-patch14",
+        "ViT-L/14@336px (large)": "large-patch14-336",
     }
 
     if "Comparison" in mode: