train.jsonl conditioning by texture and adjust dataset metadata to download texture instead of yaw angles

multiview-controlnet.py

@@ -5,7 +5,7 @@ import os
 
 _VERSION = datasets.Version("0.0.2")
 
-_DESCRIPTION = "Images rendered using DeepFashion3D meshes by varying yaw angle. Conditioning on the yaw angle naively encoded as a image"
+_DESCRIPTION = "Images rendered using DeepFashion3D meshes by varying yaw angle. Conditioning on the yaw angle naively encoded as a image. Also in the dataset conditioning by texture is possible"
 _HOMEPAGE = "TODO"
 _LICENSE = "TODO"
 _CITATION = "TODO"
@@ -32,7 +32,7 @@ IMAGES_URL = hf_hub_url(
 
 CONDITIONING_IMAGES_URL = hf_hub_url(
     "eduardoprea44/multiview-controlnet",
-    filename="yaw_angle_as_img.zip",
+    filename="texture.zip",
     repo_type="dataset",
 )