Update README.md

README.md

@@ -112,20 +112,19 @@ import datasets as ds
 import PIL.Image as Image
 import io
 
-# full data streaming
-commfor_train_stream = ds.load_dataset("OwensLab/CommunityForensics", split='Systematic+Manual', streaming=True)
+# steaming only the systematic set. Note that when streaming, you can only load specific splits
+commfor_sys_stream = ds.load_dataset("OwensLab/CommunityForensics", split='Systematic', streaming=True)
 
 # streaming only the evaluation set
 commfor_eval_stream = ds.load_dataset("OwensLab/CommunityForensics", split='PublicEval', streaming=True)
 
-# streaming only 10% of training data. Note that this does not contain the full set of models!
-commfor_train_stream_10p = ds.load_dataset("OwensLab/CommunityForensics", split='Systematic[:10%]+Manual[:10%]', streaming=True)
-
 # optionally shuffle the streaming dataset
-commfor_train_stream_10p = commfor_train_stream_10p.shuffle(seed=123, buffer_size=3000)
+commfor_sys_stream = commfor_sys_stream.shuffle(seed=123, buffer_size=3000)
 
 # usage example
-for i, data in enumerate(commfor_train_stream_10p):
+for i, data in enumerate(commfor_sys_stream):
+  if i>=10000: # use only first 10000 samples
+    break
   img, label = Image.open(io.BytesIO(data['image_data'])), data['label']
   ## Your operations here ##
   # e.g., img_torch = torchvision.transforms.functional.pil_to_tensor(img)