fix(coda): replace incorrect dataset checksums

Replace incorrect checksums in `dataset_info.json`. It seems there was a
bug when creating the original checksums, the number of bytes per file
is correct, and the data remains unchanged. The exact cause remains
unclear, but the files can be regenerated and verified from the original
annotations and templates (see https://github.com/nala-cub/coda for
details).

Fixes nala-cub/coda#1.

GitOrigin-RevId: a475c67310ce7bc52912556cb0430dee815c5442

coda.py

@@ -57,7 +57,7 @@ _URLs = {
 
 class Coda(datasets.GeneratorBasedBuilder):
 
-  VERSION = datasets.Version('1.0.0')
+  VERSION = datasets.Version('1.0.1')
 
   # TODO(corypaik): add object and annotation configs.
 

dataset_infos.json

@@ -1 +1 @@
-{"default": {"description": "*The Color Dataset* (CoDa) is a probing dataset to evaluate the representation of visual properties in language models. CoDa consists of color distributions for 521 common objects, which are split into 3 groups: Single, Multi, and Any.\n", "citation": "@misc{paik2021world,\n      title={The World of an Octopus: How Reporting Bias Influences a Language Model's Perception of Color},\n      author={Cory Paik and St\u00e9phane Aroca-Ouellette and Alessandro Roncone and Katharina Kann},\n      year={2021},\n      eprint={2110.08182},\n      archivePrefix={arXiv},\n      primaryClass={cs.CL}\n}\n", "homepage": "https://github.com/nala-cub/coda", "license": "Apache 2.0", "features": {"class_id": {"dtype": "string", "id": null, "_type": "Value"}, "display_name": {"dtype": "string", "id": null, "_type": "Value"}, "ngram": {"dtype": "string", "id": null, "_type": "Value"}, "label": {"feature": {"dtype": "float32", "id": null, "_type": "Value"}, "length": -1, "id": null, "_type": "Sequence"}, "object_group": {"num_classes": 3, "names": ["Single", "Multi", "Any"], "names_file": null, "id": null, "_type": "ClassLabel"}, "text": {"dtype": "string", "id": null, "_type": "Value"}, "template_group": {"num_classes": 2, "names": ["clip-imagenet", "text-masked"], "names_file": null, "id": null, "_type": "ClassLabel"}, "template_idx": {"dtype": "int32", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "task_templates": null, "builder_name": "coda", "config_name": "default", "version": {"version_str": "1.0.0", "description": null, "major": 1, "minor": 0, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 852459, "num_examples": 6164, "dataset_name": "coda"}, "validation": {"name": "validation", "num_bytes": 285136, "num_examples": 2060, "dataset_name": "coda"}, "test": {"name": "test", "num_bytes": 292745, "num_examples": 2133, "dataset_name": "coda"}}, "download_checksums": {"https://huggingface.co/datasets/corypaik/coda/resolve/main/data/default_train.jsonl": {"num_bytes": 1709553, "checksum": "5157bf47368ea33056b8d1ecaf688ea713ff0d6fea22002918fa6f70a5240436"}, "https://huggingface.co/datasets/corypaik/coda/resolve/main/data/default_validation.jsonl": {"num_bytes": 574026, "checksum": "76557b184e18cd0d4edebe13e0ad86e1e685b60a2d53bee2b554acf27f6d2e55"}, "https://huggingface.co/datasets/corypaik/coda/resolve/main/data/default_test.jsonl": {"num_bytes": 589396, "checksum": "24fa65a27a9dd6fa4f9b8e1dfeca0dbb26915c26c66b8392c0be4c5929c76a19"}}, "download_size": 2872975, "post_processing_size": null, "dataset_size": 1430340, "size_in_bytes": 4303315}}
+{"default": {"description": "*The Color Dataset* (CoDa) is a probing dataset to evaluate the representation of visual properties in language models. CoDa consists of color distributions for 521 common objects, which are split into 3 groups: Single, Multi, and Any.\n", "citation": "@misc{paik2021world,\n      title={The World of an Octopus: How Reporting Bias Influences a Language Model's Perception of Color},\n      author={Cory Paik and St\u00e9phane Aroca-Ouellette and Alessandro Roncone and Katharina Kann},\n      year={2021},\n      eprint={2110.08182},\n      archivePrefix={arXiv},\n      primaryClass={cs.CL}\n}\n", "homepage": "https://github.com/nala-cub/coda", "license": "Apache 2.0", "features": {"class_id": {"dtype": "string", "id": null, "_type": "Value"}, "display_name": {"dtype": "string", "id": null, "_type": "Value"}, "ngram": {"dtype": "string", "id": null, "_type": "Value"}, "label": {"feature": {"dtype": "float32", "id": null, "_type": "Value"}, "length": -1, "id": null, "_type": "Sequence"}, "object_group": {"num_classes": 3, "names": ["Single", "Multi", "Any"], "names_file": null, "id": null, "_type": "ClassLabel"}, "text": {"dtype": "string", "id": null, "_type": "Value"}, "template_group": {"num_classes": 2, "names": ["clip-imagenet", "text-masked"], "names_file": null, "id": null, "_type": "ClassLabel"}, "template_idx": {"dtype": "int32", "id": null, "_type": "Value"}}, "post_processed": null, "supervised_keys": null, "task_templates": null, "builder_name": "coda", "config_name": "default", "version": {"version_str": "1.0.1", "description": null, "major": 1, "minor": 0, "patch": 1}, "splits": {"train": {"name": "train", "num_bytes": 852459, "num_examples": 6164, "dataset_name": "coda"}, "validation": {"name": "validation", "num_bytes": 285136, "num_examples": 2060, "dataset_name": "coda"}, "test": {"name": "test", "num_bytes": 292745, "num_examples": 2133, "dataset_name": "coda"}}, "download_checksums": {"https://huggingface.co/datasets/corypaik/coda/resolve/main/data/default_train.jsonl": {"num_bytes": 1709553, "checksum": "72ed83901c3b38d72651758fc237610782e523268c7e9f17a803de88f533caf1"}, "https://huggingface.co/datasets/corypaik/coda/resolve/main/data/default_validation.jsonl": {"num_bytes": 574026, "checksum": "6a387349f66c7d04b3d3059e4d1e7d52aea16e251a53e80f18d495d1d0c5fc1e"}, "https://huggingface.co/datasets/corypaik/coda/resolve/main/data/default_test.jsonl": {"num_bytes": 589396, "checksum": "2e1271cb99f4d2989a32bcc9a4d39f01680c15523476dbf01f6cd19e3f14f8aa"}}, "download_size": 2872975, "post_processing_size": null, "dataset_size": 1430340, "size_in_bytes": 4303315}}