update dataset

dataset_infos.json

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:4ebbf9875e6989bb510fde0ed00fa3bab307a18a3d18ff26277bcaa9190d351e
-size 3249
+oid sha256:3e5b24f8456e9215167843b2858caee2bfbd4304f839e4050d88f1f1c092f15b
+size 3066

gtsrb.py

@@ -15,10 +15,7 @@
 
 # Lint as: python3
 """GTSRB: German Traffic Sign Recognition Benchmark."""
-
-import tarfile
 import csv
-import pandas as pd
 
 import datasets
 from datasets import Dataset, DatasetDict
@@ -76,7 +73,8 @@ _FEATURES = datasets.Features({
     "Roi.X2": datasets.Value("uint8"),
     "Roi.Y2": datasets.Value("uint8"),
     "ClassId": datasets.Value("uint8"),
-    "Path": datasets.Image("png"),
+    # "Path": datasets.Image("png"),
+    "Path": datasets.Value("string"),
 })
 
 _URL = "https://github.com/bazylhorsey/gtsrb/archive/refs/tags/0.0.0.tar.gz"
@@ -124,7 +122,7 @@ class GTSRB(datasets.GeneratorBasedBuilder):
         )
 
     def _split_generators(self, dl_manager):
-        files = dl_manager.download_and_extract(_URL)
+
         print("success")
         return [
             datasets.SplitGenerator(