Use repo_type instead of deprecated datasets repo IDs (#19202)

* Use repo_type instead of deprecated datasets repo IDs

* Add missing one in doc

docs/source/en/tasks/semantic_segmentation.mdx

@@ -67,9 +67,9 @@ You'll also want to create a dictionary that maps a label id to a label class wh
 >>> import json
 >>> from huggingface_hub import cached_download, hf_hub_url
 
->>> repo_id = "datasets/huggingface/label-files"
+>>> repo_id = "huggingface/label-files"
 >>> filename = "ade20k-id2label.json"
->>> id2label = json.load(open(cached_download(hf_hub_url(repo_id, filename)), "r"))
+>>> id2label = json.load(open(cached_download(hf_hub_url(repo_id, filename, repo_type="dataset")), "r"))
 >>> id2label = {int(k): v for k, v in id2label.items()}
 >>> label2id = {v: k for k, v in id2label.items()}
 >>> num_labels = len(id2label)

examples/pytorch/semantic-segmentation/run_semantic_segmentation.py

@@ -327,12 +327,12 @@ def main():
     # Prepare label mappings.
     # We'll include these in the model's config to get human readable labels in the Inference API.
     if data_args.dataset_name == "scene_parse_150":
-        repo_id = "datasets/huggingface/label-files"
+        repo_id = "huggingface/label-files"
         filename = "ade20k-id2label.json"
     else:
-        repo_id = f"datasets/{data_args.dataset_name}"
+        repo_id = data_args.dataset_name
         filename = "id2label.json"
-    id2label = json.load(open(hf_hub_download(repo_id, filename), "r"))
+    id2label = json.load(open(hf_hub_download(repo_id, filename, repo_type="dataset"), "r"))
     id2label = {int(k): v for k, v in id2label.items()}
     label2id = {v: str(k) for k, v in id2label.items()}
 

examples/pytorch/semantic-segmentation/run_semantic_segmentation_no_trainer.py

@@ -387,12 +387,12 @@ def main():
     # Prepare label mappings.
     # We'll include these in the model's config to get human readable labels in the Inference API.
     if args.dataset_name == "scene_parse_150":
-        repo_id = "datasets/huggingface/label-files"
+        repo_id = "huggingface/label-files"
         filename = "ade20k-id2label.json"
     else:
-        repo_id = f"datasets/{args.dataset_name}"
+        repo_id = args.dataset_name
         filename = "id2label.json"
-    id2label = json.load(open(hf_hub_download(repo_id, filename), "r"))
+    id2label = json.load(open(hf_hub_download(repo_id, filename, repo_type="dataset"), "r"))
     id2label = {int(k): v for k, v in id2label.items()}
     label2id = {v: k for k, v in id2label.items()}
 

src/transformers/models/beit/convert_beit_unilm_to_pytorch.py

@@ -176,7 +176,7 @@ def convert_beit_checkpoint(checkpoint_url, pytorch_dump_folder_path):
     config = BeitConfig()
     has_lm_head = False
     is_semantic = False
-    repo_id = "datasets/huggingface/label-files"
+    repo_id = "huggingface/label-files"
     # set config parameters based on URL
     if checkpoint_url[-9:-4] == "pt22k":
         # masked image modeling
@@ -188,7 +188,7 @@ def convert_beit_checkpoint(checkpoint_url, pytorch_dump_folder_path):
         config.use_relative_position_bias = True
         config.num_labels = 21841
         filename = "imagenet-22k-id2label.json"
-        id2label = json.load(open(hf_hub_download(repo_id, filename), "r"))
+        id2label = json.load(open(hf_hub_download(repo_id, filename, repo_type="dataset"), "r"))
         id2label = {int(k): v for k, v in id2label.items()}
         # this dataset contains 21843 labels but the model only has 21841
         # we delete the classes as mentioned in https://github.com/google-research/big_transfer/issues/18
@@ -201,7 +201,7 @@ def convert_beit_checkpoint(checkpoint_url, pytorch_dump_folder_path):
         config.use_relative_position_bias = True
         config.num_labels = 1000
         filename = "imagenet-1k-id2label.json"
-        id2label = json.load(open(hf_hub_download(repo_id, filename), "r"))
+        id2label = json.load(open(hf_hub_download(repo_id, filename, repo_type="dataset"), "r"))
         id2label = {int(k): v for k, v in id2label.items()}
         config.id2label = id2label
         config.label2id = {v: k for k, v in id2label.items()}
@@ -214,7 +214,7 @@ def convert_beit_checkpoint(checkpoint_url, pytorch_dump_folder_path):
         config.use_relative_position_bias = True
         config.num_labels = 150
         filename = "ade20k-id2label.json"
-        id2label = json.load(open(hf_hub_download(repo_id, filename), "r"))
+        id2label = json.load(open(hf_hub_download(repo_id, filename, repo_type="dataset"), "r"))
         id2label = {int(k): v for k, v in id2label.items()}
         config.id2label = id2label
         config.label2id = {v: k for k, v in id2label.items()}

src/transformers/models/conditional_detr/convert_conditional_detr_original_pytorch_checkpoint_to_pytorch.py

@@ -237,9 +237,9 @@ def convert_conditional_detr_checkpoint(model_name, pytorch_dump_folder_path):
         config.num_labels = 250
     else:
         config.num_labels = 91
-        repo_id = "datasets/huggingface/label-files"
+        repo_id = "huggingface/label-files"
         filename = "coco-detection-id2label.json"
-        id2label = json.load(open(hf_hub_download(repo_id, filename), "r"))
+        id2label = json.load(open(hf_hub_download(repo_id, filename, repo_type="dataset"), "r"))
         id2label = {int(k): v for k, v in id2label.items()}
         config.id2label = id2label
         config.label2id = {v: k for k, v in id2label.items()}

src/transformers/models/convnext/convert_convnext_to_pytorch.py

@@ -62,9 +62,9 @@ def get_convnext_config(checkpoint_url):
         filename = "imagenet-22k-id2label.json"
         expected_shape = (1, 21841)
 
-    repo_id = "datasets/huggingface/label-files"
+    repo_id = "huggingface/label-files"
     config.num_labels = num_labels
-    id2label = json.load(open(hf_hub_download(repo_id, filename), "r"))
+    id2label = json.load(open(hf_hub_download(repo_id, filename, repo_type="dataset"), "r"))
     id2label = {int(k): v for k, v in id2label.items()}
     if "1k" not in checkpoint_url:
         # this dataset contains 21843 labels but the model only has 21841

src/transformers/models/cvt/convert_cvt_original_pytorch_checkpoint_to_pytorch.py

@@ -282,9 +282,9 @@ def convert_cvt_checkpoint(cvt_model, image_size, cvt_file_name, pytorch_dump_fo
     img_labels_file = "imagenet-1k-id2label.json"
     num_labels = 1000
 
-    repo_id = "datasets/huggingface/label-files"
+    repo_id = "huggingface/label-files"
     num_labels = num_labels
-    id2label = json.load(open(cached_download(hf_hub_url(repo_id, img_labels_file)), "r"))
+    id2label = json.load(open(cached_download(hf_hub_url(repo_id, img_labels_file, repo_type="dataset")), "r"))
     id2label = {int(k): v for k, v in id2label.items()}
 
     id2label = id2label

src/transformers/models/data2vec/convert_data2vec_vision_original_pytorch_checkpoint_to_pytorch.py

@@ -282,9 +282,9 @@ def main():
         config.use_mean_pooling = True
         config.num_labels = 1000
 
-        repo_id = "datasets/huggingface/label-files"
+        repo_id = "huggingface/label-files"
         filename = "imagenet-1k-id2label.json"
-        id2label = json.load(open(hf_hub_download(repo_id, filename), "r"))
+        id2label = json.load(open(hf_hub_download(repo_id, filename, repo_type="dataset"), "r"))
         id2label = {int(k): v for k, v in id2label.items()}
         config.id2label = id2label
         config.label2id = {v: k for k, v in id2label.items()}

src/transformers/models/deformable_detr/convert_deformable_detr_to_pytorch.py

@@ -108,9 +108,9 @@ def convert_deformable_detr_checkpoint(
     config.two_stage = two_stage
     # set labels
     config.num_labels = 91
-    repo_id = "datasets/huggingface/label-files"
+    repo_id = "huggingface/label-files"
     filename = "coco-detection-id2label.json"
-    id2label = json.load(open(cached_download(hf_hub_url(repo_id, filename)), "r"))
+    id2label = json.load(open(cached_download(hf_hub_url(repo_id, filename, repo_type="dataset")), "r"))
     id2label = {int(k): v for k, v in id2label.items()}
     config.id2label = id2label
     config.label2id = {v: k for k, v in id2label.items()}

src/transformers/models/deit/convert_deit_timm_to_pytorch.py

@@ -140,9 +140,9 @@ def convert_deit_checkpoint(deit_name, pytorch_dump_folder_path):
     base_model = False
     # dataset (fine-tuned on ImageNet 2012), patch_size and image_size
     config.num_labels = 1000
-    repo_id = "datasets/huggingface/label-files"
+    repo_id = "huggingface/label-files"
     filename = "imagenet-1k-id2label.json"
-    id2label = json.load(open(hf_hub_download(repo_id, filename), "r"))
+    id2label = json.load(open(hf_hub_download(repo_id, filename, repo_type="dataset"), "r"))
     id2label = {int(k): v for k, v in id2label.items()}
     config.id2label = id2label
     config.label2id = {v: k for k, v in id2label.items()}

src/transformers/models/detr/convert_detr_original_pytorch_checkpoint_to_pytorch.py

@@ -194,9 +194,9 @@ def convert_detr_checkpoint(model_name, pytorch_dump_folder_path):
         config.num_labels = 250
     else:
         config.num_labels = 91
-        repo_id = "datasets/huggingface/label-files"
+        repo_id = "huggingface/label-files"
         filename = "coco-detection-id2label.json"
-        id2label = json.load(open(hf_hub_download(repo_id, filename), "r"))
+        id2label = json.load(open(hf_hub_download(repo_id, filename, repo_type="dataset"), "r"))
         id2label = {int(k): v for k, v in id2label.items()}
         config.id2label = id2label
         config.label2id = {v: k for k, v in id2label.items()}

src/transformers/models/dit/convert_dit_unilm_to_pytorch.py

@@ -149,9 +149,9 @@ def convert_dit_checkpoint(checkpoint_url, pytorch_dump_folder_path, push_to_hub
     # labels
     if "rvlcdip" in checkpoint_url:
         config.num_labels = 16
-        repo_id = "datasets/huggingface/label-files"
+        repo_id = "huggingface/label-files"
         filename = "rvlcdip-id2label.json"
-        id2label = json.load(open(hf_hub_download(repo_id, filename), "r"))
+        id2label = json.load(open(hf_hub_download(repo_id, filename, repo_type="dataset"), "r"))
         id2label = {int(k): v for k, v in id2label.items()}
         config.id2label = id2label
         config.label2id = {v: k for k, v in id2label.items()}

src/transformers/models/dpt/convert_dpt_to_pytorch.py

@@ -48,9 +48,9 @@ def get_dpt_config(checkpoint_url):
         config.use_batch_norm_in_fusion_residual = True
 
         config.num_labels = 150
-        repo_id = "datasets/huggingface/label-files"
+        repo_id = "huggingface/label-files"
         filename = "ade20k-id2label.json"
-        id2label = json.load(open(cached_download(hf_hub_url(repo_id, filename)), "r"))
+        id2label = json.load(open(cached_download(hf_hub_url(repo_id, filename, repo_type="dataset")), "r"))
         id2label = {int(k): v for k, v in id2label.items()}
         config.id2label = id2label
         config.label2id = {v: k for k, v in id2label.items()}

src/transformers/models/levit/convert_levit_timm_to_pytorch.py

@@ -85,9 +85,9 @@ def convert_weights_and_push(save_directory: Path, model_name: str = None, push_
     num_labels = 1000
     expected_shape = (1, num_labels)
 
-    repo_id = "datasets/huggingface/label-files"
+    repo_id = "huggingface/label-files"
     num_labels = num_labels
-    id2label = json.load(open(hf_hub_download(repo_id, filename), "r"))
+    id2label = json.load(open(hf_hub_download(repo_id, filename, repo_type="dataset"), "r"))
     id2label = {int(k): v for k, v in id2label.items()}
 
     id2label = id2label

src/transformers/models/mobilevit/convert_mlcvnets_to_pytorch.py

@@ -62,8 +62,8 @@ def get_mobilevit_config(mobilevit_name):
         config.num_labels = 1000
         filename = "imagenet-1k-id2label.json"
 
-    repo_id = "datasets/huggingface/label-files"
-    id2label = json.load(open(hf_hub_download(repo_id, filename), "r"))
+    repo_id = "huggingface/label-files"
+    id2label = json.load(open(hf_hub_download(repo_id, filename, repo_type="dataset"), "r"))
     id2label = {int(k): v for k, v in id2label.items()}
     config.id2label = id2label
     config.label2id = {v: k for k, v in id2label.items()}

src/transformers/models/perceiver/convert_perceiver_haiku_to_pytorch.py

@@ -300,7 +300,7 @@ def convert_perceiver_checkpoint(pickle_file, pytorch_dump_folder_path, architec
     # load HuggingFace model
     config = PerceiverConfig()
     subsampling = None
-    repo_id = "datasets/huggingface/label-files"
+    repo_id = "huggingface/label-files"
     if architecture == "MLM":
         config.qk_channels = 8 * 32
         config.v_channels = 1280
@@ -318,7 +318,7 @@ def convert_perceiver_checkpoint(pickle_file, pytorch_dump_folder_path, architec
         # set labels
         config.num_labels = 1000
         filename = "imagenet-1k-id2label.json"
-        id2label = json.load(open(hf_hub_download(repo_id, filename), "r"))
+        id2label = json.load(open(hf_hub_download(repo_id, filename, repo_type="dataset"), "r"))
         id2label = {int(k): v for k, v in id2label.items()}
         config.id2label = id2label
         config.label2id = {v: k for k, v in id2label.items()}
@@ -367,7 +367,7 @@ def convert_perceiver_checkpoint(pickle_file, pytorch_dump_folder_path, architec
         model = PerceiverForMultimodalAutoencoding(config)
         # set labels
         filename = "kinetics700-id2label.json"
-        id2label = json.load(open(hf_hub_download(repo_id, filename), "r"))
+        id2label = json.load(open(hf_hub_download(repo_id, filename, repo_type="dataset"), "r"))
         id2label = {int(k): v for k, v in id2label.items()}
         config.id2label = id2label
         config.label2id = {v: k for k, v in id2label.items()}

src/transformers/models/poolformer/convert_poolformer_original_to_pytorch.py

@@ -99,14 +99,14 @@ def convert_poolformer_checkpoint(model_name, checkpoint_path, pytorch_dump_fold
     config = PoolFormerConfig()
 
     # set attributes based on model_name
-    repo_id = "datasets/huggingface/label-files"
+    repo_id = "huggingface/label-files"
     size = model_name[-3:]
     config.num_labels = 1000
     filename = "imagenet-1k-id2label.json"
     expected_shape = (1, 1000)
 
     # set config attributes
-    id2label = json.load(open(hf_hub_download(repo_id, filename), "r"))
+    id2label = json.load(open(hf_hub_download(repo_id, filename, repo_type="dataset"), "r"))
     id2label = {int(k): v for k, v in id2label.items()}
     config.id2label = id2label
     config.label2id = {v: k for k, v in id2label.items()}

src/transformers/models/regnet/convert_regnet_seer_10b_to_pytorch.py

@@ -163,9 +163,9 @@ def convert_weights_and_push(save_directory: Path, model_name: str = None, push_
     filename = "imagenet-1k-id2label.json"
     num_labels = 1000
 
-    repo_id = "datasets/huggingface/label-files"
+    repo_id = "huggingface/label-files"
     num_labels = num_labels
-    id2label = json.load(open(cached_download(hf_hub_url(repo_id, filename)), "r"))
+    id2label = json.load(open(cached_download(hf_hub_url(repo_id, filename, repo_type="dataset")), "r"))
     id2label = {int(k): v for k, v in id2label.items()}
 
     id2label = id2label

src/transformers/models/regnet/convert_regnet_to_pytorch.py

@@ -224,9 +224,9 @@ def convert_weights_and_push(save_directory: Path, model_name: str = None, push_
     num_labels = 1000
     expected_shape = (1, num_labels)
 
-    repo_id = "datasets/huggingface/label-files"
+    repo_id = "huggingface/label-files"
     num_labels = num_labels
-    id2label = json.load(open(cached_download(hf_hub_url(repo_id, filename)), "r"))
+    id2label = json.load(open(cached_download(hf_hub_url(repo_id, filename, repo_type="dataset")), "r"))
     id2label = {int(k): v for k, v in id2label.items()}
 
     id2label = id2label

src/transformers/models/resnet/convert_resnet_to_pytorch.py

@@ -128,9 +128,9 @@ def convert_weights_and_push(save_directory: Path, model_name: str = None, push_
     num_labels = 1000
     expected_shape = (1, num_labels)
 
-    repo_id = "datasets/huggingface/label-files"
+    repo_id = "huggingface/label-files"
     num_labels = num_labels
-    id2label = json.load(open(hf_hub_download(repo_id, filename), "r"))
+    id2label = json.load(open(hf_hub_download(repo_id, filename, repo_type="dataset"), "r"))
     id2label = {int(k): v for k, v in id2label.items()}
 
     id2label = id2label