feat: graffiti-database.com palette

README.md

@@ -9,10 +9,10 @@ dataset_info:
     dtype: string
   splits:
   - name: train
-    num_bytes: 10454347970
+    num_bytes: 8373338923
     num_examples: 77478
-  download_size: 16111025975
-  dataset_size: 10454347970
+  download_size: 16124853637
+  dataset_size: 8373338923
 ---
 # Dataset Card for Graffiti
 

data/graffiti-database.com/metadata.jsonl

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:8aeece507dde6a9ea31c746226a4133ed03cf344cb5513918be1fb7018efb2d8
-size 13270223
+oid sha256:5da51677a29f6db9ce29b0e07368e9c2efd5e36e15d1342db420f5a2270275e0
+size 21153581

graffiti.py

@@ -119,11 +119,20 @@ class Graffiti(datasets.GeneratorBasedBuilder):
                 conditioning_file_obj = conditioning_file[1]
 
                 text = data["caption"]
+                if data["palette"] != None:
+                    text += ", in the colors "
+                    colors = []
+                    for color in data["palette"]:
+                        if color[1] in colors:
+                            continue
+                        colors.append(color[1])
+                    text += " and ".join(colors)
                 if data["artist"] != None:
                     # text += ", with text " + data["artist"]
-                    text += ", in the art of " + data["artist"]
+                    text += ", by " + data["artist"]
                 if data["city"] != None:
                     text += ", located in " + data["city"]
+
                 yield idx, {
                         "image": {"path": file_path, "bytes": file_obj.read()},
                         "conditioning_image": {"path": conditioning_file_path, "bytes": conditioning_file_obj.read()},