Datasets:

---

ssupinma

/

github-issues

like 0

[ "The docs for this PR live [here](https://moon-ci-docs.huggingface.co/docs/datasets/pr_6839). All of your documentation changes will be reflected on that endpoint. The docs are available until 30 days after the last update.", "<details>\n<summary>Show benchmarks</summary>\n\nPyArrow==8.0.0\n\n<details>\n<summary>Show updated benchmarks!</summary>\n\n### Benchmark: benchmark_array_xd.json\n\n| metric | read_batch_formatted_as_numpy after write_array2d | read_batch_formatted_as_numpy after write_flattened_sequence | read_batch_formatted_as_numpy after write_nested_sequence | read_batch_unformated after write_array2d | read_batch_unformated after write_flattened_sequence | read_batch_unformated after write_nested_sequence | read_col_formatted_as_numpy after write_array2d | read_col_formatted_as_numpy after write_flattened_sequence | read_col_formatted_as_numpy after write_nested_sequence | read_col_unformated after write_array2d | read_col_unformated after write_flattened_sequence | read_col_unformated after write_nested_sequence | read_formatted_as_numpy after write_array2d | read_formatted_as_numpy after write_flattened_sequence | read_formatted_as_numpy after write_nested_sequence | read_unformated after write_array2d | read_unformated after write_flattened_sequence | read_unformated after write_nested_sequence | write_array2d | write_flattened_sequence | write_nested_sequence |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.005311 / 0.011353 (-0.006042) | 0.003691 / 0.011008 (-0.007317) | 0.063714 / 0.038508 (0.025206) | 0.030875 / 0.023109 (0.007766) | 0.251210 / 0.275898 (-0.024688) | 0.280539 / 0.323480 (-0.042941) | 0.004262 / 0.007986 (-0.003724) | 0.002723 / 0.004328 (-0.001606) | 0.049487 / 0.004250 (0.045237) | 0.045655 / 0.037052 (0.008603) | 0.264399 / 0.258489 (0.005910) | 0.306613 / 0.293841 (0.012772) | 0.028513 / 0.128546 (-0.100033) | 0.010726 / 0.075646 (-0.064921) | 0.210601 / 0.419271 (-0.208670) | 0.036918 / 0.043533 (-0.006614) | 0.257872 / 0.255139 (0.002733) | 0.278951 / 0.283200 (-0.004249) | 0.017900 / 0.141683 (-0.123783) | 1.096749 / 1.452155 (-0.355406) | 1.152603 / 1.492716 (-0.340113) |\n\n### Benchmark: benchmark_getitem\\_100B.json\n\n| metric | get_batch_of\\_1024\\_random_rows | get_batch_of\\_1024\\_rows | get_first_row | get_last_row |\n|--------|---|---|---|---|\n| new / old (diff) | 0.095193 / 0.018006 (0.077187) | 0.303919 / 0.000490 (0.303429) | 0.000226 / 0.000200 (0.000026) | 0.000052 / 0.000054 (-0.000002) |\n\n### Benchmark: benchmark_indices_mapping.json\n\n| metric | select | shard | shuffle | sort | train_test_split |\n|--------|---|---|---|---|---|\n| new / old (diff) | 0.018558 / 0.037411 (-0.018853) | 0.061106 / 0.014526 (0.046580) | 0.076233 / 0.176557 (-0.100323) | 0.122402 / 0.737135 (-0.614734) | 0.075579 / 0.296338 (-0.220760) |\n\n### Benchmark: benchmark_iterating.json\n\n| metric | read 5000 | read 50000 | read_batch 50000 10 | read_batch 50000 100 | read_batch 50000 1000 | read_formatted numpy 5000 | read_formatted pandas 5000 | read_formatted tensorflow 5000 | read_formatted torch 5000 | read_formatted_batch numpy 5000 10 | read_formatted_batch numpy 5000 1000 | shuffled read 5000 | shuffled read 50000 | shuffled read_batch 50000 10 | shuffled read_batch 50000 100 | shuffled read_batch 50000 1000 | shuffled read_formatted numpy 5000 | shuffled read_formatted_batch numpy 5000 10 | shuffled read_formatted_batch numpy 5000 1000 |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.283586 / 0.215209 (0.068377) | 2.766179 / 2.077655 (0.688524) | 1.481069 / 1.504120 (-0.023051) | 1.355004 / 1.541195 (-0.186191) | 1.392940 / 1.468490 (-0.075550) | 0.578878 / 4.584777 (-4.005899) | 2.432890 / 3.745712 (-1.312822) | 2.837912 / 5.269862 (-2.431949) | 1.762803 / 4.565676 (-2.802873) | 0.063339 / 0.424275 (-0.360937) | 0.005392 / 0.007607 (-0.002215) | 0.340271 / 0.226044 (0.114227) | 3.388371 / 2.268929 (1.119443) | 1.862622 / 55.444624 (-53.582002) | 1.543209 / 6.876477 (-5.333268) | 1.569858 / 2.142072 (-0.572215) | 0.651487 / 4.805227 (-4.153740) | 0.119048 / 6.500664 (-6.381616) | 0.042309 / 0.075469 (-0.033160) |\n\n### Benchmark: benchmark_map_filter.json\n\n| metric | filter | map fast-tokenizer batched | map identity | map identity batched | map no-op batched | map no-op batched numpy | map no-op batched pandas | map no-op batched pytorch | map no-op batched tensorflow |\n|--------|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.991161 / 1.841788 (-0.850627) | 11.778857 / 8.074308 (3.704549) | 9.586019 / 10.191392 (-0.605373) | 0.148093 / 0.680424 (-0.532331) | 0.014301 / 0.534201 (-0.519900) | 0.287983 / 0.579283 (-0.291301) | 0.266070 / 0.434364 (-0.168293) | 0.328261 / 0.540337 (-0.212076) | 0.417908 / 1.386936 (-0.969028) |\n\n</details>\nPyArrow==latest\n\n<details>\n<summary>Show updated benchmarks!</summary>\n\n### Benchmark: benchmark_array_xd.json\n\n| metric | read_batch_formatted_as_numpy after write_array2d | read_batch_formatted_as_numpy after write_flattened_sequence | read_batch_formatted_as_numpy after write_nested_sequence | read_batch_unformated after write_array2d | read_batch_unformated after write_flattened_sequence | read_batch_unformated after write_nested_sequence | read_col_formatted_as_numpy after write_array2d | read_col_formatted_as_numpy after write_flattened_sequence | read_col_formatted_as_numpy after write_nested_sequence | read_col_unformated after write_array2d | read_col_unformated after write_flattened_sequence | read_col_unformated after write_nested_sequence | read_formatted_as_numpy after write_array2d | read_formatted_as_numpy after write_flattened_sequence | read_formatted_as_numpy after write_nested_sequence | read_unformated after write_array2d | read_unformated after write_flattened_sequence | read_unformated after write_nested_sequence | write_array2d | write_flattened_sequence | write_nested_sequence |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.005252 / 0.011353 (-0.006100) | 0.003740 / 0.011008 (-0.007268) | 0.049622 / 0.038508 (0.011114) | 0.030040 / 0.023109 (0.006931) | 0.262224 / 0.275898 (-0.013674) | 0.312216 / 0.323480 (-0.011264) | 0.004213 / 0.007986 (-0.003773) | 0.002737 / 0.004328 (-0.001592) | 0.049159 / 0.004250 (0.044908) | 0.041060 / 0.037052 (0.004008) | 0.275826 / 0.258489 (0.017337) | 0.301879 / 0.293841 (0.008038) | 0.029364 / 0.128546 (-0.099182) | 0.010453 / 0.075646 (-0.065193) | 0.058095 / 0.419271 (-0.361176) | 0.032898 / 0.043533 (-0.010635) | 0.263876 / 0.255139 (0.008737) | 0.281686 / 0.283200 (-0.001514) | 0.018711 / 0.141683 (-0.122971) | 1.126056 / 1.452155 (-0.326098) | 1.185125 / 1.492716 (-0.307591) |\n\n### Benchmark: benchmark_getitem\\_100B.json\n\n| metric | get_batch_of\\_1024\\_random_rows | get_batch_of\\_1024\\_rows | get_first_row | get_last_row |\n|--------|---|---|---|---|\n| new / old (diff) | 0.094153 / 0.018006 (0.076147) | 0.300719 / 0.000490 (0.300229) | 0.000207 / 0.000200 (0.000007) | 0.000048 / 0.000054 (-0.000006) |\n\n### Benchmark: benchmark_indices_mapping.json\n\n| metric | select | shard | shuffle | sort | train_test_split |\n|--------|---|---|---|---|---|\n| new / old (diff) | 0.022610 / 0.037411 (-0.014801) | 0.075502 / 0.014526 (0.060977) | 0.088858 / 0.176557 (-0.087699) | 0.129421 / 0.737135 (-0.607714) | 0.089331 / 0.296338 (-0.207007) |\n\n### Benchmark: benchmark_iterating.json\n\n| metric | read 5000 | read 50000 | read_batch 50000 10 | read_batch 50000 100 | read_batch 50000 1000 | read_formatted numpy 5000 | read_formatted pandas 5000 | read_formatted tensorflow 5000 | read_formatted torch 5000 | read_formatted_batch numpy 5000 10 | read_formatted_batch numpy 5000 1000 | shuffled read 5000 | shuffled read 50000 | shuffled read_batch 50000 10 | shuffled read_batch 50000 100 | shuffled read_batch 50000 1000 | shuffled read_formatted numpy 5000 | shuffled read_formatted_batch numpy 5000 10 | shuffled read_formatted_batch numpy 5000 1000 |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.291595 / 0.215209 (0.076386) | 2.864377 / 2.077655 (0.786722) | 1.543387 / 1.504120 (0.039267) | 1.404273 / 1.541195 (-0.136922) | 1.421964 / 1.468490 (-0.046526) | 0.579275 / 4.584777 (-4.005502) | 0.979212 / 3.745712 (-2.766500) | 2.822043 / 5.269862 (-2.447818) | 1.745015 / 4.565676 (-2.820661) | 0.064626 / 0.424275 (-0.359649) | 0.005006 / 0.007607 (-0.002601) | 0.345509 / 0.226044 (0.119464) | 3.410369 / 2.268929 (1.141440) | 1.875930 / 55.444624 (-53.568694) | 1.600841 / 6.876477 (-5.275636) | 1.611818 / 2.142072 (-0.530254) | 0.662277 / 4.805227 (-4.142950) | 0.117861 / 6.500664 (-6.382803) | 0.041061 / 0.075469 (-0.034408) |\n\n### Benchmark: benchmark_map_filter.json\n\n| metric | filter | map fast-tokenizer batched | map identity | map identity batched | map no-op batched | map no-op batched numpy | map no-op batched pandas | map no-op batched pytorch | map no-op batched tensorflow |\n|--------|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 1.007834 / 1.841788 (-0.833954) | 12.345653 / 8.074308 (4.271345) | 9.775237 / 10.191392 (-0.416155) | 0.135166 / 0.680424 (-0.545258) | 0.016799 / 0.534201 (-0.517402) | 0.289235 / 0.579283 (-0.290048) | 0.126196 / 0.434364 (-0.308168) | 0.382905 / 0.540337 (-0.157432) | 0.435248 / 1.386936 (-0.951688) |\n\n</details>\n</details>\n\n![](https://cml.dev/watermark.png#22bf5388748611a9255d8e17218d36d2f799f182 \"CML watermark\")\n" ]

[]

[ "There are 2 workarounds, tho:\r\n1. Download datasets from web and just load them locally\r\n2. Use metadata directly (temporal solution, since metadata can change)\r\n```\r\nimport datasets\r\nfrom datasets.data_files import DataFilesDict, DataFilesList\r\n\r\ndata_files_list = DataFilesList(\r\n [\r\n \"hf://datasets/allenai/c4@1588ec454efa1a09f29cd18ddd04fe05fc8653a2/en/c4-train.00000-of-01024.json.gz\"\r\n ],\r\n [(\"allenai/c4\", \"1588ec454efa1a09f29cd18ddd04fe05fc8653a2\")],\r\n)\r\ndata_files = DataFilesDict({\"train\": data_files_list})\r\nc4_dataset = datasets.load_dataset(\r\n path=\"allenai/c4\",\r\n data_files=data_files,\r\n split=\"train\",\r\n cache_dir=\"/datesets/cache\",\r\n download_mode=\"reuse_cache_if_exists\",\r\n token=False,\r\n)\r\n```\r\nSecond solution also shows where to find the bug. I suggest that the hashing functions should always use only original parameter `data_files`, and not the one they get after connecting to the server and creating `DataFilesDict`", "Hi! You need to set the `HF_DATASETS_OFFLINE` env variable to `1` to load cached datasets offline, as explained in the docs [here](https://huggingface.co/docs/datasets/v2.19.0/en/loading#offline).", "Just tested. It doesn't work, because of the exact problem I described above: hash of dataset config is different.\r\nThe only error difference is the reason why it cannot connect to HuggingFace (now it's 'offline mode is enabled')\r\n\r\n" ]

[ "Get same error on same datasets too.", "+1", "same error" ]

[ "The docs for this PR live [here](https://moon-ci-docs.huggingface.co/docs/datasets/pr_6835). All of your documentation changes will be reflected on that endpoint. The docs are available until 30 days after the last update.", "Fixed the conversion from `pyarrow` to `python` `Sequence` features. \r\n\r\nThere is still an issue that if `features` are passed the `Sequence` always forces conversion to `ListArray`.\r\nThis probably causes issues if the `LargeListArray` is actually needed.\r\n\r\nThere doesn't seem to be a great solution since this list is created solely on the `schema` for `Sequence`.\r\nOne solution would be to always use `LargeListArray` instead.\r\n", "I am retaking this PR because we would like to have this feature implemented." ]

[]

[ "Similar issue in text process \r\n\r\n```python\r\n\r\ntokenizer=AutoTokenizer.from_pretrained(model_dir[args.model])\r\ntrain_dataset=datasets.load_from_disk(dataset_dir[args.dataset],keep_in_memory=True)['train']\r\ntrain_dataset=train_dataset.map(partial(dname2func[args.dataset],tokenizer=tokenizer),batched=True,num_proc =50,remove_columns=train_dataset.features.keys(),desc='tokenize',keep_in_memory=True)\r\n\r\n```\r\nAfter this train_dataset will be like\r\n```python\r\nDataset({\r\n features: ['input_ids', 'labels'],\r\n num_rows: 51760\r\n})\r\n```\r\nIn which input_ids and labels are both List[int]\r\nHowever, per iter on dataset cost 7.412479639053345s ……？\r\n```python\r\nfor j in tqdm(range(len(train_dataset)),desc='first stage'):\r\n input_id,label=train_dataset['input_ids'][j],train_dataset['labels'][j]\r\n\r\n``` ", "The transform currently replaces the numpy formatting.\r\n\r\nSo you're back to copying data to long python lists which is super slow.\r\n\r\nIt would be cool for the transform to not remove the formatting in this case, but this requires a few changes in the lib" ]

[ "The docs for this PR live [here](https://moon-ci-docs.huggingface.co/docs/datasets/pr_6832). All of your documentation changes will be reflected on that endpoint. The docs are available until 30 days after the last update." ]

[ "The docs for this PR live [here](https://moon-ci-docs.huggingface.co/docs/datasets/pr_6831). All of your documentation changes will be reflected on that endpoint. The docs are available until 30 days after the last update.", "Concretely, the docs about convert_to_parquet are here: https://moon-ci-docs.huggingface.co/docs/datasets/pr_6831/en/cli#convert-to-parquet", "There is an issue with the example snippet when copy/pasting it: the leading shell dollar sign is also copied. I guess they will not like to fix it in the backend: currently they only support Python code snippets (with leading `>>>` or `...`), as they appear in the IPython interactive console.\r\n\r\nWhat do you suggest, @severo?" ]

[]

[]

[ "The docs for this PR live [here](https://moon-ci-docs.huggingface.co/docs/datasets/pr_6828). All of your documentation changes will be reflected on that endpoint. The docs are available until 30 days after the last update." ]

[]

[ "The docs for this PR live [here](https://moon-ci-docs.huggingface.co/docs/datasets/pr_6826). All of your documentation changes will be reflected on that endpoint. The docs are available until 30 days after the last update.", "<details>\n<summary>Show benchmarks</summary>\n\nPyArrow==8.0.0\n\n<details>\n<summary>Show updated benchmarks!</summary>\n\n### Benchmark: benchmark_array_xd.json\n\n| metric | read_batch_formatted_as_numpy after write_array2d | read_batch_formatted_as_numpy after write_flattened_sequence | read_batch_formatted_as_numpy after write_nested_sequence | read_batch_unformated after write_array2d | read_batch_unformated after write_flattened_sequence | read_batch_unformated after write_nested_sequence | read_col_formatted_as_numpy after write_array2d | read_col_formatted_as_numpy after write_flattened_sequence | read_col_formatted_as_numpy after write_nested_sequence | read_col_unformated after write_array2d | read_col_unformated after write_flattened_sequence | read_col_unformated after write_nested_sequence | read_formatted_as_numpy after write_array2d | read_formatted_as_numpy after write_flattened_sequence | read_formatted_as_numpy after write_nested_sequence | read_unformated after write_array2d | read_unformated after write_flattened_sequence | read_unformated after write_nested_sequence | write_array2d | write_flattened_sequence | write_nested_sequence |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.004893 / 0.011353 (-0.006460) | 0.003238 / 0.011008 (-0.007771) | 0.063143 / 0.038508 (0.024635) | 0.029770 / 0.023109 (0.006661) | 0.229052 / 0.275898 (-0.046846) | 0.254534 / 0.323480 (-0.068945) | 0.003083 / 0.007986 (-0.004903) | 0.002615 / 0.004328 (-0.001714) | 0.049684 / 0.004250 (0.045434) | 0.043745 / 0.037052 (0.006693) | 0.248985 / 0.258489 (-0.009504) | 0.275957 / 0.293841 (-0.017884) | 0.027323 / 0.128546 (-0.101223) | 0.010372 / 0.075646 (-0.065275) | 0.206494 / 0.419271 (-0.212778) | 0.035230 / 0.043533 (-0.008303) | 0.234235 / 0.255139 (-0.020904) | 0.252395 / 0.283200 (-0.030805) | 0.019442 / 0.141683 (-0.122240) | 1.130677 / 1.452155 (-0.321478) | 1.161721 / 1.492716 (-0.330996) |\n\n### Benchmark: benchmark_getitem\\_100B.json\n\n| metric | get_batch_of\\_1024\\_random_rows | get_batch_of\\_1024\\_rows | get_first_row | get_last_row |\n|--------|---|---|---|---|\n| new / old (diff) | 0.091659 / 0.018006 (0.073653) | 0.301323 / 0.000490 (0.300833) | 0.000212 / 0.000200 (0.000012) | 0.000049 / 0.000054 (-0.000005) |\n\n### Benchmark: benchmark_indices_mapping.json\n\n| metric | select | shard | shuffle | sort | train_test_split |\n|--------|---|---|---|---|---|\n| new / old (diff) | 0.018360 / 0.037411 (-0.019051) | 0.061101 / 0.014526 (0.046575) | 0.072383 / 0.176557 (-0.104174) | 0.117656 / 0.737135 (-0.619479) | 0.073903 / 0.296338 (-0.222436) |\n\n### Benchmark: benchmark_iterating.json\n\n| metric | read 5000 | read 50000 | read_batch 50000 10 | read_batch 50000 100 | read_batch 50000 1000 | read_formatted numpy 5000 | read_formatted pandas 5000 | read_formatted tensorflow 5000 | read_formatted torch 5000 | read_formatted_batch numpy 5000 10 | read_formatted_batch numpy 5000 1000 | shuffled read 5000 | shuffled read 50000 | shuffled read_batch 50000 10 | shuffled read_batch 50000 100 | shuffled read_batch 50000 1000 | shuffled read_formatted numpy 5000 | shuffled read_formatted_batch numpy 5000 10 | shuffled read_formatted_batch numpy 5000 1000 |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.272768 / 0.215209 (0.057558) | 2.655714 / 2.077655 (0.578059) | 1.446254 / 1.504120 (-0.057866) | 1.330543 / 1.541195 (-0.210652) | 1.352527 / 1.468490 (-0.115964) | 0.561428 / 4.584777 (-4.023349) | 2.368182 / 3.745712 (-1.377530) | 2.746508 / 5.269862 (-2.523353) | 1.713972 / 4.565676 (-2.851705) | 0.062046 / 0.424275 (-0.362229) | 0.005427 / 0.007607 (-0.002180) | 0.321652 / 0.226044 (0.095607) | 3.181812 / 2.268929 (0.912883) | 1.766778 / 55.444624 (-53.677846) | 1.492502 / 6.876477 (-5.383975) | 1.534658 / 2.142072 (-0.607415) | 0.640372 / 4.805227 (-4.164856) | 0.118180 / 6.500664 (-6.382484) | 0.042698 / 0.075469 (-0.032771) |\n\n### Benchmark: benchmark_map_filter.json\n\n| metric | filter | map fast-tokenizer batched | map identity | map identity batched | map no-op batched | map no-op batched numpy | map no-op batched pandas | map no-op batched pytorch | map no-op batched tensorflow |\n|--------|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.993262 / 1.841788 (-0.848525) | 11.512827 / 8.074308 (3.438518) | 9.602140 / 10.191392 (-0.589252) | 0.144723 / 0.680424 (-0.535701) | 0.014122 / 0.534201 (-0.520079) | 0.302211 / 0.579283 (-0.277072) | 0.268026 / 0.434364 (-0.166338) | 0.326524 / 0.540337 (-0.213813) | 0.423781 / 1.386936 (-0.963155) |\n\n</details>\nPyArrow==latest\n\n<details>\n<summary>Show updated benchmarks!</summary>\n\n### Benchmark: benchmark_array_xd.json\n\n| metric | read_batch_formatted_as_numpy after write_array2d | read_batch_formatted_as_numpy after write_flattened_sequence | read_batch_formatted_as_numpy after write_nested_sequence | read_batch_unformated after write_array2d | read_batch_unformated after write_flattened_sequence | read_batch_unformated after write_nested_sequence | read_col_formatted_as_numpy after write_array2d | read_col_formatted_as_numpy after write_flattened_sequence | read_col_formatted_as_numpy after write_nested_sequence | read_col_unformated after write_array2d | read_col_unformated after write_flattened_sequence | read_col_unformated after write_nested_sequence | read_formatted_as_numpy after write_array2d | read_formatted_as_numpy after write_flattened_sequence | read_formatted_as_numpy after write_nested_sequence | read_unformated after write_array2d | read_unformated after write_flattened_sequence | read_unformated after write_nested_sequence | write_array2d | write_flattened_sequence | write_nested_sequence |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.005388 / 0.011353 (-0.005965) | 0.003535 / 0.011008 (-0.007473) | 0.050139 / 0.038508 (0.011631) | 0.031813 / 0.023109 (0.008704) | 0.269501 / 0.275898 (-0.006397) | 0.294355 / 0.323480 (-0.029125) | 0.004128 / 0.007986 (-0.003858) | 0.002684 / 0.004328 (-0.001644) | 0.049295 / 0.004250 (0.045045) | 0.040129 / 0.037052 (0.003077) | 0.282406 / 0.258489 (0.023917) | 0.309822 / 0.293841 (0.015981) | 0.028506 / 0.128546 (-0.100040) | 0.010434 / 0.075646 (-0.065213) | 0.057890 / 0.419271 (-0.361382) | 0.032487 / 0.043533 (-0.011046) | 0.270631 / 0.255139 (0.015492) | 0.288734 / 0.283200 (0.005534) | 0.018710 / 0.141683 (-0.122973) | 1.151571 / 1.452155 (-0.300583) | 1.195222 / 1.492716 (-0.297494) |\n\n### Benchmark: benchmark_getitem\\_100B.json\n\n| metric | get_batch_of\\_1024\\_random_rows | get_batch_of\\_1024\\_rows | get_first_row | get_last_row |\n|--------|---|---|---|---|\n| new / old (diff) | 0.090939 / 0.018006 (0.072932) | 0.300278 / 0.000490 (0.299788) | 0.000202 / 0.000200 (0.000002) | 0.000052 / 0.000054 (-0.000003) |\n\n### Benchmark: benchmark_indices_mapping.json\n\n| metric | select | shard | shuffle | sort | train_test_split |\n|--------|---|---|---|---|---|\n| new / old (diff) | 0.022036 / 0.037411 (-0.015376) | 0.075131 / 0.014526 (0.060605) | 0.087775 / 0.176557 (-0.088782) | 0.125719 / 0.737135 (-0.611416) | 0.088491 / 0.296338 (-0.207848) |\n\n### Benchmark: benchmark_iterating.json\n\n| metric | read 5000 | read 50000 | read_batch 50000 10 | read_batch 50000 100 | read_batch 50000 1000 | read_formatted numpy 5000 | read_formatted pandas 5000 | read_formatted tensorflow 5000 | read_formatted torch 5000 | read_formatted_batch numpy 5000 10 | read_formatted_batch numpy 5000 1000 | shuffled read 5000 | shuffled read 50000 | shuffled read_batch 50000 10 | shuffled read_batch 50000 100 | shuffled read_batch 50000 1000 | shuffled read_formatted numpy 5000 | shuffled read_formatted_batch numpy 5000 10 | shuffled read_formatted_batch numpy 5000 1000 |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.300363 / 0.215209 (0.085154) | 2.931852 / 2.077655 (0.854197) | 1.633688 / 1.504120 (0.129568) | 1.512641 / 1.541195 (-0.028554) | 1.527703 / 1.468490 (0.059213) | 0.572781 / 4.584777 (-4.011996) | 2.445950 / 3.745712 (-1.299762) | 2.883667 / 5.269862 (-2.386195) | 1.761396 / 4.565676 (-2.804280) | 0.064422 / 0.424275 (-0.359853) | 0.005332 / 0.007607 (-0.002275) | 0.346730 / 0.226044 (0.120686) | 3.443815 / 2.268929 (1.174886) | 1.988677 / 55.444624 (-53.455948) | 1.707688 / 6.876477 (-5.168789) | 1.694216 / 2.142072 (-0.447856) | 0.634834 / 4.805227 (-4.170393) | 0.115044 / 6.500664 (-6.385620) | 0.040853 / 0.075469 (-0.034616) |\n\n### Benchmark: benchmark_map_filter.json\n\n| metric | filter | map fast-tokenizer batched | map identity | map identity batched | map no-op batched | map no-op batched numpy | map no-op batched pandas | map no-op batched pytorch | map no-op batched tensorflow |\n|--------|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 1.009382 / 1.841788 (-0.832405) | 12.327511 / 8.074308 (4.253203) | 10.123296 / 10.191392 (-0.068097) | 0.130770 / 0.680424 (-0.549654) | 0.015548 / 0.534201 (-0.518653) | 0.286650 / 0.579283 (-0.292633) | 0.270267 / 0.434364 (-0.164097) | 0.333485 / 0.540337 (-0.206852) | 0.428288 / 1.386936 (-0.958648) |\n\n</details>\n</details>\n\n![](https://cml.dev/watermark.png#f96e74d5c633cd5435dd526adb4a74631eb05c43 \"CML watermark\")\n" ]

[ "The docs for this PR live [here](https://moon-ci-docs.huggingface.co/docs/datasets/pr_6825). All of your documentation changes will be reflected on that endpoint. The docs are available until 30 days after the last update.", "<details>\n<summary>Show benchmarks</summary>\n\nPyArrow==8.0.0\n\n<details>\n<summary>Show updated benchmarks!</summary>\n\n### Benchmark: benchmark_array_xd.json\n\n| metric | read_batch_formatted_as_numpy after write_array2d | read_batch_formatted_as_numpy after write_flattened_sequence | read_batch_formatted_as_numpy after write_nested_sequence | read_batch_unformated after write_array2d | read_batch_unformated after write_flattened_sequence | read_batch_unformated after write_nested_sequence | read_col_formatted_as_numpy after write_array2d | read_col_formatted_as_numpy after write_flattened_sequence | read_col_formatted_as_numpy after write_nested_sequence | read_col_unformated after write_array2d | read_col_unformated after write_flattened_sequence | read_col_unformated after write_nested_sequence | read_formatted_as_numpy after write_array2d | read_formatted_as_numpy after write_flattened_sequence | read_formatted_as_numpy after write_nested_sequence | read_unformated after write_array2d | read_unformated after write_flattened_sequence | read_unformated after write_nested_sequence | write_array2d | write_flattened_sequence | write_nested_sequence |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.004945 / 0.011353 (-0.006407) | 0.003290 / 0.011008 (-0.007718) | 0.062404 / 0.038508 (0.023896) | 0.040056 / 0.023109 (0.016946) | 0.246574 / 0.275898 (-0.029324) | 0.275074 / 0.323480 (-0.048406) | 0.004118 / 0.007986 (-0.003867) | 0.002604 / 0.004328 (-0.001724) | 0.048618 / 0.004250 (0.044367) | 0.044088 / 0.037052 (0.007035) | 0.263059 / 0.258489 (0.004570) | 0.294602 / 0.293841 (0.000761) | 0.027425 / 0.128546 (-0.101121) | 0.010263 / 0.075646 (-0.065383) | 0.205925 / 0.419271 (-0.213346) | 0.048917 / 0.043533 (0.005384) | 0.264227 / 0.255139 (0.009088) | 0.273339 / 0.283200 (-0.009860) | 0.017783 / 0.141683 (-0.123900) | 1.137526 / 1.452155 (-0.314629) | 1.179551 / 1.492716 (-0.313165) |\n\n### Benchmark: benchmark_getitem\\_100B.json\n\n| metric | get_batch_of\\_1024\\_random_rows | get_batch_of\\_1024\\_rows | get_first_row | get_last_row |\n|--------|---|---|---|---|\n| new / old (diff) | 0.096809 / 0.018006 (0.078802) | 0.303854 / 0.000490 (0.303364) | 0.000207 / 0.000200 (0.000007) | 0.000042 / 0.000054 (-0.000012) |\n\n### Benchmark: benchmark_indices_mapping.json\n\n| metric | select | shard | shuffle | sort | train_test_split |\n|--------|---|---|---|---|---|\n| new / old (diff) | 0.017756 / 0.037411 (-0.019655) | 0.061005 / 0.014526 (0.046479) | 0.072986 / 0.176557 (-0.103571) | 0.119851 / 0.737135 (-0.617284) | 0.074733 / 0.296338 (-0.221605) |\n\n### Benchmark: benchmark_iterating.json\n\n| metric | read 5000 | read 50000 | read_batch 50000 10 | read_batch 50000 100 | read_batch 50000 1000 | read_formatted numpy 5000 | read_formatted pandas 5000 | read_formatted tensorflow 5000 | read_formatted torch 5000 | read_formatted_batch numpy 5000 10 | read_formatted_batch numpy 5000 1000 | shuffled read 5000 | shuffled read 50000 | shuffled read_batch 50000 10 | shuffled read_batch 50000 100 | shuffled read_batch 50000 1000 | shuffled read_formatted numpy 5000 | shuffled read_formatted_batch numpy 5000 10 | shuffled read_formatted_batch numpy 5000 1000 |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.278270 / 0.215209 (0.063061) | 2.737874 / 2.077655 (0.660219) | 1.460658 / 1.504120 (-0.043462) | 1.337695 / 1.541195 (-0.203499) | 1.364376 / 1.468490 (-0.104114) | 0.565622 / 4.584777 (-4.019155) | 2.365167 / 3.745712 (-1.380546) | 2.694544 / 5.269862 (-2.575317) | 1.699689 / 4.565676 (-2.865987) | 0.062564 / 0.424275 (-0.361712) | 0.005296 / 0.007607 (-0.002311) | 0.340122 / 0.226044 (0.114077) | 3.382133 / 2.268929 (1.113204) | 1.816907 / 55.444624 (-53.627718) | 1.530825 / 6.876477 (-5.345652) | 1.533266 / 2.142072 (-0.608807) | 0.638215 / 4.805227 (-4.167012) | 0.116227 / 6.500664 (-6.384437) | 0.041548 / 0.075469 (-0.033921) |\n\n### Benchmark: benchmark_map_filter.json\n\n| metric | filter | map fast-tokenizer batched | map identity | map identity batched | map no-op batched | map no-op batched numpy | map no-op batched pandas | map no-op batched pytorch | map no-op batched tensorflow |\n|--------|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.971031 / 1.841788 (-0.870757) | 11.117905 / 8.074308 (3.043597) | 9.358159 / 10.191392 (-0.833233) | 0.127954 / 0.680424 (-0.552470) | 0.013634 / 0.534201 (-0.520567) | 0.285399 / 0.579283 (-0.293885) | 0.267980 / 0.434364 (-0.166383) | 0.320219 / 0.540337 (-0.220119) | 0.416035 / 1.386936 (-0.970901) |\n\n</details>\nPyArrow==latest\n\n<details>\n<summary>Show updated benchmarks!</summary>\n\n### Benchmark: benchmark_array_xd.json\n\n| metric | read_batch_formatted_as_numpy after write_array2d | read_batch_formatted_as_numpy after write_flattened_sequence | read_batch_formatted_as_numpy after write_nested_sequence | read_batch_unformated after write_array2d | read_batch_unformated after write_flattened_sequence | read_batch_unformated after write_nested_sequence | read_col_formatted_as_numpy after write_array2d | read_col_formatted_as_numpy after write_flattened_sequence | read_col_formatted_as_numpy after write_nested_sequence | read_col_unformated after write_array2d | read_col_unformated after write_flattened_sequence | read_col_unformated after write_nested_sequence | read_formatted_as_numpy after write_array2d | read_formatted_as_numpy after write_flattened_sequence | read_formatted_as_numpy after write_nested_sequence | read_unformated after write_array2d | read_unformated after write_flattened_sequence | read_unformated after write_nested_sequence | write_array2d | write_flattened_sequence | write_nested_sequence |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.005177 / 0.011353 (-0.006176) | 0.003078 / 0.011008 (-0.007930) | 0.049650 / 0.038508 (0.011142) | 0.030897 / 0.023109 (0.007787) | 0.271186 / 0.275898 (-0.004712) | 0.296050 / 0.323480 (-0.027430) | 0.004204 / 0.007986 (-0.003781) | 0.002755 / 0.004328 (-0.001574) | 0.049550 / 0.004250 (0.045300) | 0.039801 / 0.037052 (0.002749) | 0.283243 / 0.258489 (0.024753) | 0.310932 / 0.293841 (0.017091) | 0.029136 / 0.128546 (-0.099410) | 0.010278 / 0.075646 (-0.065368) | 0.059300 / 0.419271 (-0.359971) | 0.032965 / 0.043533 (-0.010568) | 0.272646 / 0.255139 (0.017507) | 0.293697 / 0.283200 (0.010497) | 0.018330 / 0.141683 (-0.123353) | 1.144251 / 1.452155 (-0.307904) | 1.209660 / 1.492716 (-0.283056) |\n\n### Benchmark: benchmark_getitem\\_100B.json\n\n| metric | get_batch_of\\_1024\\_random_rows | get_batch_of\\_1024\\_rows | get_first_row | get_last_row |\n|--------|---|---|---|---|\n| new / old (diff) | 0.091020 / 0.018006 (0.073014) | 0.298294 / 0.000490 (0.297804) | 0.000214 / 0.000200 (0.000014) | 0.000053 / 0.000054 (-0.000002) |\n\n### Benchmark: benchmark_indices_mapping.json\n\n| metric | select | shard | shuffle | sort | train_test_split |\n|--------|---|---|---|---|---|\n| new / old (diff) | 0.021879 / 0.037411 (-0.015532) | 0.074728 / 0.014526 (0.060202) | 0.085499 / 0.176557 (-0.091057) | 0.125743 / 0.737135 (-0.611392) | 0.086130 / 0.296338 (-0.210208) |\n\n### Benchmark: benchmark_iterating.json\n\n| metric | read 5000 | read 50000 | read_batch 50000 10 | read_batch 50000 100 | read_batch 50000 1000 | read_formatted numpy 5000 | read_formatted pandas 5000 | read_formatted tensorflow 5000 | read_formatted torch 5000 | read_formatted_batch numpy 5000 10 | read_formatted_batch numpy 5000 1000 | shuffled read 5000 | shuffled read 50000 | shuffled read_batch 50000 10 | shuffled read_batch 50000 100 | shuffled read_batch 50000 1000 | shuffled read_formatted numpy 5000 | shuffled read_formatted_batch numpy 5000 10 | shuffled read_formatted_batch numpy 5000 1000 |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.292311 / 0.215209 (0.077102) | 2.861240 / 2.077655 (0.783585) | 1.590426 / 1.504120 (0.086306) | 1.472288 / 1.541195 (-0.068907) | 1.472901 / 1.468490 (0.004411) | 0.574924 / 4.584777 (-4.009853) | 2.450817 / 3.745712 (-1.294895) | 2.781903 / 5.269862 (-2.487959) | 1.747110 / 4.565676 (-2.818566) | 0.064680 / 0.424275 (-0.359595) | 0.005376 / 0.007607 (-0.002231) | 0.356846 / 0.226044 (0.130802) | 3.457851 / 2.268929 (1.188922) | 1.952678 / 55.444624 (-53.491946) | 1.670824 / 6.876477 (-5.205653) | 1.655872 / 2.142072 (-0.486200) | 0.655874 / 4.805227 (-4.149353) | 0.117098 / 6.500664 (-6.383566) | 0.040230 / 0.075469 (-0.035239) |\n\n### Benchmark: benchmark_map_filter.json\n\n| metric | filter | map fast-tokenizer batched | map identity | map identity batched | map no-op batched | map no-op batched numpy | map no-op batched pandas | map no-op batched pytorch | map no-op batched tensorflow |\n|--------|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 1.007423 / 1.841788 (-0.834365) | 11.818228 / 8.074308 (3.743920) | 10.153699 / 10.191392 (-0.037693) | 0.132073 / 0.680424 (-0.548351) | 0.015101 / 0.534201 (-0.519100) | 0.286555 / 0.579283 (-0.292728) | 0.281953 / 0.434364 (-0.152411) | 0.323647 / 0.540337 (-0.216691) | 0.418698 / 1.386936 (-0.968238) |\n\n</details>\n</details>\n\n![](https://cml.dev/watermark.png#0d3c7462bc67407c42d3ad102b7f9d5914219d9d \"CML watermark\")\n" ]

[ "Hi ! Do you mean 2.18 ? Can you try to update `fsspec` and `huggingface_hub` ?\r\n\r\n```\r\npip install -U fsspec huggingface_hub\r\n```", "Yes I meant 2.18, and it works after updating `fsspec` and `huggingface_hub`. Thanks!" ]

[]

[ "The docs for this PR live [here](https://moon-ci-docs.huggingface.co/docs/datasets/pr_6822). All of your documentation changes will be reflected on that endpoint. The docs are available until 30 days after the last update.", "<details>\n<summary>Show benchmarks</summary>\n\nPyArrow==8.0.0\n\n<details>\n<summary>Show updated benchmarks!</summary>\n\n### Benchmark: benchmark_array_xd.json\n\n| metric | read_batch_formatted_as_numpy after write_array2d | read_batch_formatted_as_numpy after write_flattened_sequence | read_batch_formatted_as_numpy after write_nested_sequence | read_batch_unformated after write_array2d | read_batch_unformated after write_flattened_sequence | read_batch_unformated after write_nested_sequence | read_col_formatted_as_numpy after write_array2d | read_col_formatted_as_numpy after write_flattened_sequence | read_col_formatted_as_numpy after write_nested_sequence | read_col_unformated after write_array2d | read_col_unformated after write_flattened_sequence | read_col_unformated after write_nested_sequence | read_formatted_as_numpy after write_array2d | read_formatted_as_numpy after write_flattened_sequence | read_formatted_as_numpy after write_nested_sequence | read_unformated after write_array2d | read_unformated after write_flattened_sequence | read_unformated after write_nested_sequence | write_array2d | write_flattened_sequence | write_nested_sequence |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.005084 / 0.011353 (-0.006269) | 0.003658 / 0.011008 (-0.007351) | 0.063369 / 0.038508 (0.024860) | 0.030739 / 0.023109 (0.007630) | 0.244335 / 0.275898 (-0.031564) | 0.271731 / 0.323480 (-0.051749) | 0.004133 / 0.007986 (-0.003853) | 0.002798 / 0.004328 (-0.001530) | 0.048790 / 0.004250 (0.044540) | 0.044054 / 0.037052 (0.007002) | 0.261514 / 0.258489 (0.003025) | 0.292155 / 0.293841 (-0.001686) | 0.027971 / 0.128546 (-0.100575) | 0.010723 / 0.075646 (-0.064923) | 0.207328 / 0.419271 (-0.211944) | 0.035928 / 0.043533 (-0.007605) | 0.245320 / 0.255139 (-0.009819) | 0.268774 / 0.283200 (-0.014426) | 0.017119 / 0.141683 (-0.124564) | 1.107052 / 1.452155 (-0.345103) | 1.151752 / 1.492716 (-0.340965) |\n\n### Benchmark: benchmark_getitem\\_100B.json\n\n| metric | get_batch_of\\_1024\\_random_rows | get_batch_of\\_1024\\_rows | get_first_row | get_last_row |\n|--------|---|---|---|---|\n| new / old (diff) | 0.089941 / 0.018006 (0.071935) | 0.299788 / 0.000490 (0.299298) | 0.000211 / 0.000200 (0.000012) | 0.000043 / 0.000054 (-0.000011) |\n\n### Benchmark: benchmark_indices_mapping.json\n\n| metric | select | shard | shuffle | sort | train_test_split |\n|--------|---|---|---|---|---|\n| new / old (diff) | 0.018159 / 0.037411 (-0.019252) | 0.061876 / 0.014526 (0.047350) | 0.074733 / 0.176557 (-0.101824) | 0.122070 / 0.737135 (-0.615065) | 0.076100 / 0.296338 (-0.220238) |\n\n### Benchmark: benchmark_iterating.json\n\n| metric | read 5000 | read 50000 | read_batch 50000 10 | read_batch 50000 100 | read_batch 50000 1000 | read_formatted numpy 5000 | read_formatted pandas 5000 | read_formatted tensorflow 5000 | read_formatted torch 5000 | read_formatted_batch numpy 5000 10 | read_formatted_batch numpy 5000 1000 | shuffled read 5000 | shuffled read 50000 | shuffled read_batch 50000 10 | shuffled read_batch 50000 100 | shuffled read_batch 50000 1000 | shuffled read_formatted numpy 5000 | shuffled read_formatted_batch numpy 5000 10 | shuffled read_formatted_batch numpy 5000 1000 |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.282209 / 0.215209 (0.067000) | 2.758098 / 2.077655 (0.680444) | 1.482454 / 1.504120 (-0.021666) | 1.372649 / 1.541195 (-0.168546) | 1.373171 / 1.468490 (-0.095319) | 0.563606 / 4.584777 (-4.021171) | 2.406760 / 3.745712 (-1.338952) | 2.796322 / 5.269862 (-2.473540) | 1.732327 / 4.565676 (-2.833350) | 0.063623 / 0.424275 (-0.360652) | 0.005338 / 0.007607 (-0.002269) | 0.337562 / 0.226044 (0.111518) | 3.345225 / 2.268929 (1.076296) | 1.844353 / 55.444624 (-53.600271) | 1.551003 / 6.876477 (-5.325474) | 1.570623 / 2.142072 (-0.571449) | 0.644843 / 4.805227 (-4.160385) | 0.118811 / 6.500664 (-6.381853) | 0.041731 / 0.075469 (-0.033738) |\n\n### Benchmark: benchmark_map_filter.json\n\n| metric | filter | map fast-tokenizer batched | map identity | map identity batched | map no-op batched | map no-op batched numpy | map no-op batched pandas | map no-op batched pytorch | map no-op batched tensorflow |\n|--------|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.970469 / 1.841788 (-0.871319) | 11.775531 / 8.074308 (3.701222) | 9.757852 / 10.191392 (-0.433540) | 0.130187 / 0.680424 (-0.550237) | 0.013654 / 0.534201 (-0.520547) | 0.328387 / 0.579283 (-0.250896) | 0.268181 / 0.434364 (-0.166183) | 0.325230 / 0.540337 (-0.215107) | 0.421055 / 1.386936 (-0.965881) |\n\n</details>\nPyArrow==latest\n\n<details>\n<summary>Show updated benchmarks!</summary>\n\n### Benchmark: benchmark_array_xd.json\n\n| metric | read_batch_formatted_as_numpy after write_array2d | read_batch_formatted_as_numpy after write_flattened_sequence | read_batch_formatted_as_numpy after write_nested_sequence | read_batch_unformated after write_array2d | read_batch_unformated after write_flattened_sequence | read_batch_unformated after write_nested_sequence | read_col_formatted_as_numpy after write_array2d | read_col_formatted_as_numpy after write_flattened_sequence | read_col_formatted_as_numpy after write_nested_sequence | read_col_unformated after write_array2d | read_col_unformated after write_flattened_sequence | read_col_unformated after write_nested_sequence | read_formatted_as_numpy after write_array2d | read_formatted_as_numpy after write_flattened_sequence | read_formatted_as_numpy after write_nested_sequence | read_unformated after write_array2d | read_unformated after write_flattened_sequence | read_unformated after write_nested_sequence | write_array2d | write_flattened_sequence | write_nested_sequence |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.005846 / 0.011353 (-0.005507) | 0.003606 / 0.011008 (-0.007402) | 0.050787 / 0.038508 (0.012279) | 0.031635 / 0.023109 (0.008526) | 0.277040 / 0.275898 (0.001142) | 0.300544 / 0.323480 (-0.022936) | 0.004200 / 0.007986 (-0.003786) | 0.002749 / 0.004328 (-0.001580) | 0.049449 / 0.004250 (0.045198) | 0.041616 / 0.037052 (0.004564) | 0.289570 / 0.258489 (0.031081) | 0.316138 / 0.293841 (0.022297) | 0.029578 / 0.128546 (-0.098969) | 0.010582 / 0.075646 (-0.065064) | 0.058284 / 0.419271 (-0.360988) | 0.033078 / 0.043533 (-0.010455) | 0.277964 / 0.255139 (0.022825) | 0.295008 / 0.283200 (0.011808) | 0.017753 / 0.141683 (-0.123930) | 1.128635 / 1.452155 (-0.323519) | 1.190142 / 1.492716 (-0.302575) |\n\n### Benchmark: benchmark_getitem\\_100B.json\n\n| metric | get_batch_of\\_1024\\_random_rows | get_batch_of\\_1024\\_rows | get_first_row | get_last_row |\n|--------|---|---|---|---|\n| new / old (diff) | 0.091504 / 0.018006 (0.073498) | 0.303875 / 0.000490 (0.303385) | 0.000221 / 0.000200 (0.000021) | 0.000052 / 0.000054 (-0.000003) |\n\n### Benchmark: benchmark_indices_mapping.json\n\n| metric | select | shard | shuffle | sort | train_test_split |\n|--------|---|---|---|---|---|\n| new / old (diff) | 0.021413 / 0.037411 (-0.015998) | 0.074825 / 0.014526 (0.060299) | 0.086329 / 0.176557 (-0.090228) | 0.125632 / 0.737135 (-0.611503) | 0.087918 / 0.296338 (-0.208420) |\n\n### Benchmark: benchmark_iterating.json\n\n| metric | read 5000 | read 50000 | read_batch 50000 10 | read_batch 50000 100 | read_batch 50000 1000 | read_formatted numpy 5000 | read_formatted pandas 5000 | read_formatted tensorflow 5000 | read_formatted torch 5000 | read_formatted_batch numpy 5000 10 | read_formatted_batch numpy 5000 1000 | shuffled read 5000 | shuffled read 50000 | shuffled read_batch 50000 10 | shuffled read_batch 50000 100 | shuffled read_batch 50000 1000 | shuffled read_formatted numpy 5000 | shuffled read_formatted_batch numpy 5000 10 | shuffled read_formatted_batch numpy 5000 1000 |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.297914 / 0.215209 (0.082705) | 2.922885 / 2.077655 (0.845230) | 1.625758 / 1.504120 (0.121638) | 1.500174 / 1.541195 (-0.041021) | 1.517162 / 1.468490 (0.048672) | 0.576885 / 4.584777 (-4.007892) | 2.458723 / 3.745712 (-1.286989) | 2.798471 / 5.269862 (-2.471391) | 1.762499 / 4.565676 (-2.803178) | 0.064736 / 0.424275 (-0.359539) | 0.005325 / 0.007607 (-0.002282) | 0.351697 / 0.226044 (0.125652) | 3.496223 / 2.268929 (1.227294) | 1.977535 / 55.444624 (-53.467090) | 1.695223 / 6.876477 (-5.181254) | 1.689692 / 2.142072 (-0.452381) | 0.656404 / 4.805227 (-4.148823) | 0.123106 / 6.500664 (-6.377558) | 0.040980 / 0.075469 (-0.034489) |\n\n### Benchmark: benchmark_map_filter.json\n\n| metric | filter | map fast-tokenizer batched | map identity | map identity batched | map no-op batched | map no-op batched numpy | map no-op batched pandas | map no-op batched pytorch | map no-op batched tensorflow |\n|--------|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 1.036972 / 1.841788 (-0.804816) | 12.163931 / 8.074308 (4.089623) | 10.297927 / 10.191392 (0.106535) | 0.144087 / 0.680424 (-0.536337) | 0.015553 / 0.534201 (-0.518648) | 0.286225 / 0.579283 (-0.293058) | 0.275567 / 0.434364 (-0.158797) | 0.332717 / 0.540337 (-0.207620) | 0.423804 / 1.386936 (-0.963132) |\n\n</details>\n</details>\n\n![](https://cml.dev/watermark.png#0bc709af303c8dc64c973a17016bd5aa5db2f3d5 \"CML watermark\")\n" ]

[ "The docs for this PR live [here](https://moon-ci-docs.huggingface.co/docs/datasets/pr_6820). All of your documentation changes will be reflected on that endpoint. The docs are available until 30 days after the last update.", "This is ready for review, @huggingface/datasets.", "I am adding a test...", "@lhoestq I am getting an error in the test and I think it happens because the CI endpoint does not have the /preupload functionality:\r\n```\r\nhuggingface_hub.utils._errors.RepositoryNotFoundError: 401 Client Error. (Request ID: Root=1-662a4de9-7134df595e29e4c073ac1298;332ff6e3-597a-4dfc-89df-4e9ac64215ad)\r\n\r\nRepository Not Found for url: https://hub-ci.huggingface.co/api/datasets/__DUMMY_TRANSFORMERS_USER__/test-dataset-6c54e2-17140484441915/preupload/main?create_pr=1.\r\nPlease make sure you specified the correct `repo_id` and `repo_type`.\r\nIf you are trying to access a private or gated repo, make sure you are authenticated.\r\nInvalid username or password.\r\nNote: Creating a commit assumes that the repo already exists on the Huggingface Hub. Please use `create_repo` if it's not the case.\r\n```", "@lhoestq, finally, I implemented the test with a mock of the call to `HfApi.create_commit`.", "<details>\n<summary>Show benchmarks</summary>\n\nPyArrow==8.0.0\n\n<details>\n<summary>Show updated benchmarks!</summary>\n\n### Benchmark: benchmark_array_xd.json\n\n| metric | read_batch_formatted_as_numpy after write_array2d | read_batch_formatted_as_numpy after write_flattened_sequence | read_batch_formatted_as_numpy after write_nested_sequence | read_batch_unformated after write_array2d | read_batch_unformated after write_flattened_sequence | read_batch_unformated after write_nested_sequence | read_col_formatted_as_numpy after write_array2d | read_col_formatted_as_numpy after write_flattened_sequence | read_col_formatted_as_numpy after write_nested_sequence | read_col_unformated after write_array2d | read_col_unformated after write_flattened_sequence | read_col_unformated after write_nested_sequence | read_formatted_as_numpy after write_array2d | read_formatted_as_numpy after write_flattened_sequence | read_formatted_as_numpy after write_nested_sequence | read_unformated after write_array2d | read_unformated after write_flattened_sequence | read_unformated after write_nested_sequence | write_array2d | write_flattened_sequence | write_nested_sequence |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.004958 / 0.011353 (-0.006395) | 0.004065 / 0.011008 (-0.006943) | 0.063499 / 0.038508 (0.024991) | 0.030260 / 0.023109 (0.007151) | 0.250910 / 0.275898 (-0.024988) | 0.276632 / 0.323480 (-0.046848) | 0.004038 / 0.007986 (-0.003948) | 0.002721 / 0.004328 (-0.001608) | 0.049098 / 0.004250 (0.044848) | 0.044418 / 0.037052 (0.007366) | 0.262189 / 0.258489 (0.003700) | 0.292426 / 0.293841 (-0.001415) | 0.027268 / 0.128546 (-0.101279) | 0.010601 / 0.075646 (-0.065045) | 0.207332 / 0.419271 (-0.211940) | 0.036102 / 0.043533 (-0.007430) | 0.252425 / 0.255139 (-0.002714) | 0.269421 / 0.283200 (-0.013779) | 0.018534 / 0.141683 (-0.123149) | 1.127869 / 1.452155 (-0.324286) | 1.179660 / 1.492716 (-0.313056) |\n\n### Benchmark: benchmark_getitem\\_100B.json\n\n| metric | get_batch_of\\_1024\\_random_rows | get_batch_of\\_1024\\_rows | get_first_row | get_last_row |\n|--------|---|---|---|---|\n| new / old (diff) | 0.092686 / 0.018006 (0.074680) | 0.299492 / 0.000490 (0.299002) | 0.000211 / 0.000200 (0.000011) | 0.000044 / 0.000054 (-0.000010) |\n\n### Benchmark: benchmark_indices_mapping.json\n\n| metric | select | shard | shuffle | sort | train_test_split |\n|--------|---|---|---|---|---|\n| new / old (diff) | 0.018385 / 0.037411 (-0.019026) | 0.060979 / 0.014526 (0.046453) | 0.073351 / 0.176557 (-0.103205) | 0.120145 / 0.737135 (-0.616990) | 0.073653 / 0.296338 (-0.222686) |\n\n### Benchmark: benchmark_iterating.json\n\n| metric | read 5000 | read 50000 | read_batch 50000 10 | read_batch 50000 100 | read_batch 50000 1000 | read_formatted numpy 5000 | read_formatted pandas 5000 | read_formatted tensorflow 5000 | read_formatted torch 5000 | read_formatted_batch numpy 5000 10 | read_formatted_batch numpy 5000 1000 | shuffled read 5000 | shuffled read 50000 | shuffled read_batch 50000 10 | shuffled read_batch 50000 100 | shuffled read_batch 50000 1000 | shuffled read_formatted numpy 5000 | shuffled read_formatted_batch numpy 5000 10 | shuffled read_formatted_batch numpy 5000 1000 |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.286175 / 0.215209 (0.070966) | 2.792698 / 2.077655 (0.715043) | 1.507442 / 1.504120 (0.003322) | 1.392531 / 1.541195 (-0.148664) | 1.387253 / 1.468490 (-0.081237) | 0.568435 / 4.584777 (-4.016342) | 2.387392 / 3.745712 (-1.358321) | 2.813695 / 5.269862 (-2.456167) | 1.747392 / 4.565676 (-2.818284) | 0.062948 / 0.424275 (-0.361328) | 0.005596 / 0.007607 (-0.002011) | 0.334357 / 0.226044 (0.108313) | 3.263289 / 2.268929 (0.994360) | 1.829553 / 55.444624 (-53.615071) | 1.552510 / 6.876477 (-5.323967) | 1.579975 / 2.142072 (-0.562098) | 0.633982 / 4.805227 (-4.171246) | 0.118752 / 6.500664 (-6.381912) | 0.042445 / 0.075469 (-0.033024) |\n\n### Benchmark: benchmark_map_filter.json\n\n| metric | filter | map fast-tokenizer batched | map identity | map identity batched | map no-op batched | map no-op batched numpy | map no-op batched pandas | map no-op batched pytorch | map no-op batched tensorflow |\n|--------|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.988062 / 1.841788 (-0.853725) | 11.615693 / 8.074308 (3.541385) | 9.728103 / 10.191392 (-0.463289) | 0.131561 / 0.680424 (-0.548862) | 0.015330 / 0.534201 (-0.518871) | 0.289617 / 0.579283 (-0.289666) | 0.265717 / 0.434364 (-0.168646) | 0.323974 / 0.540337 (-0.216363) | 0.419523 / 1.386936 (-0.967413) |\n\n</details>\nPyArrow==latest\n\n<details>\n<summary>Show updated benchmarks!</summary>\n\n### Benchmark: benchmark_array_xd.json\n\n| metric | read_batch_formatted_as_numpy after write_array2d | read_batch_formatted_as_numpy after write_flattened_sequence | read_batch_formatted_as_numpy after write_nested_sequence | read_batch_unformated after write_array2d | read_batch_unformated after write_flattened_sequence | read_batch_unformated after write_nested_sequence | read_col_formatted_as_numpy after write_array2d | read_col_formatted_as_numpy after write_flattened_sequence | read_col_formatted_as_numpy after write_nested_sequence | read_col_unformated after write_array2d | read_col_unformated after write_flattened_sequence | read_col_unformated after write_nested_sequence | read_formatted_as_numpy after write_array2d | read_formatted_as_numpy after write_flattened_sequence | read_formatted_as_numpy after write_nested_sequence | read_unformated after write_array2d | read_unformated after write_flattened_sequence | read_unformated after write_nested_sequence | write_array2d | write_flattened_sequence | write_nested_sequence |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.005385 / 0.011353 (-0.005968) | 0.003753 / 0.011008 (-0.007255) | 0.049821 / 0.038508 (0.011313) | 0.030490 / 0.023109 (0.007381) | 0.260550 / 0.275898 (-0.015348) | 0.284598 / 0.323480 (-0.038881) | 0.004165 / 0.007986 (-0.003821) | 0.002741 / 0.004328 (-0.001588) | 0.048567 / 0.004250 (0.044317) | 0.045185 / 0.037052 (0.008133) | 0.273164 / 0.258489 (0.014674) | 0.301995 / 0.293841 (0.008155) | 0.028802 / 0.128546 (-0.099744) | 0.010539 / 0.075646 (-0.065108) | 0.057967 / 0.419271 (-0.361305) | 0.032826 / 0.043533 (-0.010706) | 0.260425 / 0.255139 (0.005286) | 0.280175 / 0.283200 (-0.003024) | 0.017202 / 0.141683 (-0.124481) | 1.129588 / 1.452155 (-0.322567) | 1.199565 / 1.492716 (-0.293152) |\n\n### Benchmark: benchmark_getitem\\_100B.json\n\n| metric | get_batch_of\\_1024\\_random_rows | get_batch_of\\_1024\\_rows | get_first_row | get_last_row |\n|--------|---|---|---|---|\n| new / old (diff) | 0.091234 / 0.018006 (0.073228) | 0.299313 / 0.000490 (0.298824) | 0.000203 / 0.000200 (0.000003) | 0.000044 / 0.000054 (-0.000010) |\n\n### Benchmark: benchmark_indices_mapping.json\n\n| metric | select | shard | shuffle | sort | train_test_split |\n|--------|---|---|---|---|---|\n| new / old (diff) | 0.022519 / 0.037411 (-0.014892) | 0.075915 / 0.014526 (0.061389) | 0.088636 / 0.176557 (-0.087920) | 0.128234 / 0.737135 (-0.608902) | 0.089782 / 0.296338 (-0.206556) |\n\n### Benchmark: benchmark_iterating.json\n\n| metric | read 5000 | read 50000 | read_batch 50000 10 | read_batch 50000 100 | read_batch 50000 1000 | read_formatted numpy 5000 | read_formatted pandas 5000 | read_formatted tensorflow 5000 | read_formatted torch 5000 | read_formatted_batch numpy 5000 10 | read_formatted_batch numpy 5000 1000 | shuffled read 5000 | shuffled read 50000 | shuffled read_batch 50000 10 | shuffled read_batch 50000 100 | shuffled read_batch 50000 1000 | shuffled read_formatted numpy 5000 | shuffled read_formatted_batch numpy 5000 10 | shuffled read_formatted_batch numpy 5000 1000 |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.291936 / 0.215209 (0.076727) | 2.864589 / 2.077655 (0.786935) | 1.575649 / 1.504120 (0.071529) | 1.452797 / 1.541195 (-0.088398) | 1.476245 / 1.468490 (0.007754) | 0.593972 / 4.584777 (-3.990804) | 0.962315 / 3.745712 (-2.783397) | 2.836496 / 5.269862 (-2.433366) | 1.758639 / 4.565676 (-2.807038) | 0.064842 / 0.424275 (-0.359433) | 0.005076 / 0.007607 (-0.002531) | 0.342568 / 0.226044 (0.116524) | 3.392753 / 2.268929 (1.123825) | 1.908305 / 55.444624 (-53.536319) | 1.632140 / 6.876477 (-5.244337) | 1.653048 / 2.142072 (-0.489024) | 0.662068 / 4.805227 (-4.143159) | 0.118326 / 6.500664 (-6.382338) | 0.041222 / 0.075469 (-0.034247) |\n\n### Benchmark: benchmark_map_filter.json\n\n| metric | filter | map fast-tokenizer batched | map identity | map identity batched | map no-op batched | map no-op batched numpy | map no-op batched pandas | map no-op batched pytorch | map no-op batched tensorflow |\n|--------|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 1.005119 / 1.841788 (-0.836669) | 12.250922 / 8.074308 (4.176614) | 9.775600 / 10.191392 (-0.415792) | 0.146230 / 0.680424 (-0.534194) | 0.015883 / 0.534201 (-0.518318) | 0.290807 / 0.579283 (-0.288476) | 0.126002 / 0.434364 (-0.308362) | 0.392332 / 0.540337 (-0.148005) | 0.435513 / 1.386936 (-0.951423) |\n\n</details>\n</details>\n\n![](https://cml.dev/watermark.png#ceb25e118f21f54b5b5c5e9c223713f14a798eb5 \"CML watermark\")\n" ]

[]

[ "The docs for this PR live [here](https://moon-ci-docs.huggingface.co/docs/datasets/pr_6817). All of your documentation changes will be reflected on that endpoint. The docs are available until 30 days after the last update.", "<details>\n<summary>Show benchmarks</summary>\n\nPyArrow==8.0.0\n\n<details>\n<summary>Show updated benchmarks!</summary>\n\n### Benchmark: benchmark_array_xd.json\n\n| metric | read_batch_formatted_as_numpy after write_array2d | read_batch_formatted_as_numpy after write_flattened_sequence | read_batch_formatted_as_numpy after write_nested_sequence | read_batch_unformated after write_array2d | read_batch_unformated after write_flattened_sequence | read_batch_unformated after write_nested_sequence | read_col_formatted_as_numpy after write_array2d | read_col_formatted_as_numpy after write_flattened_sequence | read_col_formatted_as_numpy after write_nested_sequence | read_col_unformated after write_array2d | read_col_unformated after write_flattened_sequence | read_col_unformated after write_nested_sequence | read_formatted_as_numpy after write_array2d | read_formatted_as_numpy after write_flattened_sequence | read_formatted_as_numpy after write_nested_sequence | read_unformated after write_array2d | read_unformated after write_flattened_sequence | read_unformated after write_nested_sequence | write_array2d | write_flattened_sequence | write_nested_sequence |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.005464 / 0.011353 (-0.005889) | 0.004174 / 0.011008 (-0.006834) | 0.064252 / 0.038508 (0.025744) | 0.033305 / 0.023109 (0.010196) | 0.245831 / 0.275898 (-0.030067) | 0.275575 / 0.323480 (-0.047905) | 0.003359 / 0.007986 (-0.004626) | 0.004196 / 0.004328 (-0.000132) | 0.049961 / 0.004250 (0.045710) | 0.048940 / 0.037052 (0.011888) | 0.261037 / 0.258489 (0.002548) | 0.295329 / 0.293841 (0.001488) | 0.028570 / 0.128546 (-0.099976) | 0.010747 / 0.075646 (-0.064900) | 0.216021 / 0.419271 (-0.203251) | 0.036885 / 0.043533 (-0.006648) | 0.251169 / 0.255139 (-0.003970) | 0.286233 / 0.283200 (0.003034) | 0.021253 / 0.141683 (-0.120429) | 1.150669 / 1.452155 (-0.301485) | 1.187577 / 1.492716 (-0.305140) |\n\n### Benchmark: benchmark_getitem\\_100B.json\n\n| metric | get_batch_of\\_1024\\_random_rows | get_batch_of\\_1024\\_rows | get_first_row | get_last_row |\n|--------|---|---|---|---|\n| new / old (diff) | 0.094443 / 0.018006 (0.076436) | 0.304410 / 0.000490 (0.303920) | 0.000213 / 0.000200 (0.000013) | 0.000041 / 0.000054 (-0.000013) |\n\n### Benchmark: benchmark_indices_mapping.json\n\n| metric | select | shard | shuffle | sort | train_test_split |\n|--------|---|---|---|---|---|\n| new / old (diff) | 0.019568 / 0.037411 (-0.017844) | 0.065734 / 0.014526 (0.051208) | 0.076042 / 0.176557 (-0.100515) | 0.123624 / 0.737135 (-0.613511) | 0.078047 / 0.296338 (-0.218291) |\n\n### Benchmark: benchmark_iterating.json\n\n| metric | read 5000 | read 50000 | read_batch 50000 10 | read_batch 50000 100 | read_batch 50000 1000 | read_formatted numpy 5000 | read_formatted pandas 5000 | read_formatted tensorflow 5000 | read_formatted torch 5000 | read_formatted_batch numpy 5000 10 | read_formatted_batch numpy 5000 1000 | shuffled read 5000 | shuffled read 50000 | shuffled read_batch 50000 10 | shuffled read_batch 50000 100 | shuffled read_batch 50000 1000 | shuffled read_formatted numpy 5000 | shuffled read_formatted_batch numpy 5000 10 | shuffled read_formatted_batch numpy 5000 1000 |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.295725 / 0.215209 (0.080515) | 2.752501 / 2.077655 (0.674846) | 1.461856 / 1.504120 (-0.042264) | 1.353692 / 1.541195 (-0.187503) | 1.391777 / 1.468490 (-0.076713) | 0.563423 / 4.584777 (-4.021354) | 2.384620 / 3.745712 (-1.361092) | 2.876092 / 5.269862 (-2.393769) | 1.803913 / 4.565676 (-2.761763) | 0.062678 / 0.424275 (-0.361597) | 0.005428 / 0.007607 (-0.002179) | 0.333797 / 0.226044 (0.107753) | 3.304458 / 2.268929 (1.035530) | 1.801768 / 55.444624 (-53.642856) | 1.569406 / 6.876477 (-5.307070) | 1.614535 / 2.142072 (-0.527538) | 0.650178 / 4.805227 (-4.155049) | 0.119693 / 6.500664 (-6.380971) | 0.042832 / 0.075469 (-0.032637) |\n\n### Benchmark: benchmark_map_filter.json\n\n| metric | filter | map fast-tokenizer batched | map identity | map identity batched | map no-op batched | map no-op batched numpy | map no-op batched pandas | map no-op batched pytorch | map no-op batched tensorflow |\n|--------|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.982035 / 1.841788 (-0.859753) | 12.390006 / 8.074308 (4.315698) | 10.127018 / 10.191392 (-0.064374) | 0.131963 / 0.680424 (-0.548461) | 0.013926 / 0.534201 (-0.520275) | 0.289587 / 0.579283 (-0.289696) | 0.270302 / 0.434364 (-0.164062) | 0.327231 / 0.540337 (-0.213107) | 0.422522 / 1.386936 (-0.964414) |\n\n</details>\nPyArrow==latest\n\n<details>\n<summary>Show updated benchmarks!</summary>\n\n### Benchmark: benchmark_array_xd.json\n\n| metric | read_batch_formatted_as_numpy after write_array2d | read_batch_formatted_as_numpy after write_flattened_sequence | read_batch_formatted_as_numpy after write_nested_sequence | read_batch_unformated after write_array2d | read_batch_unformated after write_flattened_sequence | read_batch_unformated after write_nested_sequence | read_col_formatted_as_numpy after write_array2d | read_col_formatted_as_numpy after write_flattened_sequence | read_col_formatted_as_numpy after write_nested_sequence | read_col_unformated after write_array2d | read_col_unformated after write_flattened_sequence | read_col_unformated after write_nested_sequence | read_formatted_as_numpy after write_array2d | read_formatted_as_numpy after write_flattened_sequence | read_formatted_as_numpy after write_nested_sequence | read_unformated after write_array2d | read_unformated after write_flattened_sequence | read_unformated after write_nested_sequence | write_array2d | write_flattened_sequence | write_nested_sequence |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.005666 / 0.011353 (-0.005687) | 0.003914 / 0.011008 (-0.007094) | 0.050315 / 0.038508 (0.011807) | 0.032367 / 0.023109 (0.009257) | 0.271732 / 0.275898 (-0.004166) | 0.297248 / 0.323480 (-0.026231) | 0.005101 / 0.007986 (-0.002884) | 0.002882 / 0.004328 (-0.001447) | 0.049651 / 0.004250 (0.045401) | 0.043773 / 0.037052 (0.006721) | 0.288011 / 0.258489 (0.029522) | 0.311863 / 0.293841 (0.018023) | 0.029147 / 0.128546 (-0.099399) | 0.010722 / 0.075646 (-0.064925) | 0.058832 / 0.419271 (-0.360440) | 0.033092 / 0.043533 (-0.010441) | 0.274686 / 0.255139 (0.019547) | 0.294174 / 0.283200 (0.010975) | 0.019196 / 0.141683 (-0.122486) | 1.126615 / 1.452155 (-0.325540) | 1.193107 / 1.492716 (-0.299609) |\n\n### Benchmark: benchmark_getitem\\_100B.json\n\n| metric | get_batch_of\\_1024\\_random_rows | get_batch_of\\_1024\\_rows | get_first_row | get_last_row |\n|--------|---|---|---|---|\n| new / old (diff) | 0.097547 / 0.018006 (0.079541) | 0.316018 / 0.000490 (0.315529) | 0.000330 / 0.000200 (0.000130) | 0.000073 / 0.000054 (0.000019) |\n\n### Benchmark: benchmark_indices_mapping.json\n\n| metric | select | shard | shuffle | sort | train_test_split |\n|--------|---|---|---|---|---|\n| new / old (diff) | 0.022336 / 0.037411 (-0.015076) | 0.077092 / 0.014526 (0.062566) | 0.088873 / 0.176557 (-0.087684) | 0.128517 / 0.737135 (-0.608619) | 0.094061 / 0.296338 (-0.202278) |\n\n### Benchmark: benchmark_iterating.json\n\n| metric | read 5000 | read 50000 | read_batch 50000 10 | read_batch 50000 100 | read_batch 50000 1000 | read_formatted numpy 5000 | read_formatted pandas 5000 | read_formatted tensorflow 5000 | read_formatted torch 5000 | read_formatted_batch numpy 5000 10 | read_formatted_batch numpy 5000 1000 | shuffled read 5000 | shuffled read 50000 | shuffled read_batch 50000 10 | shuffled read_batch 50000 100 | shuffled read_batch 50000 1000 | shuffled read_formatted numpy 5000 | shuffled read_formatted_batch numpy 5000 10 | shuffled read_formatted_batch numpy 5000 1000 |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.300100 / 0.215209 (0.084891) | 2.893114 / 2.077655 (0.815460) | 1.570541 / 1.504120 (0.066421) | 1.453538 / 1.541195 (-0.087657) | 1.505325 / 1.468490 (0.036835) | 0.567955 / 4.584777 (-4.016822) | 2.458547 / 3.745712 (-1.287166) | 2.969181 / 5.269862 (-2.300680) | 1.850082 / 4.565676 (-2.715594) | 0.063811 / 0.424275 (-0.360464) | 0.005378 / 0.007607 (-0.002229) | 0.348219 / 0.226044 (0.122175) | 3.443986 / 2.268929 (1.175057) | 1.943005 / 55.444624 (-53.501620) | 1.686541 / 6.876477 (-5.189935) | 1.715552 / 2.142072 (-0.426520) | 0.641361 / 4.805227 (-4.163866) | 0.116652 / 6.500664 (-6.384012) | 0.042216 / 0.075469 (-0.033253) |\n\n### Benchmark: benchmark_map_filter.json\n\n| metric | filter | map fast-tokenizer batched | map identity | map identity batched | map no-op batched | map no-op batched numpy | map no-op batched pandas | map no-op batched pytorch | map no-op batched tensorflow |\n|--------|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 1.020102 / 1.841788 (-0.821686) | 12.966127 / 8.074308 (4.891819) | 10.748397 / 10.191392 (0.557005) | 0.132601 / 0.680424 (-0.547823) | 0.016643 / 0.534201 (-0.517558) | 0.289422 / 0.579283 (-0.289861) | 0.275524 / 0.434364 (-0.158840) | 0.332835 / 0.540337 (-0.207503) | 0.427867 / 1.386936 (-0.959069) |\n\n</details>\n</details>\n\n![](https://cml.dev/watermark.png#5eb93f61f9f6e7fefba5d800defe21e50ddf8c58 \"CML watermark\")\n" ]

[ "The docs for this PR live [here](https://moon-ci-docs.huggingface.co/docs/datasets/pr_6816). All of your documentation changes will be reflected on that endpoint. The docs are available until 30 days after the last update.", "Hi! This is a breaking change. A better solution is to check for \"indexable\" types in `__getitem__` to support keys such as `np.int64`:\r\n```python\r\nimport operator\r\n\r\ndef _query_table_with_indices_mapping(...): # or _query_table\r\n ...\r\n try:\r\n operator.index(key)\r\n except TypeError:\r\n pass\r\n \r\n _raise_bad_key_type(key)\r\n```", "Sounds good! We should still update type annotations for SearchResult in my opinion." ]

[ "The docs for this PR live [here](https://moon-ci-docs.huggingface.co/docs/datasets/pr_6815). All of your documentation changes will be reflected on that endpoint. The docs are available until 30 days after the last update.", "<details>\n<summary>Show benchmarks</summary>\n\nPyArrow==8.0.0\n\n<details>\n<summary>Show updated benchmarks!</summary>\n\n### Benchmark: benchmark_array_xd.json\n\n| metric | read_batch_formatted_as_numpy after write_array2d | read_batch_formatted_as_numpy after write_flattened_sequence | read_batch_formatted_as_numpy after write_nested_sequence | read_batch_unformated after write_array2d | read_batch_unformated after write_flattened_sequence | read_batch_unformated after write_nested_sequence | read_col_formatted_as_numpy after write_array2d | read_col_formatted_as_numpy after write_flattened_sequence | read_col_formatted_as_numpy after write_nested_sequence | read_col_unformated after write_array2d | read_col_unformated after write_flattened_sequence | read_col_unformated after write_nested_sequence | read_formatted_as_numpy after write_array2d | read_formatted_as_numpy after write_flattened_sequence | read_formatted_as_numpy after write_nested_sequence | read_unformated after write_array2d | read_unformated after write_flattened_sequence | read_unformated after write_nested_sequence | write_array2d | write_flattened_sequence | write_nested_sequence |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.005101 / 0.011353 (-0.006252) | 0.003478 / 0.011008 (-0.007531) | 0.063634 / 0.038508 (0.025126) | 0.030670 / 0.023109 (0.007561) | 0.240057 / 0.275898 (-0.035841) | 0.258726 / 0.323480 (-0.064754) | 0.004136 / 0.007986 (-0.003849) | 0.002667 / 0.004328 (-0.001662) | 0.048968 / 0.004250 (0.044718) | 0.043125 / 0.037052 (0.006073) | 0.249033 / 0.258489 (-0.009456) | 0.282630 / 0.293841 (-0.011211) | 0.027528 / 0.128546 (-0.101018) | 0.009987 / 0.075646 (-0.065660) | 0.210614 / 0.419271 (-0.208657) | 0.034965 / 0.043533 (-0.008567) | 0.239199 / 0.255139 (-0.015940) | 0.276891 / 0.283200 (-0.006309) | 0.017781 / 0.141683 (-0.123902) | 1.142795 / 1.452155 (-0.309360) | 1.184171 / 1.492716 (-0.308545) |\n\n### Benchmark: benchmark_getitem\\_100B.json\n\n| metric | get_batch_of\\_1024\\_random_rows | get_batch_of\\_1024\\_rows | get_first_row | get_last_row |\n|--------|---|---|---|---|\n| new / old (diff) | 0.092075 / 0.018006 (0.074068) | 0.300709 / 0.000490 (0.300220) | 0.000217 / 0.000200 (0.000017) | 0.000046 / 0.000054 (-0.000008) |\n\n### Benchmark: benchmark_indices_mapping.json\n\n| metric | select | shard | shuffle | sort | train_test_split |\n|--------|---|---|---|---|---|\n| new / old (diff) | 0.017887 / 0.037411 (-0.019525) | 0.061134 / 0.014526 (0.046608) | 0.077075 / 0.176557 (-0.099482) | 0.118808 / 0.737135 (-0.618327) | 0.074961 / 0.296338 (-0.221377) |\n\n### Benchmark: benchmark_iterating.json\n\n| metric | read 5000 | read 50000 | read_batch 50000 10 | read_batch 50000 100 | read_batch 50000 1000 | read_formatted numpy 5000 | read_formatted pandas 5000 | read_formatted tensorflow 5000 | read_formatted torch 5000 | read_formatted_batch numpy 5000 10 | read_formatted_batch numpy 5000 1000 | shuffled read 5000 | shuffled read 50000 | shuffled read_batch 50000 10 | shuffled read_batch 50000 100 | shuffled read_batch 50000 1000 | shuffled read_formatted numpy 5000 | shuffled read_formatted_batch numpy 5000 10 | shuffled read_formatted_batch numpy 5000 1000 |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.280404 / 0.215209 (0.065194) | 2.759453 / 2.077655 (0.681798) | 1.437552 / 1.504120 (-0.066568) | 1.318703 / 1.541195 (-0.222492) | 1.313075 / 1.468490 (-0.155416) | 0.564876 / 4.584777 (-4.019901) | 2.381595 / 3.745712 (-1.364118) | 2.759171 / 5.269862 (-2.510691) | 1.725878 / 4.565676 (-2.839799) | 0.062627 / 0.424275 (-0.361648) | 0.005295 / 0.007607 (-0.002312) | 0.335245 / 0.226044 (0.109201) | 3.276266 / 2.268929 (1.007337) | 1.843272 / 55.444624 (-53.601353) | 1.519948 / 6.876477 (-5.356529) | 1.519626 / 2.142072 (-0.622447) | 0.637891 / 4.805227 (-4.167336) | 0.116260 / 6.500664 (-6.384404) | 0.041768 / 0.075469 (-0.033701) |\n\n### Benchmark: benchmark_map_filter.json\n\n| metric | filter | map fast-tokenizer batched | map identity | map identity batched | map no-op batched | map no-op batched numpy | map no-op batched pandas | map no-op batched pytorch | map no-op batched tensorflow |\n|--------|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.981739 / 1.841788 (-0.860049) | 11.354768 / 8.074308 (3.280460) | 9.900585 / 10.191392 (-0.290807) | 0.130683 / 0.680424 (-0.549741) | 0.014122 / 0.534201 (-0.520079) | 0.297451 / 0.579283 (-0.281832) | 0.264786 / 0.434364 (-0.169577) | 0.337559 / 0.540337 (-0.202778) | 0.425131 / 1.386936 (-0.961805) |\n\n</details>\nPyArrow==latest\n\n<details>\n<summary>Show updated benchmarks!</summary>\n\n### Benchmark: benchmark_array_xd.json\n\n| metric | read_batch_formatted_as_numpy after write_array2d | read_batch_formatted_as_numpy after write_flattened_sequence | read_batch_formatted_as_numpy after write_nested_sequence | read_batch_unformated after write_array2d | read_batch_unformated after write_flattened_sequence | read_batch_unformated after write_nested_sequence | read_col_formatted_as_numpy after write_array2d | read_col_formatted_as_numpy after write_flattened_sequence | read_col_formatted_as_numpy after write_nested_sequence | read_col_unformated after write_array2d | read_col_unformated after write_flattened_sequence | read_col_unformated after write_nested_sequence | read_formatted_as_numpy after write_array2d | read_formatted_as_numpy after write_flattened_sequence | read_formatted_as_numpy after write_nested_sequence | read_unformated after write_array2d | read_unformated after write_flattened_sequence | read_unformated after write_nested_sequence | write_array2d | write_flattened_sequence | write_nested_sequence |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.005182 / 0.011353 (-0.006171) | 0.003355 / 0.011008 (-0.007653) | 0.049842 / 0.038508 (0.011334) | 0.031094 / 0.023109 (0.007985) | 0.270080 / 0.275898 (-0.005818) | 0.291602 / 0.323480 (-0.031878) | 0.004210 / 0.007986 (-0.003776) | 0.002720 / 0.004328 (-0.001608) | 0.048986 / 0.004250 (0.044736) | 0.055187 / 0.037052 (0.018135) | 0.280085 / 0.258489 (0.021595) | 0.308148 / 0.293841 (0.014308) | 0.029300 / 0.128546 (-0.099246) | 0.009976 / 0.075646 (-0.065670) | 0.057930 / 0.419271 (-0.361341) | 0.032543 / 0.043533 (-0.010990) | 0.277485 / 0.255139 (0.022346) | 0.289345 / 0.283200 (0.006145) | 0.018070 / 0.141683 (-0.123613) | 1.140977 / 1.452155 (-0.311178) | 1.190543 / 1.492716 (-0.302173) |\n\n### Benchmark: benchmark_getitem\\_100B.json\n\n| metric | get_batch_of\\_1024\\_random_rows | get_batch_of\\_1024\\_rows | get_first_row | get_last_row |\n|--------|---|---|---|---|\n| new / old (diff) | 0.093416 / 0.018006 (0.075410) | 0.298732 / 0.000490 (0.298242) | 0.000224 / 0.000200 (0.000024) | 0.000051 / 0.000054 (-0.000004) |\n\n### Benchmark: benchmark_indices_mapping.json\n\n| metric | select | shard | shuffle | sort | train_test_split |\n|--------|---|---|---|---|---|\n| new / old (diff) | 0.022167 / 0.037411 (-0.015244) | 0.074970 / 0.014526 (0.060444) | 0.086047 / 0.176557 (-0.090509) | 0.125228 / 0.737135 (-0.611907) | 0.088330 / 0.296338 (-0.208008) |\n\n### Benchmark: benchmark_iterating.json\n\n| metric | read 5000 | read 50000 | read_batch 50000 10 | read_batch 50000 100 | read_batch 50000 1000 | read_formatted numpy 5000 | read_formatted pandas 5000 | read_formatted tensorflow 5000 | read_formatted torch 5000 | read_formatted_batch numpy 5000 10 | read_formatted_batch numpy 5000 1000 | shuffled read 5000 | shuffled read 50000 | shuffled read_batch 50000 10 | shuffled read_batch 50000 100 | shuffled read_batch 50000 1000 | shuffled read_formatted numpy 5000 | shuffled read_formatted_batch numpy 5000 10 | shuffled read_formatted_batch numpy 5000 1000 |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.292016 / 0.215209 (0.076807) | 2.845712 / 2.077655 (0.768057) | 1.576951 / 1.504120 (0.072831) | 1.452298 / 1.541195 (-0.088897) | 1.456918 / 1.468490 (-0.011572) | 0.560529 / 4.584777 (-4.024248) | 2.425333 / 3.745712 (-1.320379) | 2.739416 / 5.269862 (-2.530445) | 1.715779 / 4.565676 (-2.849898) | 0.062568 / 0.424275 (-0.361707) | 0.005327 / 0.007607 (-0.002280) | 0.351376 / 0.226044 (0.125332) | 3.401855 / 2.268929 (1.132927) | 1.921844 / 55.444624 (-53.522780) | 1.648423 / 6.876477 (-5.228054) | 1.642003 / 2.142072 (-0.500069) | 0.640789 / 4.805227 (-4.164438) | 0.114699 / 6.500664 (-6.385965) | 0.040451 / 0.075469 (-0.035018) |\n\n### Benchmark: benchmark_map_filter.json\n\n| metric | filter | map fast-tokenizer batched | map identity | map identity batched | map no-op batched | map no-op batched numpy | map no-op batched pandas | map no-op batched pytorch | map no-op batched tensorflow |\n|--------|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 1.004186 / 1.841788 (-0.837602) | 11.879918 / 8.074308 (3.805609) | 9.981852 / 10.191392 (-0.209540) | 0.141298 / 0.680424 (-0.539126) | 0.015005 / 0.534201 (-0.519196) | 0.291537 / 0.579283 (-0.287746) | 0.272093 / 0.434364 (-0.162271) | 0.331361 / 0.540337 (-0.208977) | 0.422940 / 1.386936 (-0.963996) |\n\n</details>\n</details>\n\n![](https://cml.dev/watermark.png#ed8860faef3e751f3b77c08e09ce723a74d2c2e5 \"CML watermark\")\n" ]

[ "Hi ! You can try to reduce `writer_batch_size`. It corresponds to the number of samples that stay in RAM before being flushed to disk" ]

[ "The docs for this PR live [here](https://moon-ci-docs.huggingface.co/docs/datasets/pr_6813). All of your documentation changes will be reflected on that endpoint. The docs are available until 30 days after the last update.", "<details>\n<summary>Show benchmarks</summary>\n\nPyArrow==8.0.0\n\n<details>\n<summary>Show updated benchmarks!</summary>\n\n### Benchmark: benchmark_array_xd.json\n\n| metric | read_batch_formatted_as_numpy after write_array2d | read_batch_formatted_as_numpy after write_flattened_sequence | read_batch_formatted_as_numpy after write_nested_sequence | read_batch_unformated after write_array2d | read_batch_unformated after write_flattened_sequence | read_batch_unformated after write_nested_sequence | read_col_formatted_as_numpy after write_array2d | read_col_formatted_as_numpy after write_flattened_sequence | read_col_formatted_as_numpy after write_nested_sequence | read_col_unformated after write_array2d | read_col_unformated after write_flattened_sequence | read_col_unformated after write_nested_sequence | read_formatted_as_numpy after write_array2d | read_formatted_as_numpy after write_flattened_sequence | read_formatted_as_numpy after write_nested_sequence | read_unformated after write_array2d | read_unformated after write_flattened_sequence | read_unformated after write_nested_sequence | write_array2d | write_flattened_sequence | write_nested_sequence |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.005153 / 0.011353 (-0.006200) | 0.003560 / 0.011008 (-0.007448) | 0.063142 / 0.038508 (0.024634) | 0.030799 / 0.023109 (0.007690) | 0.241754 / 0.275898 (-0.034144) | 0.264874 / 0.323480 (-0.058606) | 0.003099 / 0.007986 (-0.004887) | 0.002629 / 0.004328 (-0.001700) | 0.049006 / 0.004250 (0.044756) | 0.044831 / 0.037052 (0.007779) | 0.258961 / 0.258489 (0.000472) | 0.286939 / 0.293841 (-0.006902) | 0.026756 / 0.128546 (-0.101791) | 0.010443 / 0.075646 (-0.065204) | 0.207264 / 0.419271 (-0.212007) | 0.035242 / 0.043533 (-0.008291) | 0.250440 / 0.255139 (-0.004699) | 0.265405 / 0.283200 (-0.017794) | 0.018924 / 0.141683 (-0.122759) | 1.138607 / 1.452155 (-0.313547) | 1.203017 / 1.492716 (-0.289700) |\n\n### Benchmark: benchmark_getitem\\_100B.json\n\n| metric | get_batch_of\\_1024\\_random_rows | get_batch_of\\_1024\\_rows | get_first_row | get_last_row |\n|--------|---|---|---|---|\n| new / old (diff) | 0.091293 / 0.018006 (0.073286) | 0.303937 / 0.000490 (0.303447) | 0.000266 / 0.000200 (0.000066) | 0.000056 / 0.000054 (0.000002) |\n\n### Benchmark: benchmark_indices_mapping.json\n\n| metric | select | shard | shuffle | sort | train_test_split |\n|--------|---|---|---|---|---|\n| new / old (diff) | 0.018667 / 0.037411 (-0.018744) | 0.061310 / 0.014526 (0.046784) | 0.073565 / 0.176557 (-0.102991) | 0.119044 / 0.737135 (-0.618091) | 0.074484 / 0.296338 (-0.221854) |\n\n### Benchmark: benchmark_iterating.json\n\n| metric | read 5000 | read 50000 | read_batch 50000 10 | read_batch 50000 100 | read_batch 50000 1000 | read_formatted numpy 5000 | read_formatted pandas 5000 | read_formatted tensorflow 5000 | read_formatted torch 5000 | read_formatted_batch numpy 5000 10 | read_formatted_batch numpy 5000 1000 | shuffled read 5000 | shuffled read 50000 | shuffled read_batch 50000 10 | shuffled read_batch 50000 100 | shuffled read_batch 50000 1000 | shuffled read_formatted numpy 5000 | shuffled read_formatted_batch numpy 5000 10 | shuffled read_formatted_batch numpy 5000 1000 |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.286324 / 0.215209 (0.071114) | 2.836637 / 2.077655 (0.758982) | 1.458531 / 1.504120 (-0.045589) | 1.333081 / 1.541195 (-0.208114) | 1.328398 / 1.468490 (-0.140092) | 0.571467 / 4.584777 (-4.013310) | 2.409869 / 3.745712 (-1.335843) | 2.760241 / 5.269862 (-2.509621) | 1.728153 / 4.565676 (-2.837523) | 0.063008 / 0.424275 (-0.361267) | 0.005375 / 0.007607 (-0.002232) | 0.338574 / 0.226044 (0.112530) | 3.355485 / 2.268929 (1.086556) | 1.812741 / 55.444624 (-53.631884) | 1.507435 / 6.876477 (-5.369041) | 1.516957 / 2.142072 (-0.625116) | 0.643790 / 4.805227 (-4.161437) | 0.117465 / 6.500664 (-6.383199) | 0.041960 / 0.075469 (-0.033509) |\n\n### Benchmark: benchmark_map_filter.json\n\n| metric | filter | map fast-tokenizer batched | map identity | map identity batched | map no-op batched | map no-op batched numpy | map no-op batched pandas | map no-op batched pytorch | map no-op batched tensorflow |\n|--------|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.993787 / 1.841788 (-0.848001) | 11.439076 / 8.074308 (3.364768) | 9.636815 / 10.191392 (-0.554577) | 0.131292 / 0.680424 (-0.549132) | 0.014916 / 0.534201 (-0.519285) | 0.287309 / 0.579283 (-0.291974) | 0.261971 / 0.434364 (-0.172392) | 0.324453 / 0.540337 (-0.215885) | 0.420306 / 1.386936 (-0.966630) |\n\n</details>\nPyArrow==latest\n\n<details>\n<summary>Show updated benchmarks!</summary>\n\n### Benchmark: benchmark_array_xd.json\n\n| metric | read_batch_formatted_as_numpy after write_array2d | read_batch_formatted_as_numpy after write_flattened_sequence | read_batch_formatted_as_numpy after write_nested_sequence | read_batch_unformated after write_array2d | read_batch_unformated after write_flattened_sequence | read_batch_unformated after write_nested_sequence | read_col_formatted_as_numpy after write_array2d | read_col_formatted_as_numpy after write_flattened_sequence | read_col_formatted_as_numpy after write_nested_sequence | read_col_unformated after write_array2d | read_col_unformated after write_flattened_sequence | read_col_unformated after write_nested_sequence | read_formatted_as_numpy after write_array2d | read_formatted_as_numpy after write_flattened_sequence | read_formatted_as_numpy after write_nested_sequence | read_unformated after write_array2d | read_unformated after write_flattened_sequence | read_unformated after write_nested_sequence | write_array2d | write_flattened_sequence | write_nested_sequence |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.005138 / 0.011353 (-0.006215) | 0.003719 / 0.011008 (-0.007289) | 0.050411 / 0.038508 (0.011903) | 0.031334 / 0.023109 (0.008225) | 0.281752 / 0.275898 (0.005854) | 0.299445 / 0.323480 (-0.024035) | 0.004194 / 0.007986 (-0.003792) | 0.002737 / 0.004328 (-0.001591) | 0.048527 / 0.004250 (0.044277) | 0.040294 / 0.037052 (0.003242) | 0.291763 / 0.258489 (0.033274) | 0.317597 / 0.293841 (0.023757) | 0.029014 / 0.128546 (-0.099532) | 0.010372 / 0.075646 (-0.065274) | 0.058704 / 0.419271 (-0.360568) | 0.033259 / 0.043533 (-0.010273) | 0.278109 / 0.255139 (0.022970) | 0.299593 / 0.283200 (0.016393) | 0.018048 / 0.141683 (-0.123635) | 1.185558 / 1.452155 (-0.266597) | 1.203481 / 1.492716 (-0.289236) |\n\n### Benchmark: benchmark_getitem\\_100B.json\n\n| metric | get_batch_of\\_1024\\_random_rows | get_batch_of\\_1024\\_rows | get_first_row | get_last_row |\n|--------|---|---|---|---|\n| new / old (diff) | 0.091149 / 0.018006 (0.073143) | 0.306152 / 0.000490 (0.305662) | 0.000246 / 0.000200 (0.000046) | 0.000052 / 0.000054 (-0.000003) |\n\n### Benchmark: benchmark_indices_mapping.json\n\n| metric | select | shard | shuffle | sort | train_test_split |\n|--------|---|---|---|---|---|\n| new / old (diff) | 0.022082 / 0.037411 (-0.015330) | 0.074487 / 0.014526 (0.059961) | 0.086112 / 0.176557 (-0.090444) | 0.124303 / 0.737135 (-0.612832) | 0.088831 / 0.296338 (-0.207508) |\n\n### Benchmark: benchmark_iterating.json\n\n| metric | read 5000 | read 50000 | read_batch 50000 10 | read_batch 50000 100 | read_batch 50000 1000 | read_formatted numpy 5000 | read_formatted pandas 5000 | read_formatted tensorflow 5000 | read_formatted torch 5000 | read_formatted_batch numpy 5000 10 | read_formatted_batch numpy 5000 1000 | shuffled read 5000 | shuffled read 50000 | shuffled read_batch 50000 10 | shuffled read_batch 50000 100 | shuffled read_batch 50000 1000 | shuffled read_formatted numpy 5000 | shuffled read_formatted_batch numpy 5000 10 | shuffled read_formatted_batch numpy 5000 1000 |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.291745 / 0.215209 (0.076536) | 2.878397 / 2.077655 (0.800742) | 1.606920 / 1.504120 (0.102801) | 1.492352 / 1.541195 (-0.048843) | 1.509725 / 1.468490 (0.041235) | 0.567087 / 4.584777 (-4.017690) | 2.436423 / 3.745712 (-1.309290) | 2.793930 / 5.269862 (-2.475932) | 1.748329 / 4.565676 (-2.817347) | 0.063424 / 0.424275 (-0.360851) | 0.005476 / 0.007607 (-0.002131) | 0.346211 / 0.226044 (0.120167) | 3.461288 / 2.268929 (1.192360) | 1.979362 / 55.444624 (-53.465262) | 1.702877 / 6.876477 (-5.173600) | 1.699087 / 2.142072 (-0.442985) | 0.645116 / 4.805227 (-4.160112) | 0.116186 / 6.500664 (-6.384478) | 0.041246 / 0.075469 (-0.034223) |\n\n### Benchmark: benchmark_map_filter.json\n\n| metric | filter | map fast-tokenizer batched | map identity | map identity batched | map no-op batched | map no-op batched numpy | map no-op batched pandas | map no-op batched pytorch | map no-op batched tensorflow |\n|--------|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 1.017540 / 1.841788 (-0.824248) | 12.016640 / 8.074308 (3.942332) | 10.234085 / 10.191392 (0.042693) | 0.147558 / 0.680424 (-0.532866) | 0.015096 / 0.534201 (-0.519105) | 0.288077 / 0.579283 (-0.291206) | 0.274629 / 0.434364 (-0.159735) | 0.334097 / 0.540337 (-0.206241) | 0.425476 / 1.386936 (-0.961460) |\n\n</details>\n</details>\n\n![](https://cml.dev/watermark.png#55eb1d9a34a91dbf2418166f9f1d92f7181e778b \"CML watermark\")\n" ]

[ "(Sorry, meant to open this against my own fork. I'm attempting to debug this issue (https://github.com/astral-sh/uv/issues/1921#issuecomment-2058056192) reported by `huggingface/datasets` on the uv repo.)" ]

[ "The docs for this PR live [here](https://moon-ci-docs.huggingface.co/docs/datasets/pr_6811). All of your documentation changes will be reflected on that endpoint. The docs are available until 30 days after the last update.", "@mariosasko pytest seems to be missing on windows?", "CI is not behaving well today 🙂 ", "I couldn't find an instance of the `allow_number_to_str` parameter (or `array_cast`/`cast_array_to_feature` more generally) being used in the wild. So, I think simply removing `allow_number_to_str` instead of deprecating it should be fine, considering `array_cast`/`cast_array_to_feature` are somewhat hidden. Do you agree @lhoestq? ", "Yup we can remove without any deprecation cycle", "<details>\n<summary>Show benchmarks</summary>\n\nPyArrow==8.0.0\n\n<details>\n<summary>Show updated benchmarks!</summary>\n\n### Benchmark: benchmark_array_xd.json\n\n| metric | read_batch_formatted_as_numpy after write_array2d | read_batch_formatted_as_numpy after write_flattened_sequence | read_batch_formatted_as_numpy after write_nested_sequence | read_batch_unformated after write_array2d | read_batch_unformated after write_flattened_sequence | read_batch_unformated after write_nested_sequence | read_col_formatted_as_numpy after write_array2d | read_col_formatted_as_numpy after write_flattened_sequence | read_col_formatted_as_numpy after write_nested_sequence | read_col_unformated after write_array2d | read_col_unformated after write_flattened_sequence | read_col_unformated after write_nested_sequence | read_formatted_as_numpy after write_array2d | read_formatted_as_numpy after write_flattened_sequence | read_formatted_as_numpy after write_nested_sequence | read_unformated after write_array2d | read_unformated after write_flattened_sequence | read_unformated after write_nested_sequence | write_array2d | write_flattened_sequence | write_nested_sequence |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.005253 / 0.011353 (-0.006100) | 0.003767 / 0.011008 (-0.007241) | 0.064599 / 0.038508 (0.026091) | 0.030758 / 0.023109 (0.007649) | 0.237437 / 0.275898 (-0.038461) | 0.277580 / 0.323480 (-0.045900) | 0.004220 / 0.007986 (-0.003766) | 0.002738 / 0.004328 (-0.001591) | 0.049393 / 0.004250 (0.045143) | 0.045283 / 0.037052 (0.008231) | 0.249907 / 0.258489 (-0.008582) | 0.283301 / 0.293841 (-0.010540) | 0.027722 / 0.128546 (-0.100825) | 0.010842 / 0.075646 (-0.064804) | 0.219197 / 0.419271 (-0.200074) | 0.036449 / 0.043533 (-0.007084) | 0.237774 / 0.255139 (-0.017365) | 0.257981 / 0.283200 (-0.025218) | 0.018098 / 0.141683 (-0.123585) | 1.161778 / 1.452155 (-0.290376) | 1.212707 / 1.492716 (-0.280010) |\n\n### Benchmark: benchmark_getitem\\_100B.json\n\n| metric | get_batch_of\\_1024\\_random_rows | get_batch_of\\_1024\\_rows | get_first_row | get_last_row |\n|--------|---|---|---|---|\n| new / old (diff) | 0.096462 / 0.018006 (0.078456) | 0.305322 / 0.000490 (0.304832) | 0.000218 / 0.000200 (0.000018) | 0.000048 / 0.000054 (-0.000006) |\n\n### Benchmark: benchmark_indices_mapping.json\n\n| metric | select | shard | shuffle | sort | train_test_split |\n|--------|---|---|---|---|---|\n| new / old (diff) | 0.018438 / 0.037411 (-0.018973) | 0.061633 / 0.014526 (0.047107) | 0.073678 / 0.176557 (-0.102879) | 0.122033 / 0.737135 (-0.615103) | 0.074846 / 0.296338 (-0.221493) |\n\n### Benchmark: benchmark_iterating.json\n\n| metric | read 5000 | read 50000 | read_batch 50000 10 | read_batch 50000 100 | read_batch 50000 1000 | read_formatted numpy 5000 | read_formatted pandas 5000 | read_formatted tensorflow 5000 | read_formatted torch 5000 | read_formatted_batch numpy 5000 10 | read_formatted_batch numpy 5000 1000 | shuffled read 5000 | shuffled read 50000 | shuffled read_batch 50000 10 | shuffled read_batch 50000 100 | shuffled read_batch 50000 1000 | shuffled read_formatted numpy 5000 | shuffled read_formatted_batch numpy 5000 10 | shuffled read_formatted_batch numpy 5000 1000 |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.279564 / 0.215209 (0.064355) | 2.756984 / 2.077655 (0.679330) | 1.486525 / 1.504120 (-0.017595) | 1.366474 / 1.541195 (-0.174721) | 1.370192 / 1.468490 (-0.098298) | 0.576940 / 4.584777 (-4.007837) | 2.414088 / 3.745712 (-1.331624) | 2.788423 / 5.269862 (-2.481439) | 1.738695 / 4.565676 (-2.826982) | 0.064456 / 0.424275 (-0.359819) | 0.005536 / 0.007607 (-0.002071) | 0.337266 / 0.226044 (0.111222) | 3.327140 / 2.268929 (1.058212) | 1.837553 / 55.444624 (-53.607072) | 1.538955 / 6.876477 (-5.337521) | 1.575624 / 2.142072 (-0.566448) | 0.639960 / 4.805227 (-4.165267) | 0.117607 / 6.500664 (-6.383057) | 0.042077 / 0.075469 (-0.033393) |\n\n### Benchmark: benchmark_map_filter.json\n\n| metric | filter | map fast-tokenizer batched | map identity | map identity batched | map no-op batched | map no-op batched numpy | map no-op batched pandas | map no-op batched pytorch | map no-op batched tensorflow |\n|--------|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.960488 / 1.841788 (-0.881300) | 11.565280 / 8.074308 (3.490972) | 9.702633 / 10.191392 (-0.488759) | 0.139106 / 0.680424 (-0.541318) | 0.013601 / 0.534201 (-0.520600) | 0.291499 / 0.579283 (-0.287784) | 0.277433 / 0.434364 (-0.156930) | 0.325700 / 0.540337 (-0.214637) | 0.421036 / 1.386936 (-0.965900) |\n\n</details>\nPyArrow==latest\n\n<details>\n<summary>Show updated benchmarks!</summary>\n\n### Benchmark: benchmark_array_xd.json\n\n| metric | read_batch_formatted_as_numpy after write_array2d | read_batch_formatted_as_numpy after write_flattened_sequence | read_batch_formatted_as_numpy after write_nested_sequence | read_batch_unformated after write_array2d | read_batch_unformated after write_flattened_sequence | read_batch_unformated after write_nested_sequence | read_col_formatted_as_numpy after write_array2d | read_col_formatted_as_numpy after write_flattened_sequence | read_col_formatted_as_numpy after write_nested_sequence | read_col_unformated after write_array2d | read_col_unformated after write_flattened_sequence | read_col_unformated after write_nested_sequence | read_formatted_as_numpy after write_array2d | read_formatted_as_numpy after write_flattened_sequence | read_formatted_as_numpy after write_nested_sequence | read_unformated after write_array2d | read_unformated after write_flattened_sequence | read_unformated after write_nested_sequence | write_array2d | write_flattened_sequence | write_nested_sequence |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.005405 / 0.011353 (-0.005948) | 0.003816 / 0.011008 (-0.007192) | 0.050422 / 0.038508 (0.011914) | 0.030473 / 0.023109 (0.007364) | 0.275975 / 0.275898 (0.000077) | 0.298002 / 0.323480 (-0.025478) | 0.004280 / 0.007986 (-0.003706) | 0.002746 / 0.004328 (-0.001583) | 0.049649 / 0.004250 (0.045398) | 0.040675 / 0.037052 (0.003623) | 0.287496 / 0.258489 (0.029007) | 0.315140 / 0.293841 (0.021299) | 0.029835 / 0.128546 (-0.098711) | 0.010443 / 0.075646 (-0.065204) | 0.058299 / 0.419271 (-0.360972) | 0.032944 / 0.043533 (-0.010588) | 0.279468 / 0.255139 (0.024329) | 0.296336 / 0.283200 (0.013136) | 0.018572 / 0.141683 (-0.123111) | 1.177622 / 1.452155 (-0.274532) | 1.238240 / 1.492716 (-0.254477) |\n\n### Benchmark: benchmark_getitem\\_100B.json\n\n| metric | get_batch_of\\_1024\\_random_rows | get_batch_of\\_1024\\_rows | get_first_row | get_last_row |\n|--------|---|---|---|---|\n| new / old (diff) | 0.091867 / 0.018006 (0.073861) | 0.299982 / 0.000490 (0.299492) | 0.000217 / 0.000200 (0.000017) | 0.000043 / 0.000054 (-0.000011) |\n\n### Benchmark: benchmark_indices_mapping.json\n\n| metric | select | shard | shuffle | sort | train_test_split |\n|--------|---|---|---|---|---|\n| new / old (diff) | 0.022649 / 0.037411 (-0.014762) | 0.074948 / 0.014526 (0.060422) | 0.087949 / 0.176557 (-0.088607) | 0.125875 / 0.737135 (-0.611261) | 0.089295 / 0.296338 (-0.207044) |\n\n### Benchmark: benchmark_iterating.json\n\n| metric | read 5000 | read 50000 | read_batch 50000 10 | read_batch 50000 100 | read_batch 50000 1000 | read_formatted numpy 5000 | read_formatted pandas 5000 | read_formatted tensorflow 5000 | read_formatted torch 5000 | read_formatted_batch numpy 5000 10 | read_formatted_batch numpy 5000 1000 | shuffled read 5000 | shuffled read 50000 | shuffled read_batch 50000 10 | shuffled read_batch 50000 100 | shuffled read_batch 50000 1000 | shuffled read_formatted numpy 5000 | shuffled read_formatted_batch numpy 5000 10 | shuffled read_formatted_batch numpy 5000 1000 |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.290387 / 0.215209 (0.075178) | 2.820969 / 2.077655 (0.743315) | 1.614607 / 1.504120 (0.110487) | 1.496959 / 1.541195 (-0.044236) | 1.526475 / 1.468490 (0.057985) | 0.570087 / 4.584777 (-4.014690) | 2.423106 / 3.745712 (-1.322606) | 2.825321 / 5.269862 (-2.444540) | 1.765580 / 4.565676 (-2.800097) | 0.063289 / 0.424275 (-0.360986) | 0.005456 / 0.007607 (-0.002151) | 0.344100 / 0.226044 (0.118055) | 3.395733 / 2.268929 (1.126804) | 1.951794 / 55.444624 (-53.492830) | 1.677689 / 6.876477 (-5.198787) | 1.684448 / 2.142072 (-0.457624) | 0.644343 / 4.805227 (-4.160885) | 0.115796 / 6.500664 (-6.384868) | 0.041052 / 0.075469 (-0.034417) |\n\n### Benchmark: benchmark_map_filter.json\n\n| metric | filter | map fast-tokenizer batched | map identity | map identity batched | map no-op batched | map no-op batched numpy | map no-op batched pandas | map no-op batched pytorch | map no-op batched tensorflow |\n|--------|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 1.031487 / 1.841788 (-0.810301) | 12.116156 / 8.074308 (4.041848) | 10.472247 / 10.191392 (0.280855) | 0.142934 / 0.680424 (-0.537490) | 0.015470 / 0.534201 (-0.518731) | 0.290402 / 0.579283 (-0.288882) | 0.272594 / 0.434364 (-0.161770) | 0.328311 / 0.540337 (-0.212027) | 0.424694 / 1.386936 (-0.962242) |\n\n</details>\n</details>\n\n![](https://cml.dev/watermark.png#8983a3b4dec315bf25331a6065cb74de9017f0e8 \"CML watermark\")\n" ]

[ "Probably best to implement this as a CLI command?", "Thanks for your comment, @mariosasko. Or maybe both (in Python and as CLI command)? The Python command would be just the reverse of `push_to_hub`...\r\n\r\nI am working on a draft implementation, so we can discuss about the API and UX." ]

[ "The docs for this PR live [here](https://moon-ci-docs.huggingface.co/docs/datasets/pr_6809). All of your documentation changes will be reflected on that endpoint. The docs are available until 30 days after the last update.", "@huggingface/datasets once this PR is merged, I would suggest making a release. Do you agree?\r\n- This PR is a follow-up of #6795", "<details>\n<summary>Show benchmarks</summary>\n\nPyArrow==8.0.0\n\n<details>\n<summary>Show updated benchmarks!</summary>\n\n### Benchmark: benchmark_array_xd.json\n\n| metric | read_batch_formatted_as_numpy after write_array2d | read_batch_formatted_as_numpy after write_flattened_sequence | read_batch_formatted_as_numpy after write_nested_sequence | read_batch_unformated after write_array2d | read_batch_unformated after write_flattened_sequence | read_batch_unformated after write_nested_sequence | read_col_formatted_as_numpy after write_array2d | read_col_formatted_as_numpy after write_flattened_sequence | read_col_formatted_as_numpy after write_nested_sequence | read_col_unformated after write_array2d | read_col_unformated after write_flattened_sequence | read_col_unformated after write_nested_sequence | read_formatted_as_numpy after write_array2d | read_formatted_as_numpy after write_flattened_sequence | read_formatted_as_numpy after write_nested_sequence | read_unformated after write_array2d | read_unformated after write_flattened_sequence | read_unformated after write_nested_sequence | write_array2d | write_flattened_sequence | write_nested_sequence |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.004963 / 0.011353 (-0.006390) | 0.003121 / 0.011008 (-0.007888) | 0.063421 / 0.038508 (0.024913) | 0.030727 / 0.023109 (0.007618) | 0.237698 / 0.275898 (-0.038200) | 0.266613 / 0.323480 (-0.056867) | 0.004237 / 0.007986 (-0.003749) | 0.002715 / 0.004328 (-0.001614) | 0.049503 / 0.004250 (0.045253) | 0.043705 / 0.037052 (0.006653) | 0.247818 / 0.258489 (-0.010671) | 0.287545 / 0.293841 (-0.006296) | 0.027232 / 0.128546 (-0.101314) | 0.009952 / 0.075646 (-0.065695) | 0.208678 / 0.419271 (-0.210593) | 0.035494 / 0.043533 (-0.008039) | 0.260900 / 0.255139 (0.005761) | 0.264738 / 0.283200 (-0.018461) | 0.018093 / 0.141683 (-0.123590) | 1.130924 / 1.452155 (-0.321231) | 1.178982 / 1.492716 (-0.313734) |\n\n### Benchmark: benchmark_getitem\\_100B.json\n\n| metric | get_batch_of\\_1024\\_random_rows | get_batch_of\\_1024\\_rows | get_first_row | get_last_row |\n|--------|---|---|---|---|\n| new / old (diff) | 0.094610 / 0.018006 (0.076604) | 0.304674 / 0.000490 (0.304184) | 0.000215 / 0.000200 (0.000015) | 0.000048 / 0.000054 (-0.000007) |\n\n### Benchmark: benchmark_indices_mapping.json\n\n| metric | select | shard | shuffle | sort | train_test_split |\n|--------|---|---|---|---|---|\n| new / old (diff) | 0.018168 / 0.037411 (-0.019243) | 0.062040 / 0.014526 (0.047514) | 0.075634 / 0.176557 (-0.100922) | 0.119488 / 0.737135 (-0.617647) | 0.074790 / 0.296338 (-0.221548) |\n\n### Benchmark: benchmark_iterating.json\n\n| metric | read 5000 | read 50000 | read_batch 50000 10 | read_batch 50000 100 | read_batch 50000 1000 | read_formatted numpy 5000 | read_formatted pandas 5000 | read_formatted tensorflow 5000 | read_formatted torch 5000 | read_formatted_batch numpy 5000 10 | read_formatted_batch numpy 5000 1000 | shuffled read 5000 | shuffled read 50000 | shuffled read_batch 50000 10 | shuffled read_batch 50000 100 | shuffled read_batch 50000 1000 | shuffled read_formatted numpy 5000 | shuffled read_formatted_batch numpy 5000 10 | shuffled read_formatted_batch numpy 5000 1000 |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.282449 / 0.215209 (0.067240) | 2.773231 / 2.077655 (0.695576) | 1.455156 / 1.504120 (-0.048964) | 1.332652 / 1.541195 (-0.208543) | 1.340795 / 1.468490 (-0.127695) | 0.576588 / 4.584777 (-4.008189) | 2.415513 / 3.745712 (-1.330199) | 2.801569 / 5.269862 (-2.468292) | 1.741039 / 4.565676 (-2.824637) | 0.064386 / 0.424275 (-0.359890) | 0.005293 / 0.007607 (-0.002314) | 0.329732 / 0.226044 (0.103688) | 3.227275 / 2.268929 (0.958347) | 1.793121 / 55.444624 (-53.651503) | 1.515115 / 6.876477 (-5.361362) | 1.518738 / 2.142072 (-0.623335) | 0.664465 / 4.805227 (-4.140762) | 0.118813 / 6.500664 (-6.381851) | 0.041715 / 0.075469 (-0.033754) |\n\n### Benchmark: benchmark_map_filter.json\n\n| metric | filter | map fast-tokenizer batched | map identity | map identity batched | map no-op batched | map no-op batched numpy | map no-op batched pandas | map no-op batched pytorch | map no-op batched tensorflow |\n|--------|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.974371 / 1.841788 (-0.867416) | 11.432869 / 8.074308 (3.358561) | 9.607939 / 10.191392 (-0.583453) | 0.143996 / 0.680424 (-0.536427) | 0.014624 / 0.534201 (-0.519577) | 0.286899 / 0.579283 (-0.292384) | 0.265965 / 0.434364 (-0.168399) | 0.324727 / 0.540337 (-0.215611) | 0.420917 / 1.386936 (-0.966019) |\n\n</details>\nPyArrow==latest\n\n<details>\n<summary>Show updated benchmarks!</summary>\n\n### Benchmark: benchmark_array_xd.json\n\n| metric | read_batch_formatted_as_numpy after write_array2d | read_batch_formatted_as_numpy after write_flattened_sequence | read_batch_formatted_as_numpy after write_nested_sequence | read_batch_unformated after write_array2d | read_batch_unformated after write_flattened_sequence | read_batch_unformated after write_nested_sequence | read_col_formatted_as_numpy after write_array2d | read_col_formatted_as_numpy after write_flattened_sequence | read_col_formatted_as_numpy after write_nested_sequence | read_col_unformated after write_array2d | read_col_unformated after write_flattened_sequence | read_col_unformated after write_nested_sequence | read_formatted_as_numpy after write_array2d | read_formatted_as_numpy after write_flattened_sequence | read_formatted_as_numpy after write_nested_sequence | read_unformated after write_array2d | read_unformated after write_flattened_sequence | read_unformated after write_nested_sequence | write_array2d | write_flattened_sequence | write_nested_sequence |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.005145 / 0.011353 (-0.006207) | 0.003723 / 0.011008 (-0.007286) | 0.050387 / 0.038508 (0.011879) | 0.030734 / 0.023109 (0.007625) | 0.274331 / 0.275898 (-0.001567) | 0.295045 / 0.323480 (-0.028435) | 0.004187 / 0.007986 (-0.003799) | 0.002781 / 0.004328 (-0.001547) | 0.049698 / 0.004250 (0.045448) | 0.040049 / 0.037052 (0.002996) | 0.284016 / 0.258489 (0.025527) | 0.309908 / 0.293841 (0.016067) | 0.028994 / 0.128546 (-0.099552) | 0.010625 / 0.075646 (-0.065021) | 0.059305 / 0.419271 (-0.359967) | 0.032982 / 0.043533 (-0.010551) | 0.273342 / 0.255139 (0.018203) | 0.291726 / 0.283200 (0.008527) | 0.018084 / 0.141683 (-0.123599) | 1.136864 / 1.452155 (-0.315290) | 1.163656 / 1.492716 (-0.329061) |\n\n### Benchmark: benchmark_getitem\\_100B.json\n\n| metric | get_batch_of\\_1024\\_random_rows | get_batch_of\\_1024\\_rows | get_first_row | get_last_row |\n|--------|---|---|---|---|\n| new / old (diff) | 0.094868 / 0.018006 (0.076862) | 0.302900 / 0.000490 (0.302410) | 0.000226 / 0.000200 (0.000026) | 0.000053 / 0.000054 (-0.000002) |\n\n### Benchmark: benchmark_indices_mapping.json\n\n| metric | select | shard | shuffle | sort | train_test_split |\n|--------|---|---|---|---|---|\n| new / old (diff) | 0.022142 / 0.037411 (-0.015269) | 0.077457 / 0.014526 (0.062932) | 0.087989 / 0.176557 (-0.088568) | 0.127354 / 0.737135 (-0.609781) | 0.092027 / 0.296338 (-0.204312) |\n\n### Benchmark: benchmark_iterating.json\n\n| metric | read 5000 | read 50000 | read_batch 50000 10 | read_batch 50000 100 | read_batch 50000 1000 | read_formatted numpy 5000 | read_formatted pandas 5000 | read_formatted tensorflow 5000 | read_formatted torch 5000 | read_formatted_batch numpy 5000 10 | read_formatted_batch numpy 5000 1000 | shuffled read 5000 | shuffled read 50000 | shuffled read_batch 50000 10 | shuffled read_batch 50000 100 | shuffled read_batch 50000 1000 | shuffled read_formatted numpy 5000 | shuffled read_formatted_batch numpy 5000 10 | shuffled read_formatted_batch numpy 5000 1000 |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.291196 / 0.215209 (0.075987) | 2.840386 / 2.077655 (0.762731) | 1.571201 / 1.504120 (0.067081) | 1.449429 / 1.541195 (-0.091765) | 1.467189 / 1.468490 (-0.001301) | 0.580991 / 4.584777 (-4.003786) | 2.422566 / 3.745712 (-1.323146) | 2.839621 / 5.269862 (-2.430240) | 1.782987 / 4.565676 (-2.782689) | 0.064765 / 0.424275 (-0.359510) | 0.005338 / 0.007607 (-0.002269) | 0.349148 / 0.226044 (0.123104) | 3.421283 / 2.268929 (1.152355) | 1.943503 / 55.444624 (-53.501122) | 1.653881 / 6.876477 (-5.222596) | 1.698141 / 2.142072 (-0.443931) | 0.667628 / 4.805227 (-4.137599) | 0.118469 / 6.500664 (-6.382195) | 0.041693 / 0.075469 (-0.033776) |\n\n### Benchmark: benchmark_map_filter.json\n\n| metric | filter | map fast-tokenizer batched | map identity | map identity batched | map no-op batched | map no-op batched numpy | map no-op batched pandas | map no-op batched pytorch | map no-op batched tensorflow |\n|--------|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 1.026385 / 1.841788 (-0.815403) | 12.225049 / 8.074308 (4.150741) | 10.363072 / 10.191392 (0.171680) | 0.142682 / 0.680424 (-0.537742) | 0.015698 / 0.534201 (-0.518502) | 0.288148 / 0.579283 (-0.291135) | 0.272639 / 0.434364 (-0.161724) | 0.325305 / 0.540337 (-0.215032) | 0.421395 / 1.386936 (-0.965541) |\n\n</details>\n</details>\n\n![](https://cml.dev/watermark.png#2a14271263da2fda9f966af41c7bd885bfa42256 \"CML watermark\")\n" ]

[]

[ "The docs for this PR live [here](https://moon-ci-docs.huggingface.co/docs/datasets/pr_6806). All of your documentation changes will be reflected on that endpoint. The docs are available until 30 days after the last update.", "<details>\n<summary>Show benchmarks</summary>\n\nPyArrow==8.0.0\n\n<details>\n<summary>Show updated benchmarks!</summary>\n\n### Benchmark: benchmark_array_xd.json\n\n| metric | read_batch_formatted_as_numpy after write_array2d | read_batch_formatted_as_numpy after write_flattened_sequence | read_batch_formatted_as_numpy after write_nested_sequence | read_batch_unformated after write_array2d | read_batch_unformated after write_flattened_sequence | read_batch_unformated after write_nested_sequence | read_col_formatted_as_numpy after write_array2d | read_col_formatted_as_numpy after write_flattened_sequence | read_col_formatted_as_numpy after write_nested_sequence | read_col_unformated after write_array2d | read_col_unformated after write_flattened_sequence | read_col_unformated after write_nested_sequence | read_formatted_as_numpy after write_array2d | read_formatted_as_numpy after write_flattened_sequence | read_formatted_as_numpy after write_nested_sequence | read_unformated after write_array2d | read_unformated after write_flattened_sequence | read_unformated after write_nested_sequence | write_array2d | write_flattened_sequence | write_nested_sequence |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.005068 / 0.011353 (-0.006285) | 0.003613 / 0.011008 (-0.007395) | 0.063226 / 0.038508 (0.024718) | 0.030653 / 0.023109 (0.007544) | 0.243981 / 0.275898 (-0.031918) | 0.268596 / 0.323480 (-0.054884) | 0.003109 / 0.007986 (-0.004876) | 0.003292 / 0.004328 (-0.001036) | 0.048857 / 0.004250 (0.044606) | 0.043929 / 0.037052 (0.006876) | 0.264002 / 0.258489 (0.005513) | 0.289028 / 0.293841 (-0.004813) | 0.028053 / 0.128546 (-0.100493) | 0.010837 / 0.075646 (-0.064809) | 0.208084 / 0.419271 (-0.211188) | 0.035592 / 0.043533 (-0.007941) | 0.252639 / 0.255139 (-0.002500) | 0.267599 / 0.283200 (-0.015600) | 0.018097 / 0.141683 (-0.123585) | 1.150811 / 1.452155 (-0.301344) | 1.219449 / 1.492716 (-0.273267) |\n\n### Benchmark: benchmark_getitem\\_100B.json\n\n| metric | get_batch_of\\_1024\\_random_rows | get_batch_of\\_1024\\_rows | get_first_row | get_last_row |\n|--------|---|---|---|---|\n| new / old (diff) | 0.095427 / 0.018006 (0.077421) | 0.307270 / 0.000490 (0.306781) | 0.000218 / 0.000200 (0.000018) | 0.000043 / 0.000054 (-0.000011) |\n\n### Benchmark: benchmark_indices_mapping.json\n\n| metric | select | shard | shuffle | sort | train_test_split |\n|--------|---|---|---|---|---|\n| new / old (diff) | 0.018713 / 0.037411 (-0.018698) | 0.065238 / 0.014526 (0.050712) | 0.074650 / 0.176557 (-0.101906) | 0.120130 / 0.737135 (-0.617005) | 0.078457 / 0.296338 (-0.217882) |\n\n### Benchmark: benchmark_iterating.json\n\n| metric | read 5000 | read 50000 | read_batch 50000 10 | read_batch 50000 100 | read_batch 50000 1000 | read_formatted numpy 5000 | read_formatted pandas 5000 | read_formatted tensorflow 5000 | read_formatted torch 5000 | read_formatted_batch numpy 5000 10 | read_formatted_batch numpy 5000 1000 | shuffled read 5000 | shuffled read 50000 | shuffled read_batch 50000 10 | shuffled read_batch 50000 100 | shuffled read_batch 50000 1000 | shuffled read_formatted numpy 5000 | shuffled read_formatted_batch numpy 5000 10 | shuffled read_formatted_batch numpy 5000 1000 |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.283666 / 0.215209 (0.068457) | 2.852818 / 2.077655 (0.775163) | 1.459790 / 1.504120 (-0.044330) | 1.326732 / 1.541195 (-0.214463) | 1.373530 / 1.468490 (-0.094960) | 0.579136 / 4.584777 (-4.005641) | 2.388369 / 3.745712 (-1.357343) | 2.813786 / 5.269862 (-2.456075) | 1.730079 / 4.565676 (-2.835597) | 0.063445 / 0.424275 (-0.360831) | 0.005355 / 0.007607 (-0.002252) | 0.340169 / 0.226044 (0.114124) | 3.391220 / 2.268929 (1.122291) | 1.838003 / 55.444624 (-53.606621) | 1.523518 / 6.876477 (-5.352959) | 1.574007 / 2.142072 (-0.568065) | 0.650265 / 4.805227 (-4.154962) | 0.117114 / 6.500664 (-6.383550) | 0.042430 / 0.075469 (-0.033039) |\n\n### Benchmark: benchmark_map_filter.json\n\n| metric | filter | map fast-tokenizer batched | map identity | map identity batched | map no-op batched | map no-op batched numpy | map no-op batched pandas | map no-op batched pytorch | map no-op batched tensorflow |\n|--------|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.955596 / 1.841788 (-0.886191) | 11.546544 / 8.074308 (3.472236) | 9.593613 / 10.191392 (-0.597779) | 0.141502 / 0.680424 (-0.538922) | 0.014251 / 0.534201 (-0.519950) | 0.293825 / 0.579283 (-0.285458) | 0.263088 / 0.434364 (-0.171276) | 0.325035 / 0.540337 (-0.215302) | 0.419372 / 1.386936 (-0.967564) |\n\n</details>\nPyArrow==latest\n\n<details>\n<summary>Show updated benchmarks!</summary>\n\n### Benchmark: benchmark_array_xd.json\n\n| metric | read_batch_formatted_as_numpy after write_array2d | read_batch_formatted_as_numpy after write_flattened_sequence | read_batch_formatted_as_numpy after write_nested_sequence | read_batch_unformated after write_array2d | read_batch_unformated after write_flattened_sequence | read_batch_unformated after write_nested_sequence | read_col_formatted_as_numpy after write_array2d | read_col_formatted_as_numpy after write_flattened_sequence | read_col_formatted_as_numpy after write_nested_sequence | read_col_unformated after write_array2d | read_col_unformated after write_flattened_sequence | read_col_unformated after write_nested_sequence | read_formatted_as_numpy after write_array2d | read_formatted_as_numpy after write_flattened_sequence | read_formatted_as_numpy after write_nested_sequence | read_unformated after write_array2d | read_unformated after write_flattened_sequence | read_unformated after write_nested_sequence | write_array2d | write_flattened_sequence | write_nested_sequence |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.005567 / 0.011353 (-0.005785) | 0.003670 / 0.011008 (-0.007338) | 0.050338 / 0.038508 (0.011830) | 0.031730 / 0.023109 (0.008621) | 0.278307 / 0.275898 (0.002409) | 0.303170 / 0.323480 (-0.020310) | 0.004276 / 0.007986 (-0.003709) | 0.002720 / 0.004328 (-0.001609) | 0.048675 / 0.004250 (0.044425) | 0.041026 / 0.037052 (0.003974) | 0.291353 / 0.258489 (0.032864) | 0.318487 / 0.293841 (0.024646) | 0.029676 / 0.128546 (-0.098870) | 0.010428 / 0.075646 (-0.065218) | 0.057443 / 0.419271 (-0.361828) | 0.032735 / 0.043533 (-0.010798) | 0.282900 / 0.255139 (0.027761) | 0.297539 / 0.283200 (0.014339) | 0.018237 / 0.141683 (-0.123446) | 1.188047 / 1.452155 (-0.264107) | 1.223283 / 1.492716 (-0.269433) |\n\n### Benchmark: benchmark_getitem\\_100B.json\n\n| metric | get_batch_of\\_1024\\_random_rows | get_batch_of\\_1024\\_rows | get_first_row | get_last_row |\n|--------|---|---|---|---|\n| new / old (diff) | 0.090629 / 0.018006 (0.072623) | 0.300898 / 0.000490 (0.300408) | 0.000212 / 0.000200 (0.000012) | 0.000133 / 0.000054 (0.000078) |\n\n### Benchmark: benchmark_indices_mapping.json\n\n| metric | select | shard | shuffle | sort | train_test_split |\n|--------|---|---|---|---|---|\n| new / old (diff) | 0.022200 / 0.037411 (-0.015211) | 0.075310 / 0.014526 (0.060784) | 0.086790 / 0.176557 (-0.089766) | 0.127392 / 0.737135 (-0.609744) | 0.088435 / 0.296338 (-0.207903) |\n\n### Benchmark: benchmark_iterating.json\n\n| metric | read 5000 | read 50000 | read_batch 50000 10 | read_batch 50000 100 | read_batch 50000 1000 | read_formatted numpy 5000 | read_formatted pandas 5000 | read_formatted tensorflow 5000 | read_formatted torch 5000 | read_formatted_batch numpy 5000 10 | read_formatted_batch numpy 5000 1000 | shuffled read 5000 | shuffled read 50000 | shuffled read_batch 50000 10 | shuffled read_batch 50000 100 | shuffled read_batch 50000 1000 | shuffled read_formatted numpy 5000 | shuffled read_formatted_batch numpy 5000 10 | shuffled read_formatted_batch numpy 5000 1000 |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.301308 / 0.215209 (0.086099) | 2.963126 / 2.077655 (0.885471) | 1.639604 / 1.504120 (0.135484) | 1.508776 / 1.541195 (-0.032419) | 1.553280 / 1.468490 (0.084789) | 0.567256 / 4.584777 (-4.017520) | 2.445231 / 3.745712 (-1.300482) | 2.884071 / 5.269862 (-2.385791) | 1.777321 / 4.565676 (-2.788355) | 0.063659 / 0.424275 (-0.360616) | 0.005435 / 0.007607 (-0.002172) | 0.361786 / 0.226044 (0.135742) | 3.624264 / 2.268929 (1.355335) | 2.022661 / 55.444624 (-53.421963) | 1.740581 / 6.876477 (-5.135896) | 1.748503 / 2.142072 (-0.393570) | 0.660783 / 4.805227 (-4.144444) | 0.118045 / 6.500664 (-6.382619) | 0.040940 / 0.075469 (-0.034529) |\n\n### Benchmark: benchmark_map_filter.json\n\n| metric | filter | map fast-tokenizer batched | map identity | map identity batched | map no-op batched | map no-op batched numpy | map no-op batched pandas | map no-op batched pytorch | map no-op batched tensorflow |\n|--------|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 1.015614 / 1.841788 (-0.826174) | 12.094985 / 8.074308 (4.020677) | 10.435581 / 10.191392 (0.244189) | 0.140239 / 0.680424 (-0.540185) | 0.014992 / 0.534201 (-0.519209) | 0.290549 / 0.579283 (-0.288735) | 0.274718 / 0.434364 (-0.159645) | 0.334783 / 0.540337 (-0.205554) | 0.426540 / 1.386936 (-0.960396) |\n\n</details>\n</details>\n\n![](https://cml.dev/watermark.png#828aff908450ac7af3a1820bb2eb7b438f2692f5 \"CML watermark\")\n" ]

[ "This seems to be hardcoded behavior in table.py `array_cast`.\r\n```python\r\nif (\r\n not allow_number_to_str\r\n and pa.types.is_string(pa_type)\r\n and (pa.types.is_floating(array.type) or pa.types.is_integer(array.type))\r\n ):\r\n raise TypeError(\r\n f\"Couldn't cast array of type {array.type} to {pa_type} since allow_number_to_str is set to {allow_number_to_str}\"\r\n )\r\n if pa.types.is_null(pa_type) and not pa.types.is_null(array.type):\r\n raise TypeError(f\"Couldn't cast array of type {array.type} to {pa_type}\")\r\n return array.cast(pa_type)\r\n```\r\nwhere floats and integers are not cast to string but booleans are.\r\nMaybe this should be extended to booleans?", "Thanks for reporting! @Modexus Do you want to open a PR with the suggested fix?", "I'll gladly create a PR but not sure what the behavior should be.\r\n\r\nShould a value returned from map be cast to the current feature?\r\nAt the moment this seems very inconsistent since `datetime `is also cast (this would only fix `boolean`) but nested structures are not.\r\n\r\n```python\r\ndset = Dataset.from_dict({\"a\": [\"Hello world!\"]})\r\ndset = dset.map(lambda x: {\"a\": date(2021, 1, 1)})\r\n# dset[0][\"a\"] == '2021-01-01'\r\n```\r\n```python\r\ndset = Dataset.from_dict({\"a\": [\"Hello world!\"]})\r\ndset = dset.map(lambda x: {\"a\": [True]})\r\n# dset[0][\"a\"] == [True]\r\n```\r\n\r\nIs there are reason to cast the value if the user doesn't specify it explicitly?\r\nSeems tricky that some things are cast and some are not.", "Indeed, it also makes sense to raise a `TypeError` for temporal and decimal types.\r\n\r\n> Is there are reason to cast the value if the user doesn't specify it explicitly?\r\n\r\nThis is how PyArrow's built-in `cast` behaves - it allows casting from primitive types to strings. Hence, we need `allow_number_to_str` to disallow such casts (e.g., in the [scenario](https://github.com/huggingface/datasets/blob/a3bc89d8bfd47c2a175c3ce16d92b7307cdeafd6/src/datasets/arrow_writer.py#L208) when we are \"trying a type\" to preserve the original type if there is a column in the output dataset with the same name as in the input one).\r\n\r\nPS: In the PR, we can introduce `allow_numeric_to_str` (for floats, integers, decimals, booleans) and `allow_temporal_to_str` (for dates, timestamps, ...) and deprecate `allow_number_to_str` to make it clear what each parameter does.", "Would just `allow_primitive_to_str` work?\r\nThis should include all `numeric`, `boolean `and `temporal`formats.\r\n\r\nNote that at least in the [ C++ implementation](https://arrow.apache.org/docs/cpp/api/utilities.html#_CPPv410is_numericRK8DataType) `numeric `seems to exclude `boolean`.\r\n[](https://arrow.apache.org/docs/cpp/api/utilities.html#_CPPv410is_numericRK8DataType)", "Indeed, `allow_primitive_to_str` sounds better.\r\n\r\nPS: PyArrow's `pa.types.is_primitive` returns `False` for decimal types, but I think is okay for us to treat decimals as primitive types (or we can have `allow_decimal_to_str` to be fully consistent with PyArrow)", "Fixed by:\r\n- #6811" ]

[ "The docs for this PR live [here](https://moon-ci-docs.huggingface.co/docs/datasets/pr_6804). All of your documentation changes will be reflected on that endpoint. The docs are available until 30 days after the last update.", "<details>\n<summary>Show benchmarks</summary>\n\nPyArrow==8.0.0\n\n<details>\n<summary>Show updated benchmarks!</summary>\n\n### Benchmark: benchmark_array_xd.json\n\n| metric | read_batch_formatted_as_numpy after write_array2d | read_batch_formatted_as_numpy after write_flattened_sequence | read_batch_formatted_as_numpy after write_nested_sequence | read_batch_unformated after write_array2d | read_batch_unformated after write_flattened_sequence | read_batch_unformated after write_nested_sequence | read_col_formatted_as_numpy after write_array2d | read_col_formatted_as_numpy after write_flattened_sequence | read_col_formatted_as_numpy after write_nested_sequence | read_col_unformated after write_array2d | read_col_unformated after write_flattened_sequence | read_col_unformated after write_nested_sequence | read_formatted_as_numpy after write_array2d | read_formatted_as_numpy after write_flattened_sequence | read_formatted_as_numpy after write_nested_sequence | read_unformated after write_array2d | read_unformated after write_flattened_sequence | read_unformated after write_nested_sequence | write_array2d | write_flattened_sequence | write_nested_sequence |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.005222 / 0.011353 (-0.006131) | 0.003306 / 0.011008 (-0.007702) | 0.063326 / 0.038508 (0.024818) | 0.031371 / 0.023109 (0.008261) | 0.244947 / 0.275898 (-0.030951) | 0.264141 / 0.323480 (-0.059339) | 0.004186 / 0.007986 (-0.003800) | 0.002676 / 0.004328 (-0.001653) | 0.048690 / 0.004250 (0.044440) | 0.045172 / 0.037052 (0.008120) | 0.256597 / 0.258489 (-0.001892) | 0.284348 / 0.293841 (-0.009493) | 0.026855 / 0.128546 (-0.101691) | 0.009947 / 0.075646 (-0.065699) | 0.206311 / 0.419271 (-0.212961) | 0.035178 / 0.043533 (-0.008355) | 0.251501 / 0.255139 (-0.003638) | 0.261314 / 0.283200 (-0.021886) | 0.018000 / 0.141683 (-0.123683) | 1.144588 / 1.452155 (-0.307566) | 1.193627 / 1.492716 (-0.299089) |\n\n### Benchmark: benchmark_getitem\\_100B.json\n\n| metric | get_batch_of\\_1024\\_random_rows | get_batch_of\\_1024\\_rows | get_first_row | get_last_row |\n|--------|---|---|---|---|\n| new / old (diff) | 0.091629 / 0.018006 (0.073623) | 0.298959 / 0.000490 (0.298469) | 0.000207 / 0.000200 (0.000007) | 0.000042 / 0.000054 (-0.000012) |\n\n### Benchmark: benchmark_indices_mapping.json\n\n| metric | select | shard | shuffle | sort | train_test_split |\n|--------|---|---|---|---|---|\n| new / old (diff) | 0.018053 / 0.037411 (-0.019358) | 0.061280 / 0.014526 (0.046754) | 0.074138 / 0.176557 (-0.102419) | 0.119048 / 0.737135 (-0.618088) | 0.074572 / 0.296338 (-0.221767) |\n\n### Benchmark: benchmark_iterating.json\n\n| metric | read 5000 | read 50000 | read_batch 50000 10 | read_batch 50000 100 | read_batch 50000 1000 | read_formatted numpy 5000 | read_formatted pandas 5000 | read_formatted tensorflow 5000 | read_formatted torch 5000 | read_formatted_batch numpy 5000 10 | read_formatted_batch numpy 5000 1000 | shuffled read 5000 | shuffled read 50000 | shuffled read_batch 50000 10 | shuffled read_batch 50000 100 | shuffled read_batch 50000 1000 | shuffled read_formatted numpy 5000 | shuffled read_formatted_batch numpy 5000 10 | shuffled read_formatted_batch numpy 5000 1000 |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.282440 / 0.215209 (0.067231) | 2.762017 / 2.077655 (0.684362) | 1.474452 / 1.504120 (-0.029668) | 1.361489 / 1.541195 (-0.179706) | 1.359696 / 1.468490 (-0.108795) | 0.569640 / 4.584777 (-4.015137) | 2.398098 / 3.745712 (-1.347614) | 2.731399 / 5.269862 (-2.538462) | 1.697432 / 4.565676 (-2.868245) | 0.063330 / 0.424275 (-0.360945) | 0.005416 / 0.007607 (-0.002191) | 0.346510 / 0.226044 (0.120465) | 3.276473 / 2.268929 (1.007544) | 1.837605 / 55.444624 (-53.607019) | 1.538654 / 6.876477 (-5.337822) | 1.553943 / 2.142072 (-0.588129) | 0.640571 / 4.805227 (-4.164657) | 0.116736 / 6.500664 (-6.383928) | 0.041701 / 0.075469 (-0.033768) |\n\n### Benchmark: benchmark_map_filter.json\n\n| metric | filter | map fast-tokenizer batched | map identity | map identity batched | map no-op batched | map no-op batched numpy | map no-op batched pandas | map no-op batched pytorch | map no-op batched tensorflow |\n|--------|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.975846 / 1.841788 (-0.865942) | 11.151727 / 8.074308 (3.077419) | 9.436281 / 10.191392 (-0.755111) | 0.141027 / 0.680424 (-0.539397) | 0.014389 / 0.534201 (-0.519812) | 0.285575 / 0.579283 (-0.293708) | 0.263753 / 0.434364 (-0.170610) | 0.321893 / 0.540337 (-0.218444) | 0.420280 / 1.386936 (-0.966656) |\n\n</details>\nPyArrow==latest\n\n<details>\n<summary>Show updated benchmarks!</summary>\n\n### Benchmark: benchmark_array_xd.json\n\n| metric | read_batch_formatted_as_numpy after write_array2d | read_batch_formatted_as_numpy after write_flattened_sequence | read_batch_formatted_as_numpy after write_nested_sequence | read_batch_unformated after write_array2d | read_batch_unformated after write_flattened_sequence | read_batch_unformated after write_nested_sequence | read_col_formatted_as_numpy after write_array2d | read_col_formatted_as_numpy after write_flattened_sequence | read_col_formatted_as_numpy after write_nested_sequence | read_col_unformated after write_array2d | read_col_unformated after write_flattened_sequence | read_col_unformated after write_nested_sequence | read_formatted_as_numpy after write_array2d | read_formatted_as_numpy after write_flattened_sequence | read_formatted_as_numpy after write_nested_sequence | read_unformated after write_array2d | read_unformated after write_flattened_sequence | read_unformated after write_nested_sequence | write_array2d | write_flattened_sequence | write_nested_sequence |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.005148 / 0.011353 (-0.006205) | 0.003264 / 0.011008 (-0.007744) | 0.049828 / 0.038508 (0.011320) | 0.031234 / 0.023109 (0.008125) | 0.271079 / 0.275898 (-0.004819) | 0.295256 / 0.323480 (-0.028224) | 0.004128 / 0.007986 (-0.003857) | 0.002637 / 0.004328 (-0.001692) | 0.048145 / 0.004250 (0.043895) | 0.039691 / 0.037052 (0.002638) | 0.287229 / 0.258489 (0.028740) | 0.310477 / 0.293841 (0.016636) | 0.028936 / 0.128546 (-0.099610) | 0.010392 / 0.075646 (-0.065254) | 0.057774 / 0.419271 (-0.361497) | 0.032557 / 0.043533 (-0.010975) | 0.275146 / 0.255139 (0.020007) | 0.291283 / 0.283200 (0.008084) | 0.017724 / 0.141683 (-0.123958) | 1.186831 / 1.452155 (-0.265324) | 1.220086 / 1.492716 (-0.272630) |\n\n### Benchmark: benchmark_getitem\\_100B.json\n\n| metric | get_batch_of\\_1024\\_random_rows | get_batch_of\\_1024\\_rows | get_first_row | get_last_row |\n|--------|---|---|---|---|\n| new / old (diff) | 0.093575 / 0.018006 (0.075569) | 0.297198 / 0.000490 (0.296709) | 0.000216 / 0.000200 (0.000016) | 0.000044 / 0.000054 (-0.000010) |\n\n### Benchmark: benchmark_indices_mapping.json\n\n| metric | select | shard | shuffle | sort | train_test_split |\n|--------|---|---|---|---|---|\n| new / old (diff) | 0.021683 / 0.037411 (-0.015728) | 0.075347 / 0.014526 (0.060821) | 0.085453 / 0.176557 (-0.091103) | 0.125422 / 0.737135 (-0.611713) | 0.087185 / 0.296338 (-0.209153) |\n\n### Benchmark: benchmark_iterating.json\n\n| metric | read 5000 | read 50000 | read_batch 50000 10 | read_batch 50000 100 | read_batch 50000 1000 | read_formatted numpy 5000 | read_formatted pandas 5000 | read_formatted tensorflow 5000 | read_formatted torch 5000 | read_formatted_batch numpy 5000 10 | read_formatted_batch numpy 5000 1000 | shuffled read 5000 | shuffled read 50000 | shuffled read_batch 50000 10 | shuffled read_batch 50000 100 | shuffled read_batch 50000 1000 | shuffled read_formatted numpy 5000 | shuffled read_formatted_batch numpy 5000 10 | shuffled read_formatted_batch numpy 5000 1000 |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.301520 / 0.215209 (0.086311) | 2.951614 / 2.077655 (0.873959) | 1.659897 / 1.504120 (0.155777) | 1.528097 / 1.541195 (-0.013097) | 1.552031 / 1.468490 (0.083541) | 0.576297 / 4.584777 (-4.008480) | 2.492349 / 3.745712 (-1.253363) | 2.805999 / 5.269862 (-2.463862) | 1.757556 / 4.565676 (-2.808121) | 0.064940 / 0.424275 (-0.359335) | 0.005314 / 0.007607 (-0.002293) | 0.358838 / 0.226044 (0.132793) | 3.576890 / 2.268929 (1.307961) | 2.030788 / 55.444624 (-53.413837) | 1.743650 / 6.876477 (-5.132826) | 1.745229 / 2.142072 (-0.396844) | 0.647840 / 4.805227 (-4.157387) | 0.116637 / 6.500664 (-6.384027) | 0.040555 / 0.075469 (-0.034915) |\n\n### Benchmark: benchmark_map_filter.json\n\n| metric | filter | map fast-tokenizer batched | map identity | map identity batched | map no-op batched | map no-op batched numpy | map no-op batched pandas | map no-op batched pytorch | map no-op batched tensorflow |\n|--------|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 1.009130 / 1.841788 (-0.832657) | 11.951145 / 8.074308 (3.876836) | 9.968355 / 10.191392 (-0.223037) | 0.139959 / 0.680424 (-0.540465) | 0.015985 / 0.534201 (-0.518216) | 0.286594 / 0.579283 (-0.292689) | 0.275805 / 0.434364 (-0.158559) | 0.328484 / 0.540337 (-0.211854) | 0.419818 / 1.386936 (-0.967118) |\n\n</details>\n</details>\n\n![](https://cml.dev/watermark.png#89a58cdfc59ecc83662a47b638cf82a5b99f4a48 \"CML watermark\")\n" ]

[ "The docs for this PR live [here](https://moon-ci-docs.huggingface.co/docs/datasets/pr_6803). All of your documentation changes will be reflected on that endpoint. The docs are available until 30 days after the last update.", "CI failures are unrelated.", "<details>\n<summary>Show benchmarks</summary>\n\nPyArrow==8.0.0\n\n<details>\n<summary>Show updated benchmarks!</summary>\n\n### Benchmark: benchmark_array_xd.json\n\n| metric | read_batch_formatted_as_numpy after write_array2d | read_batch_formatted_as_numpy after write_flattened_sequence | read_batch_formatted_as_numpy after write_nested_sequence | read_batch_unformated after write_array2d | read_batch_unformated after write_flattened_sequence | read_batch_unformated after write_nested_sequence | read_col_formatted_as_numpy after write_array2d | read_col_formatted_as_numpy after write_flattened_sequence | read_col_formatted_as_numpy after write_nested_sequence | read_col_unformated after write_array2d | read_col_unformated after write_flattened_sequence | read_col_unformated after write_nested_sequence | read_formatted_as_numpy after write_array2d | read_formatted_as_numpy after write_flattened_sequence | read_formatted_as_numpy after write_nested_sequence | read_unformated after write_array2d | read_unformated after write_flattened_sequence | read_unformated after write_nested_sequence | write_array2d | write_flattened_sequence | write_nested_sequence |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.005063 / 0.011353 (-0.006290) | 0.003598 / 0.011008 (-0.007410) | 0.062929 / 0.038508 (0.024421) | 0.031723 / 0.023109 (0.008614) | 0.246503 / 0.275898 (-0.029395) | 0.268742 / 0.323480 (-0.054738) | 0.003249 / 0.007986 (-0.004737) | 0.002613 / 0.004328 (-0.001715) | 0.049001 / 0.004250 (0.044751) | 0.045740 / 0.037052 (0.008687) | 0.261182 / 0.258489 (0.002693) | 0.297328 / 0.293841 (0.003487) | 0.026925 / 0.128546 (-0.101621) | 0.010588 / 0.075646 (-0.065059) | 0.208954 / 0.419271 (-0.210317) | 0.035286 / 0.043533 (-0.008246) | 0.277678 / 0.255139 (0.022539) | 0.269313 / 0.283200 (-0.013887) | 0.019865 / 0.141683 (-0.121818) | 1.145883 / 1.452155 (-0.306272) | 1.196766 / 1.492716 (-0.295950) |\n\n### Benchmark: benchmark_getitem\\_100B.json\n\n| metric | get_batch_of\\_1024\\_random_rows | get_batch_of\\_1024\\_rows | get_first_row | get_last_row |\n|--------|---|---|---|---|\n| new / old (diff) | 0.093886 / 0.018006 (0.075879) | 0.305118 / 0.000490 (0.304629) | 0.000207 / 0.000200 (0.000008) | 0.000048 / 0.000054 (-0.000006) |\n\n### Benchmark: benchmark_indices_mapping.json\n\n| metric | select | shard | shuffle | sort | train_test_split |\n|--------|---|---|---|---|---|\n| new / old (diff) | 0.018473 / 0.037411 (-0.018938) | 0.061719 / 0.014526 (0.047193) | 0.074980 / 0.176557 (-0.101577) | 0.122354 / 0.737135 (-0.614781) | 0.076111 / 0.296338 (-0.220227) |\n\n### Benchmark: benchmark_iterating.json\n\n| metric | read 5000 | read 50000 | read_batch 50000 10 | read_batch 50000 100 | read_batch 50000 1000 | read_formatted numpy 5000 | read_formatted pandas 5000 | read_formatted tensorflow 5000 | read_formatted torch 5000 | read_formatted_batch numpy 5000 10 | read_formatted_batch numpy 5000 1000 | shuffled read 5000 | shuffled read 50000 | shuffled read_batch 50000 10 | shuffled read_batch 50000 100 | shuffled read_batch 50000 1000 | shuffled read_formatted numpy 5000 | shuffled read_formatted_batch numpy 5000 10 | shuffled read_formatted_batch numpy 5000 1000 |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.280222 / 0.215209 (0.065013) | 2.692820 / 2.077655 (0.615165) | 1.440897 / 1.504120 (-0.063223) | 1.313829 / 1.541195 (-0.227366) | 1.324392 / 1.468490 (-0.144098) | 0.570114 / 4.584777 (-4.014662) | 2.373946 / 3.745712 (-1.371766) | 2.804485 / 5.269862 (-2.465377) | 1.753595 / 4.565676 (-2.812081) | 0.062660 / 0.424275 (-0.361615) | 0.005267 / 0.007607 (-0.002340) | 0.323108 / 0.226044 (0.097063) | 3.257302 / 2.268929 (0.988373) | 1.802613 / 55.444624 (-53.642011) | 1.510590 / 6.876477 (-5.365886) | 1.567452 / 2.142072 (-0.574621) | 0.649872 / 4.805227 (-4.155355) | 0.117245 / 6.500664 (-6.383419) | 0.042260 / 0.075469 (-0.033209) |\n\n### Benchmark: benchmark_map_filter.json\n\n| metric | filter | map fast-tokenizer batched | map identity | map identity batched | map no-op batched | map no-op batched numpy | map no-op batched pandas | map no-op batched pytorch | map no-op batched tensorflow |\n|--------|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.976068 / 1.841788 (-0.865720) | 11.565981 / 8.074308 (3.491672) | 9.598650 / 10.191392 (-0.592742) | 0.129903 / 0.680424 (-0.550520) | 0.014925 / 0.534201 (-0.519276) | 0.290732 / 0.579283 (-0.288551) | 0.271236 / 0.434364 (-0.163128) | 0.325450 / 0.540337 (-0.214888) | 0.420218 / 1.386936 (-0.966718) |\n\n</details>\nPyArrow==latest\n\n<details>\n<summary>Show updated benchmarks!</summary>\n\n### Benchmark: benchmark_array_xd.json\n\n| metric | read_batch_formatted_as_numpy after write_array2d | read_batch_formatted_as_numpy after write_flattened_sequence | read_batch_formatted_as_numpy after write_nested_sequence | read_batch_unformated after write_array2d | read_batch_unformated after write_flattened_sequence | read_batch_unformated after write_nested_sequence | read_col_formatted_as_numpy after write_array2d | read_col_formatted_as_numpy after write_flattened_sequence | read_col_formatted_as_numpy after write_nested_sequence | read_col_unformated after write_array2d | read_col_unformated after write_flattened_sequence | read_col_unformated after write_nested_sequence | read_formatted_as_numpy after write_array2d | read_formatted_as_numpy after write_flattened_sequence | read_formatted_as_numpy after write_nested_sequence | read_unformated after write_array2d | read_unformated after write_flattened_sequence | read_unformated after write_nested_sequence | write_array2d | write_flattened_sequence | write_nested_sequence |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.005404 / 0.011353 (-0.005949) | 0.003710 / 0.011008 (-0.007298) | 0.050982 / 0.038508 (0.012474) | 0.031340 / 0.023109 (0.008231) | 0.279221 / 0.275898 (0.003323) | 0.300936 / 0.323480 (-0.022544) | 0.004251 / 0.007986 (-0.003735) | 0.002697 / 0.004328 (-0.001631) | 0.049335 / 0.004250 (0.045085) | 0.040979 / 0.037052 (0.003926) | 0.287121 / 0.258489 (0.028632) | 0.315100 / 0.293841 (0.021259) | 0.029093 / 0.128546 (-0.099454) | 0.010618 / 0.075646 (-0.065028) | 0.059095 / 0.419271 (-0.360177) | 0.032953 / 0.043533 (-0.010580) | 0.274861 / 0.255139 (0.019722) | 0.292284 / 0.283200 (0.009085) | 0.017882 / 0.141683 (-0.123801) | 1.150590 / 1.452155 (-0.301565) | 1.203501 / 1.492716 (-0.289215) |\n\n### Benchmark: benchmark_getitem\\_100B.json\n\n| metric | get_batch_of\\_1024\\_random_rows | get_batch_of\\_1024\\_rows | get_first_row | get_last_row |\n|--------|---|---|---|---|\n| new / old (diff) | 0.096868 / 0.018006 (0.078862) | 0.306460 / 0.000490 (0.305971) | 0.000230 / 0.000200 (0.000030) | 0.000058 / 0.000054 (0.000003) |\n\n### Benchmark: benchmark_indices_mapping.json\n\n| metric | select | shard | shuffle | sort | train_test_split |\n|--------|---|---|---|---|---|\n| new / old (diff) | 0.022031 / 0.037411 (-0.015381) | 0.074847 / 0.014526 (0.060321) | 0.086951 / 0.176557 (-0.089606) | 0.125706 / 0.737135 (-0.611429) | 0.088244 / 0.296338 (-0.208094) |\n\n### Benchmark: benchmark_iterating.json\n\n| metric | read 5000 | read 50000 | read_batch 50000 10 | read_batch 50000 100 | read_batch 50000 1000 | read_formatted numpy 5000 | read_formatted pandas 5000 | read_formatted tensorflow 5000 | read_formatted torch 5000 | read_formatted_batch numpy 5000 10 | read_formatted_batch numpy 5000 1000 | shuffled read 5000 | shuffled read 50000 | shuffled read_batch 50000 10 | shuffled read_batch 50000 100 | shuffled read_batch 50000 1000 | shuffled read_formatted numpy 5000 | shuffled read_formatted_batch numpy 5000 10 | shuffled read_formatted_batch numpy 5000 1000 |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.297861 / 0.215209 (0.082652) | 2.923172 / 2.077655 (0.845518) | 1.628511 / 1.504120 (0.124391) | 1.499907 / 1.541195 (-0.041288) | 1.490060 / 1.468490 (0.021570) | 0.564087 / 4.584777 (-4.020690) | 2.441201 / 3.745712 (-1.304511) | 2.805283 / 5.269862 (-2.464578) | 1.762703 / 4.565676 (-2.802974) | 0.063038 / 0.424275 (-0.361237) | 0.005276 / 0.007607 (-0.002331) | 0.343413 / 0.226044 (0.117369) | 3.400858 / 2.268929 (1.131930) | 2.039937 / 55.444624 (-53.404687) | 1.674622 / 6.876477 (-5.201855) | 1.688371 / 2.142072 (-0.453702) | 0.635321 / 4.805227 (-4.169907) | 0.120235 / 6.500664 (-6.380429) | 0.041106 / 0.075469 (-0.034363) |\n\n### Benchmark: benchmark_map_filter.json\n\n| metric | filter | map fast-tokenizer batched | map identity | map identity batched | map no-op batched | map no-op batched numpy | map no-op batched pandas | map no-op batched pytorch | map no-op batched tensorflow |\n|--------|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 1.017469 / 1.841788 (-0.824319) | 12.383734 / 8.074308 (4.309426) | 10.352393 / 10.191392 (0.161001) | 0.131981 / 0.680424 (-0.548443) | 0.015204 / 0.534201 (-0.518997) | 0.286157 / 0.579283 (-0.293126) | 0.278270 / 0.434364 (-0.156094) | 0.325105 / 0.540337 (-0.215233) | 0.422301 / 1.386936 (-0.964635) |\n\n</details>\n</details>\n\n![](https://cml.dev/watermark.png#9323521505b7fab098fbe2a304389ee2d59783ff \"CML watermark\")\n" ]

[ "The docs for this PR live [here](https://moon-ci-docs.huggingface.co/docs/datasets/pr_6802). All of your documentation changes will be reflected on that endpoint. The docs are available until 30 days after the last update.", "<details>\n<summary>Show benchmarks</summary>\n\nPyArrow==8.0.0\n\n<details>\n<summary>Show updated benchmarks!</summary>\n\n### Benchmark: benchmark_array_xd.json\n\n| metric | read_batch_formatted_as_numpy after write_array2d | read_batch_formatted_as_numpy after write_flattened_sequence | read_batch_formatted_as_numpy after write_nested_sequence | read_batch_unformated after write_array2d | read_batch_unformated after write_flattened_sequence | read_batch_unformated after write_nested_sequence | read_col_formatted_as_numpy after write_array2d | read_col_formatted_as_numpy after write_flattened_sequence | read_col_formatted_as_numpy after write_nested_sequence | read_col_unformated after write_array2d | read_col_unformated after write_flattened_sequence | read_col_unformated after write_nested_sequence | read_formatted_as_numpy after write_array2d | read_formatted_as_numpy after write_flattened_sequence | read_formatted_as_numpy after write_nested_sequence | read_unformated after write_array2d | read_unformated after write_flattened_sequence | read_unformated after write_nested_sequence | write_array2d | write_flattened_sequence | write_nested_sequence |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.004991 / 0.011353 (-0.006362) | 0.003574 / 0.011008 (-0.007434) | 0.062369 / 0.038508 (0.023861) | 0.029966 / 0.023109 (0.006857) | 0.256140 / 0.275898 (-0.019758) | 0.283705 / 0.323480 (-0.039775) | 0.003170 / 0.007986 (-0.004816) | 0.002732 / 0.004328 (-0.001597) | 0.048048 / 0.004250 (0.043798) | 0.044497 / 0.037052 (0.007445) | 0.273206 / 0.258489 (0.014717) | 0.294593 / 0.293841 (0.000752) | 0.027251 / 0.128546 (-0.101295) | 0.010205 / 0.075646 (-0.065441) | 0.205979 / 0.419271 (-0.213293) | 0.035416 / 0.043533 (-0.008117) | 0.256260 / 0.255139 (0.001121) | 0.270580 / 0.283200 (-0.012620) | 0.019659 / 0.141683 (-0.122024) | 1.138722 / 1.452155 (-0.313432) | 1.170535 / 1.492716 (-0.322182) |\n\n### Benchmark: benchmark_getitem\\_100B.json\n\n| metric | get_batch_of\\_1024\\_random_rows | get_batch_of\\_1024\\_rows | get_first_row | get_last_row |\n|--------|---|---|---|---|\n| new / old (diff) | 0.091588 / 0.018006 (0.073582) | 0.301280 / 0.000490 (0.300791) | 0.000209 / 0.000200 (0.000009) | 0.000048 / 0.000054 (-0.000006) |\n\n### Benchmark: benchmark_indices_mapping.json\n\n| metric | select | shard | shuffle | sort | train_test_split |\n|--------|---|---|---|---|---|\n| new / old (diff) | 0.019684 / 0.037411 (-0.017727) | 0.061166 / 0.014526 (0.046640) | 0.072999 / 0.176557 (-0.103558) | 0.119264 / 0.737135 (-0.617871) | 0.074555 / 0.296338 (-0.221784) |\n\n### Benchmark: benchmark_iterating.json\n\n| metric | read 5000 | read 50000 | read_batch 50000 10 | read_batch 50000 100 | read_batch 50000 1000 | read_formatted numpy 5000 | read_formatted pandas 5000 | read_formatted tensorflow 5000 | read_formatted torch 5000 | read_formatted_batch numpy 5000 10 | read_formatted_batch numpy 5000 1000 | shuffled read 5000 | shuffled read 50000 | shuffled read_batch 50000 10 | shuffled read_batch 50000 100 | shuffled read_batch 50000 1000 | shuffled read_formatted numpy 5000 | shuffled read_formatted_batch numpy 5000 10 | shuffled read_formatted_batch numpy 5000 1000 |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.283210 / 0.215209 (0.068001) | 2.762284 / 2.077655 (0.684629) | 1.472700 / 1.504120 (-0.031420) | 1.352734 / 1.541195 (-0.188461) | 1.363287 / 1.468490 (-0.105203) | 0.558175 / 4.584777 (-4.026602) | 2.391648 / 3.745712 (-1.354064) | 2.787109 / 5.269862 (-2.482752) | 1.725635 / 4.565676 (-2.840042) | 0.061827 / 0.424275 (-0.362448) | 0.005351 / 0.007607 (-0.002256) | 0.337540 / 0.226044 (0.111496) | 3.353181 / 2.268929 (1.084252) | 1.829599 / 55.444624 (-53.615026) | 1.567691 / 6.876477 (-5.308786) | 1.605680 / 2.142072 (-0.536393) | 0.642182 / 4.805227 (-4.163045) | 0.117321 / 6.500664 (-6.383343) | 0.042555 / 0.075469 (-0.032915) |\n\n### Benchmark: benchmark_map_filter.json\n\n| metric | filter | map fast-tokenizer batched | map identity | map identity batched | map no-op batched | map no-op batched numpy | map no-op batched pandas | map no-op batched pytorch | map no-op batched tensorflow |\n|--------|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.991099 / 1.841788 (-0.850689) | 11.545219 / 8.074308 (3.470911) | 9.777574 / 10.191392 (-0.413818) | 0.130237 / 0.680424 (-0.550186) | 0.015068 / 0.534201 (-0.519133) | 0.286029 / 0.579283 (-0.293254) | 0.266778 / 0.434364 (-0.167586) | 0.321468 / 0.540337 (-0.218869) | 0.425371 / 1.386936 (-0.961565) |\n\n</details>\nPyArrow==latest\n\n<details>\n<summary>Show updated benchmarks!</summary>\n\n### Benchmark: benchmark_array_xd.json\n\n| metric | read_batch_formatted_as_numpy after write_array2d | read_batch_formatted_as_numpy after write_flattened_sequence | read_batch_formatted_as_numpy after write_nested_sequence | read_batch_unformated after write_array2d | read_batch_unformated after write_flattened_sequence | read_batch_unformated after write_nested_sequence | read_col_formatted_as_numpy after write_array2d | read_col_formatted_as_numpy after write_flattened_sequence | read_col_formatted_as_numpy after write_nested_sequence | read_col_unformated after write_array2d | read_col_unformated after write_flattened_sequence | read_col_unformated after write_nested_sequence | read_formatted_as_numpy after write_array2d | read_formatted_as_numpy after write_flattened_sequence | read_formatted_as_numpy after write_nested_sequence | read_unformated after write_array2d | read_unformated after write_flattened_sequence | read_unformated after write_nested_sequence | write_array2d | write_flattened_sequence | write_nested_sequence |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.005144 / 0.011353 (-0.006208) | 0.004046 / 0.011008 (-0.006962) | 0.050552 / 0.038508 (0.012043) | 0.030716 / 0.023109 (0.007607) | 0.273462 / 0.275898 (-0.002436) | 0.290649 / 0.323480 (-0.032831) | 0.004093 / 0.007986 (-0.003893) | 0.002700 / 0.004328 (-0.001628) | 0.048833 / 0.004250 (0.044582) | 0.040059 / 0.037052 (0.003007) | 0.282496 / 0.258489 (0.024007) | 0.309176 / 0.293841 (0.015335) | 0.029207 / 0.128546 (-0.099339) | 0.010740 / 0.075646 (-0.064907) | 0.057692 / 0.419271 (-0.361580) | 0.032570 / 0.043533 (-0.010963) | 0.269048 / 0.255139 (0.013909) | 0.287351 / 0.283200 (0.004151) | 0.017565 / 0.141683 (-0.124118) | 1.161628 / 1.452155 (-0.290526) | 1.187236 / 1.492716 (-0.305480) |\n\n### Benchmark: benchmark_getitem\\_100B.json\n\n| metric | get_batch_of\\_1024\\_random_rows | get_batch_of\\_1024\\_rows | get_first_row | get_last_row |\n|--------|---|---|---|---|\n| new / old (diff) | 0.095552 / 0.018006 (0.077546) | 0.312449 / 0.000490 (0.311959) | 0.000219 / 0.000200 (0.000019) | 0.000052 / 0.000054 (-0.000003) |\n\n### Benchmark: benchmark_indices_mapping.json\n\n| metric | select | shard | shuffle | sort | train_test_split |\n|--------|---|---|---|---|---|\n| new / old (diff) | 0.022425 / 0.037411 (-0.014986) | 0.074941 / 0.014526 (0.060416) | 0.086784 / 0.176557 (-0.089772) | 0.125630 / 0.737135 (-0.611506) | 0.088632 / 0.296338 (-0.207706) |\n\n### Benchmark: benchmark_iterating.json\n\n| metric | read 5000 | read 50000 | read_batch 50000 10 | read_batch 50000 100 | read_batch 50000 1000 | read_formatted numpy 5000 | read_formatted pandas 5000 | read_formatted tensorflow 5000 | read_formatted torch 5000 | read_formatted_batch numpy 5000 10 | read_formatted_batch numpy 5000 1000 | shuffled read 5000 | shuffled read 50000 | shuffled read_batch 50000 10 | shuffled read_batch 50000 100 | shuffled read_batch 50000 1000 | shuffled read_formatted numpy 5000 | shuffled read_formatted_batch numpy 5000 10 | shuffled read_formatted_batch numpy 5000 1000 |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.293003 / 0.215209 (0.077794) | 2.881826 / 2.077655 (0.804172) | 1.612840 / 1.504120 (0.108720) | 1.492727 / 1.541195 (-0.048468) | 1.520023 / 1.468490 (0.051532) | 0.558715 / 4.584777 (-4.026062) | 2.431093 / 3.745712 (-1.314619) | 2.782672 / 5.269862 (-2.487189) | 1.721611 / 4.565676 (-2.844065) | 0.063466 / 0.424275 (-0.360809) | 0.005221 / 0.007607 (-0.002386) | 0.352917 / 0.226044 (0.126873) | 3.443742 / 2.268929 (1.174814) | 1.981190 / 55.444624 (-53.463435) | 1.695396 / 6.876477 (-5.181081) | 1.709959 / 2.142072 (-0.432113) | 0.649267 / 4.805227 (-4.155960) | 0.116604 / 6.500664 (-6.384060) | 0.040688 / 0.075469 (-0.034781) |\n\n### Benchmark: benchmark_map_filter.json\n\n| metric | filter | map fast-tokenizer batched | map identity | map identity batched | map no-op batched | map no-op batched numpy | map no-op batched pandas | map no-op batched pytorch | map no-op batched tensorflow |\n|--------|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 1.023182 / 1.841788 (-0.818605) | 12.046760 / 8.074308 (3.972452) | 10.294706 / 10.191392 (0.103314) | 0.132323 / 0.680424 (-0.548101) | 0.016141 / 0.534201 (-0.518060) | 0.286620 / 0.579283 (-0.292663) | 0.272299 / 0.434364 (-0.162065) | 0.320995 / 0.540337 (-0.219343) | 0.424138 / 1.386936 (-0.962798) |\n\n</details>\n</details>\n\n![](https://cml.dev/watermark.png#873b7c8e354bfbd1873272a03d1392550d2cac39 \"CML watermark\")\n", "> Should it also be applied to this example a few lines later ?\r\n\r\nYes!", "done in https://github.com/huggingface/datasets/pull/6804" ]

[ "Hi! I'll open a PR on the Hub to fix this, but please use the Hub's [Community tab](https://huggingface.co/datasets/nyanko7/danbooru2023/discussions) to report such issues in the future.", "I've opened a [PR](https://huggingface.co/datasets/nyanko7/danbooru2023/discussions/8) in the repo, so let's continue the discussion there" ]

[ "Hi !\r\n\r\nIt's possible to multiple files at once:\r\n\r\n```python\r\ndata_files = \"data/*.jsonl\"\r\n# Or pass a list of files\r\nlangs = ['ka-ml', 'br-sr', 'ka-pt', 'id-ko', ..., 'fi-ze_zh', 'he-kk', 'ka-tr']\r\ndata_files = [f\"data/{lang}.jsonl\" for lang in langs]\r\nds = load_dataset(\"loicmagne/open-subtitles-250-bitext-mining\", data_files=data_files, split=\"train\")\r\n```\r\n\r\nAlso maybe you can add a subset called \"all\" for people that want to load all the data without having to list all the languages ?\r\n\r\n```yaml\r\n - config_name: all\r\n data_files: data/*.jsonl\r\n```\r\n", "Thanks for your reply, it is indeed much faster, however the result is a dataset where all the subsets are \"merged\" together, the language pair is lost:\r\n```\r\nDatasetDict({\r\n train: Dataset({\r\n features: ['sentence1', 'sentence2'],\r\n num_rows: 247809\r\n })\r\n})\r\n```\r\nI guess I could add a 'lang' feature for each row in the dataset, is there a better way to do it ?", "Hi @lhoestq over at https://github.com/embeddings-benchmark/mteb/issues/530 we have started examining these issues and would love to make a PR for datasets if we believe there is a way to improve the speed. As I assume you have a better overview than me @lhoestq, would you be interested in a PR, and might you have an idea about where we would start working on it?\r\n\r\nWe see a speed comparison of \r\n1. 15 minutes (for ~20% of the languages) when loaded using a for loop\r\n2. 17 minutes using the your suggestion\r\n3. ~30 seconds when using @loicmagne \"merged\" method.\r\n\r\nWorth mentioning is that solution 2 looses the language information.", "Can you retry using `datasets` 2.19 ? We improved a lot the speed of downloading datasets with tons of small files.\r\n\r\n```\r\npip install -U datasets\r\n```\r\n\r\nNow this takes 17sec on my side instead of the 17min minutes @loicmagne mentioned :)\r\n\r\n```python\r\n>>> %time ds = load_dataset(\"loicmagne/open-subtitles-250-bitext-mining\", data_files=\"data/*.jsonl\")\r\nDownloading readme: 100%|█████████████████████████████████| 13.7k/13.7k [00:00<00:00, 5.47MB/s]\r\nResolving data files: 100%|█████████████████████████████████| 250/250 [00:00<00:00, 612.51it/s]\r\nDownloading data: 100%|██████████████████████████████████| 250/250 [00:12<00:00, 19.68files/s]\r\nGenerating train split: 247809 examples [00:00, 1057071.08 examples/s]\r\nCPU times: user 4.95 s, sys: 3.1 s, total: 8.05 s\r\nWall time: 17.4 s\r\n```", "> Can you retry using `datasets` 2.19 ? We improved a lot the speed of downloading datasets with tons of small files.\r\n> \r\n> ```\r\n> pip install -U datasets\r\n> ```\r\n> \r\n> Now this takes 17sec on my side instead of the 17min minutes @loicmagne mentioned :)\r\n> \r\n> ```python\r\n> >>> %time ds = load_dataset(\"loicmagne/open-subtitles-250-bitext-mining\", data_files=\"data/*.jsonl\")\r\n> Downloading readme: 100%|█████████████████████████████████| 13.7k/13.7k [00:00<00:00, 5.47MB/s]\r\n> Resolving data files: 100%|█████████████████████████████████| 250/250 [00:00<00:00, 612.51it/s]\r\n> Downloading data: 100%|██████████████████████████████████| 250/250 [00:12<00:00, 19.68files/s]\r\n> Generating train split: 247809 examples [00:00, 1057071.08 examples/s]\r\n> CPU times: user 4.95 s, sys: 3.1 s, total: 8.05 s\r\n> Wall time: 17.4 s\r\n> ```\r\n\r\nI was actually just noticing that, I bumped from 2.18 to 2.19 and got a massive speedup, amazing!\r\n\r\nAbout the fact that subset names are lost when loading all files at once, currently my solution is to add a 'lang' feature to each rows, convert to polars and use:\r\n\r\n```python\r\nds_split = ds.to_polars().group_by('lang')\r\n```\r\n\r\nIt's fast so I think it's an acceptable solution, but is there a better way to do it ?", "It's the fastest way I think :)\r\n\r\nAlternatively you can download the dataset repository locally using [huggingface_hub](https://huggingface.co/docs/huggingface_hub/guides/download) (either via CLI or in python) and load the subsets one by one locally using a for loop as you were doing before (just pass the directory path to load_dataset instead of the dataset_id). " ]

[ "The docs for this PR live [here](https://moon-ci-docs.huggingface.co/docs/datasets/pr_6799). All of your documentation changes will be reflected on that endpoint. The docs are available until 30 days after the last update.", "The CI failures are unrelated to the changes", "<details>\n<summary>Show benchmarks</summary>\n\nPyArrow==8.0.0\n\n<details>\n<summary>Show updated benchmarks!</summary>\n\n### Benchmark: benchmark_array_xd.json\n\n| metric | read_batch_formatted_as_numpy after write_array2d | read_batch_formatted_as_numpy after write_flattened_sequence | read_batch_formatted_as_numpy after write_nested_sequence | read_batch_unformated after write_array2d | read_batch_unformated after write_flattened_sequence | read_batch_unformated after write_nested_sequence | read_col_formatted_as_numpy after write_array2d | read_col_formatted_as_numpy after write_flattened_sequence | read_col_formatted_as_numpy after write_nested_sequence | read_col_unformated after write_array2d | read_col_unformated after write_flattened_sequence | read_col_unformated after write_nested_sequence | read_formatted_as_numpy after write_array2d | read_formatted_as_numpy after write_flattened_sequence | read_formatted_as_numpy after write_nested_sequence | read_unformated after write_array2d | read_unformated after write_flattened_sequence | read_unformated after write_nested_sequence | write_array2d | write_flattened_sequence | write_nested_sequence |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.004974 / 0.011353 (-0.006378) | 0.003153 / 0.011008 (-0.007856) | 0.062785 / 0.038508 (0.024277) | 0.029504 / 0.023109 (0.006395) | 0.245558 / 0.275898 (-0.030340) | 0.274022 / 0.323480 (-0.049457) | 0.003173 / 0.007986 (-0.004813) | 0.002643 / 0.004328 (-0.001686) | 0.048917 / 0.004250 (0.044667) | 0.042965 / 0.037052 (0.005912) | 0.261266 / 0.258489 (0.002777) | 0.291546 / 0.293841 (-0.002295) | 0.027860 / 0.128546 (-0.100686) | 0.010397 / 0.075646 (-0.065249) | 0.205981 / 0.419271 (-0.213290) | 0.035663 / 0.043533 (-0.007870) | 0.250466 / 0.255139 (-0.004673) | 0.273947 / 0.283200 (-0.009253) | 0.016659 / 0.141683 (-0.125023) | 1.147884 / 1.452155 (-0.304270) | 1.187609 / 1.492716 (-0.305107) |\n\n### Benchmark: benchmark_getitem\\_100B.json\n\n| metric | get_batch_of\\_1024\\_random_rows | get_batch_of\\_1024\\_rows | get_first_row | get_last_row |\n|--------|---|---|---|---|\n| new / old (diff) | 0.095564 / 0.018006 (0.077558) | 0.300086 / 0.000490 (0.299597) | 0.000212 / 0.000200 (0.000012) | 0.000049 / 0.000054 (-0.000005) |\n\n### Benchmark: benchmark_indices_mapping.json\n\n| metric | select | shard | shuffle | sort | train_test_split |\n|--------|---|---|---|---|---|\n| new / old (diff) | 0.018100 / 0.037411 (-0.019311) | 0.061342 / 0.014526 (0.046816) | 0.073747 / 0.176557 (-0.102810) | 0.120577 / 0.737135 (-0.616559) | 0.075797 / 0.296338 (-0.220541) |\n\n### Benchmark: benchmark_iterating.json\n\n| metric | read 5000 | read 50000 | read_batch 50000 10 | read_batch 50000 100 | read_batch 50000 1000 | read_formatted numpy 5000 | read_formatted pandas 5000 | read_formatted tensorflow 5000 | read_formatted torch 5000 | read_formatted_batch numpy 5000 10 | read_formatted_batch numpy 5000 1000 | shuffled read 5000 | shuffled read 50000 | shuffled read_batch 50000 10 | shuffled read_batch 50000 100 | shuffled read_batch 50000 1000 | shuffled read_formatted numpy 5000 | shuffled read_formatted_batch numpy 5000 10 | shuffled read_formatted_batch numpy 5000 1000 |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.288766 / 0.215209 (0.073557) | 2.835274 / 2.077655 (0.757620) | 1.515288 / 1.504120 (0.011168) | 1.396097 / 1.541195 (-0.145098) | 1.424293 / 1.468490 (-0.044197) | 0.568356 / 4.584777 (-4.016421) | 2.393171 / 3.745712 (-1.352541) | 2.756219 / 5.269862 (-2.513642) | 1.731343 / 4.565676 (-2.834334) | 0.062542 / 0.424275 (-0.361733) | 0.005385 / 0.007607 (-0.002223) | 0.340876 / 0.226044 (0.114832) | 3.376649 / 2.268929 (1.107720) | 1.856135 / 55.444624 (-53.588490) | 1.581802 / 6.876477 (-5.294675) | 1.591081 / 2.142072 (-0.550992) | 0.647963 / 4.805227 (-4.157264) | 0.119218 / 6.500664 (-6.381446) | 0.042660 / 0.075469 (-0.032809) |\n\n### Benchmark: benchmark_map_filter.json\n\n| metric | filter | map fast-tokenizer batched | map identity | map identity batched | map no-op batched | map no-op batched numpy | map no-op batched pandas | map no-op batched pytorch | map no-op batched tensorflow |\n|--------|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 1.005017 / 1.841788 (-0.836770) | 11.670779 / 8.074308 (3.596471) | 9.533790 / 10.191392 (-0.657602) | 0.141571 / 0.680424 (-0.538853) | 0.013987 / 0.534201 (-0.520214) | 0.286598 / 0.579283 (-0.292685) | 0.260123 / 0.434364 (-0.174240) | 0.324186 / 0.540337 (-0.216151) | 0.421246 / 1.386936 (-0.965690) |\n\n</details>\nPyArrow==latest\n\n<details>\n<summary>Show updated benchmarks!</summary>\n\n### Benchmark: benchmark_array_xd.json\n\n| metric | read_batch_formatted_as_numpy after write_array2d | read_batch_formatted_as_numpy after write_flattened_sequence | read_batch_formatted_as_numpy after write_nested_sequence | read_batch_unformated after write_array2d | read_batch_unformated after write_flattened_sequence | read_batch_unformated after write_nested_sequence | read_col_formatted_as_numpy after write_array2d | read_col_formatted_as_numpy after write_flattened_sequence | read_col_formatted_as_numpy after write_nested_sequence | read_col_unformated after write_array2d | read_col_unformated after write_flattened_sequence | read_col_unformated after write_nested_sequence | read_formatted_as_numpy after write_array2d | read_formatted_as_numpy after write_flattened_sequence | read_formatted_as_numpy after write_nested_sequence | read_unformated after write_array2d | read_unformated after write_flattened_sequence | read_unformated after write_nested_sequence | write_array2d | write_flattened_sequence | write_nested_sequence |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.005196 / 0.011353 (-0.006157) | 0.003697 / 0.011008 (-0.007311) | 0.049530 / 0.038508 (0.011022) | 0.030892 / 0.023109 (0.007783) | 0.284787 / 0.275898 (0.008889) | 0.302833 / 0.323480 (-0.020647) | 0.004203 / 0.007986 (-0.003783) | 0.002736 / 0.004328 (-0.001592) | 0.050203 / 0.004250 (0.045953) | 0.040335 / 0.037052 (0.003283) | 0.292508 / 0.258489 (0.034019) | 0.317918 / 0.293841 (0.024077) | 0.029144 / 0.128546 (-0.099403) | 0.010171 / 0.075646 (-0.065475) | 0.058130 / 0.419271 (-0.361141) | 0.032743 / 0.043533 (-0.010790) | 0.281354 / 0.255139 (0.026215) | 0.296951 / 0.283200 (0.013751) | 0.018399 / 0.141683 (-0.123284) | 1.158852 / 1.452155 (-0.293303) | 1.189750 / 1.492716 (-0.302966) |\n\n### Benchmark: benchmark_getitem\\_100B.json\n\n| metric | get_batch_of\\_1024\\_random_rows | get_batch_of\\_1024\\_rows | get_first_row | get_last_row |\n|--------|---|---|---|---|\n| new / old (diff) | 0.093073 / 0.018006 (0.075066) | 0.301779 / 0.000490 (0.301290) | 0.000209 / 0.000200 (0.000009) | 0.000051 / 0.000054 (-0.000003) |\n\n### Benchmark: benchmark_indices_mapping.json\n\n| metric | select | shard | shuffle | sort | train_test_split |\n|--------|---|---|---|---|---|\n| new / old (diff) | 0.021565 / 0.037411 (-0.015846) | 0.075237 / 0.014526 (0.060711) | 0.087368 / 0.176557 (-0.089188) | 0.126955 / 0.737135 (-0.610180) | 0.088456 / 0.296338 (-0.207883) |\n\n### Benchmark: benchmark_iterating.json\n\n| metric | read 5000 | read 50000 | read_batch 50000 10 | read_batch 50000 100 | read_batch 50000 1000 | read_formatted numpy 5000 | read_formatted pandas 5000 | read_formatted tensorflow 5000 | read_formatted torch 5000 | read_formatted_batch numpy 5000 10 | read_formatted_batch numpy 5000 1000 | shuffled read 5000 | shuffled read 50000 | shuffled read_batch 50000 10 | shuffled read_batch 50000 100 | shuffled read_batch 50000 1000 | shuffled read_formatted numpy 5000 | shuffled read_formatted_batch numpy 5000 10 | shuffled read_formatted_batch numpy 5000 1000 |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.291225 / 0.215209 (0.076016) | 2.863220 / 2.077655 (0.785565) | 1.616936 / 1.504120 (0.112817) | 1.500553 / 1.541195 (-0.040641) | 1.501693 / 1.468490 (0.033203) | 0.560118 / 4.584777 (-4.024659) | 2.439241 / 3.745712 (-1.306472) | 2.786804 / 5.269862 (-2.483058) | 1.737772 / 4.565676 (-2.827905) | 0.063668 / 0.424275 (-0.360607) | 0.005320 / 0.007607 (-0.002287) | 0.344539 / 0.226044 (0.118495) | 3.418803 / 2.268929 (1.149874) | 1.981791 / 55.444624 (-53.462834) | 1.698484 / 6.876477 (-5.177993) | 1.686815 / 2.142072 (-0.455258) | 0.646911 / 4.805227 (-4.158316) | 0.116969 / 6.500664 (-6.383696) | 0.040380 / 0.075469 (-0.035089) |\n\n### Benchmark: benchmark_map_filter.json\n\n| metric | filter | map fast-tokenizer batched | map identity | map identity batched | map no-op batched | map no-op batched numpy | map no-op batched pandas | map no-op batched pytorch | map no-op batched tensorflow |\n|--------|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 1.017337 / 1.841788 (-0.824451) | 11.858212 / 8.074308 (3.783904) | 10.270287 / 10.191392 (0.078895) | 0.154266 / 0.680424 (-0.526158) | 0.014886 / 0.534201 (-0.519315) | 0.292354 / 0.579283 (-0.286929) | 0.270888 / 0.434364 (-0.163476) | 0.333289 / 0.540337 (-0.207049) | 0.423001 / 1.386936 (-0.963935) |\n\n</details>\n</details>\n\n![](https://cml.dev/watermark.png#d9cc95f6d0513bbc692bb73c669346e3d1825cb0 \"CML watermark\")\n" ]

[ "Good catch! Feel free to open a PR with the suggested fix :).", "There is also the [`MockDownloadManager`](https://github.com/JonasLoos/datasets/blob/main/src/datasets/download/mock_download_manager.py#L33), which seems like it might get passed here too. However, to me, it doesn't really seem relevant to the users of the datasets library, so I would just ignore it. What do you think, @mariosasko?", "The API (`dummy_data` CLI command ) that uses the `MockDownloadManager` has been deprecated, so ignoring it sounds good!" ]

[ "The docs for this PR live [here](https://moon-ci-docs.huggingface.co/docs/datasets/pr_6797). All of your documentation changes will be reflected on that endpoint. The docs are available until 30 days after the last update.", "Finally:\r\n- the initial issue seems it was temporary\r\n- there is a different issue now \r\n\r\n```\r\nFAILED tests/test_load.py::ModuleFactoryTest::test_HubDatasetModuleFactoryWithParquetExport - datasets.utils._dataset_viewer.DatasetViewerError: No exported Parquet files available.\r\nFAILED tests/test_load.py::ModuleFactoryTest::test_HubDatasetModuleFactoryWithParquetExport_errors_on_wrong_sha - datasets.utils._dataset_viewer.DatasetViewerError: No exported Parquet files available.\r\nFAILED tests/test_load.py::test_load_dataset_builder_for_community_dataset_with_script - AssertionError: assert 'dataset_with_script' == 'parquet'\r\n \r\n - parquet\r\n + dataset_with_script\r\n```" ]

[ "Finally:\r\n- the initial issue seems it was temporary\r\n- there is a different issue now: https://github.com/huggingface/datasets/actions/runs/8627153993/job/23646584590?pr=6797\r\n```\r\nFAILED tests/test_load.py::ModuleFactoryTest::test_HubDatasetModuleFactoryWithParquetExport - datasets.utils._dataset_viewer.DatasetViewerError: No exported Parquet files available.\r\nFAILED tests/test_load.py::ModuleFactoryTest::test_HubDatasetModuleFactoryWithParquetExport_errors_on_wrong_sha - datasets.utils._dataset_viewer.DatasetViewerError: No exported Parquet files available.\r\nFAILED tests/test_load.py::test_load_dataset_builder_for_community_dataset_with_script - AssertionError: assert 'dataset_with_script' == 'parquet'\r\n \r\n - parquet\r\n + dataset_with_script\r\n```\r\n\r\nMaybe related to `hf-internal-testing/dataset_with_script` dataset: https://huggingface.co/datasets/hf-internal-testing/dataset_with_script", "This URL: https://datasets-server.huggingface.co/parquet?dataset=hf-internal-testing/dataset_with_script\r\nraises:\r\n> {\"error\":\"The dataset viewer doesn't support this dataset because it runs arbitrary python code. Please open a discussion in the discussion tab if you think this is an error and tag @lhoestq and @severo.\"}\r\n\r\nWas there a recent change on the Hub enforcing this behavior?", "OK, I just saw this PR:\r\n- https://github.com/huggingface/dataset-viewer/pull/2689\r\n\r\nOnce merged and deployed, it should fix the issue.", "Once the script-dataset has been allowed in the dataset-viewer, we should fix our test to make the CI pass.\r\n\r\nI am addressing this." ]

[ "The docs for this PR live [here](https://moon-ci-docs.huggingface.co/docs/datasets/pr_6795). All of your documentation changes will be reflected on that endpoint. The docs are available until 30 days after the last update.", "@huggingface/datasets once this PR is merged, I would suggest making a release. Do you agree?", "<details>\n<summary>Show benchmarks</summary>\n\nPyArrow==8.0.0\n\n<details>\n<summary>Show updated benchmarks!</summary>\n\n### Benchmark: benchmark_array_xd.json\n\n| metric | read_batch_formatted_as_numpy after write_array2d | read_batch_formatted_as_numpy after write_flattened_sequence | read_batch_formatted_as_numpy after write_nested_sequence | read_batch_unformated after write_array2d | read_batch_unformated after write_flattened_sequence | read_batch_unformated after write_nested_sequence | read_col_formatted_as_numpy after write_array2d | read_col_formatted_as_numpy after write_flattened_sequence | read_col_formatted_as_numpy after write_nested_sequence | read_col_unformated after write_array2d | read_col_unformated after write_flattened_sequence | read_col_unformated after write_nested_sequence | read_formatted_as_numpy after write_array2d | read_formatted_as_numpy after write_flattened_sequence | read_formatted_as_numpy after write_nested_sequence | read_unformated after write_array2d | read_unformated after write_flattened_sequence | read_unformated after write_nested_sequence | write_array2d | write_flattened_sequence | write_nested_sequence |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.005367 / 0.011353 (-0.005986) | 0.003161 / 0.011008 (-0.007847) | 0.063259 / 0.038508 (0.024751) | 0.030550 / 0.023109 (0.007441) | 0.243789 / 0.275898 (-0.032109) | 0.262474 / 0.323480 (-0.061006) | 0.003157 / 0.007986 (-0.004829) | 0.002586 / 0.004328 (-0.001742) | 0.049336 / 0.004250 (0.045085) | 0.046434 / 0.037052 (0.009382) | 0.249142 / 0.258489 (-0.009347) | 0.282953 / 0.293841 (-0.010888) | 0.027881 / 0.128546 (-0.100666) | 0.010069 / 0.075646 (-0.065578) | 0.207937 / 0.419271 (-0.211334) | 0.036005 / 0.043533 (-0.007528) | 0.251850 / 0.255139 (-0.003288) | 0.265156 / 0.283200 (-0.018044) | 0.019780 / 0.141683 (-0.121903) | 1.124301 / 1.452155 (-0.327853) | 1.177392 / 1.492716 (-0.315324) |\n\n### Benchmark: benchmark_getitem\\_100B.json\n\n| metric | get_batch_of\\_1024\\_random_rows | get_batch_of\\_1024\\_rows | get_first_row | get_last_row |\n|--------|---|---|---|---|\n| new / old (diff) | 0.091045 / 0.018006 (0.073039) | 0.301258 / 0.000490 (0.300769) | 0.000214 / 0.000200 (0.000014) | 0.000048 / 0.000054 (-0.000006) |\n\n### Benchmark: benchmark_indices_mapping.json\n\n| metric | select | shard | shuffle | sort | train_test_split |\n|--------|---|---|---|---|---|\n| new / old (diff) | 0.018726 / 0.037411 (-0.018686) | 0.061623 / 0.014526 (0.047097) | 0.073905 / 0.176557 (-0.102651) | 0.119444 / 0.737135 (-0.617692) | 0.074614 / 0.296338 (-0.221725) |\n\n### Benchmark: benchmark_iterating.json\n\n| metric | read 5000 | read 50000 | read_batch 50000 10 | read_batch 50000 100 | read_batch 50000 1000 | read_formatted numpy 5000 | read_formatted pandas 5000 | read_formatted tensorflow 5000 | read_formatted torch 5000 | read_formatted_batch numpy 5000 10 | read_formatted_batch numpy 5000 1000 | shuffled read 5000 | shuffled read 50000 | shuffled read_batch 50000 10 | shuffled read_batch 50000 100 | shuffled read_batch 50000 1000 | shuffled read_formatted numpy 5000 | shuffled read_formatted_batch numpy 5000 10 | shuffled read_formatted_batch numpy 5000 1000 |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.287313 / 0.215209 (0.072104) | 2.772864 / 2.077655 (0.695209) | 1.465267 / 1.504120 (-0.038853) | 1.343666 / 1.541195 (-0.197528) | 1.329390 / 1.468490 (-0.139100) | 0.570222 / 4.584777 (-4.014555) | 2.421835 / 3.745712 (-1.323877) | 2.747282 / 5.269862 (-2.522579) | 1.728733 / 4.565676 (-2.836943) | 0.063671 / 0.424275 (-0.360604) | 0.005343 / 0.007607 (-0.002264) | 0.335078 / 0.226044 (0.109033) | 3.334305 / 2.268929 (1.065376) | 1.779496 / 55.444624 (-53.665129) | 1.496475 / 6.876477 (-5.380002) | 1.507848 / 2.142072 (-0.634224) | 0.653653 / 4.805227 (-4.151575) | 0.118373 / 6.500664 (-6.382291) | 0.041727 / 0.075469 (-0.033742) |\n\n### Benchmark: benchmark_map_filter.json\n\n| metric | filter | map fast-tokenizer batched | map identity | map identity batched | map no-op batched | map no-op batched numpy | map no-op batched pandas | map no-op batched pytorch | map no-op batched tensorflow |\n|--------|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.981985 / 1.841788 (-0.859803) | 11.290978 / 8.074308 (3.216670) | 9.499217 / 10.191392 (-0.692175) | 0.131353 / 0.680424 (-0.549071) | 0.014416 / 0.534201 (-0.519785) | 0.288381 / 0.579283 (-0.290902) | 0.265483 / 0.434364 (-0.168880) | 0.323438 / 0.540337 (-0.216900) | 0.417946 / 1.386936 (-0.968990) |\n\n</details>\nPyArrow==latest\n\n<details>\n<summary>Show updated benchmarks!</summary>\n\n### Benchmark: benchmark_array_xd.json\n\n| metric | read_batch_formatted_as_numpy after write_array2d | read_batch_formatted_as_numpy after write_flattened_sequence | read_batch_formatted_as_numpy after write_nested_sequence | read_batch_unformated after write_array2d | read_batch_unformated after write_flattened_sequence | read_batch_unformated after write_nested_sequence | read_col_formatted_as_numpy after write_array2d | read_col_formatted_as_numpy after write_flattened_sequence | read_col_formatted_as_numpy after write_nested_sequence | read_col_unformated after write_array2d | read_col_unformated after write_flattened_sequence | read_col_unformated after write_nested_sequence | read_formatted_as_numpy after write_array2d | read_formatted_as_numpy after write_flattened_sequence | read_formatted_as_numpy after write_nested_sequence | read_unformated after write_array2d | read_unformated after write_flattened_sequence | read_unformated after write_nested_sequence | write_array2d | write_flattened_sequence | write_nested_sequence |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.005272 / 0.011353 (-0.006081) | 0.003551 / 0.011008 (-0.007457) | 0.050173 / 0.038508 (0.011665) | 0.031291 / 0.023109 (0.008182) | 0.278658 / 0.275898 (0.002760) | 0.301812 / 0.323480 (-0.021668) | 0.004237 / 0.007986 (-0.003748) | 0.002713 / 0.004328 (-0.001615) | 0.049483 / 0.004250 (0.045233) | 0.039995 / 0.037052 (0.002943) | 0.293101 / 0.258489 (0.034612) | 0.319956 / 0.293841 (0.026116) | 0.029127 / 0.128546 (-0.099419) | 0.010247 / 0.075646 (-0.065400) | 0.057929 / 0.419271 (-0.361342) | 0.032942 / 0.043533 (-0.010591) | 0.281677 / 0.255139 (0.026538) | 0.297937 / 0.283200 (0.014737) | 0.018285 / 0.141683 (-0.123398) | 1.272858 / 1.452155 (-0.179297) | 1.213375 / 1.492716 (-0.279342) |\n\n### Benchmark: benchmark_getitem\\_100B.json\n\n| metric | get_batch_of\\_1024\\_random_rows | get_batch_of\\_1024\\_rows | get_first_row | get_last_row |\n|--------|---|---|---|---|\n| new / old (diff) | 0.091110 / 0.018006 (0.073104) | 0.302589 / 0.000490 (0.302099) | 0.000214 / 0.000200 (0.000014) | 0.000070 / 0.000054 (0.000015) |\n\n### Benchmark: benchmark_indices_mapping.json\n\n| metric | select | shard | shuffle | sort | train_test_split |\n|--------|---|---|---|---|---|\n| new / old (diff) | 0.021520 / 0.037411 (-0.015891) | 0.075013 / 0.014526 (0.060487) | 0.088695 / 0.176557 (-0.087862) | 0.128281 / 0.737135 (-0.608854) | 0.090611 / 0.296338 (-0.205727) |\n\n### Benchmark: benchmark_iterating.json\n\n| metric | read 5000 | read 50000 | read_batch 50000 10 | read_batch 50000 100 | read_batch 50000 1000 | read_formatted numpy 5000 | read_formatted pandas 5000 | read_formatted tensorflow 5000 | read_formatted torch 5000 | read_formatted_batch numpy 5000 10 | read_formatted_batch numpy 5000 1000 | shuffled read 5000 | shuffled read 50000 | shuffled read_batch 50000 10 | shuffled read_batch 50000 100 | shuffled read_batch 50000 1000 | shuffled read_formatted numpy 5000 | shuffled read_formatted_batch numpy 5000 10 | shuffled read_formatted_batch numpy 5000 1000 |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.297457 / 0.215209 (0.082248) | 2.928612 / 2.077655 (0.850957) | 1.613245 / 1.504120 (0.109125) | 1.485263 / 1.541195 (-0.055931) | 1.496885 / 1.468490 (0.028395) | 0.570120 / 4.584777 (-4.014657) | 2.487532 / 3.745712 (-1.258180) | 2.761552 / 5.269862 (-2.508309) | 1.731864 / 4.565676 (-2.833812) | 0.062989 / 0.424275 (-0.361286) | 0.005428 / 0.007607 (-0.002179) | 0.354932 / 0.226044 (0.128888) | 3.524475 / 2.268929 (1.255547) | 1.977684 / 55.444624 (-53.466941) | 1.692568 / 6.876477 (-5.183909) | 1.673003 / 2.142072 (-0.469069) | 0.643976 / 4.805227 (-4.161251) | 0.116499 / 6.500664 (-6.384165) | 0.040772 / 0.075469 (-0.034697) |\n\n### Benchmark: benchmark_map_filter.json\n\n| metric | filter | map fast-tokenizer batched | map identity | map identity batched | map no-op batched | map no-op batched numpy | map no-op batched pandas | map no-op batched pytorch | map no-op batched tensorflow |\n|--------|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 1.020354 / 1.841788 (-0.821434) | 12.143991 / 8.074308 (4.069683) | 10.354058 / 10.191392 (0.162666) | 0.145460 / 0.680424 (-0.534964) | 0.015356 / 0.534201 (-0.518845) | 0.307190 / 0.579283 (-0.272093) | 0.276664 / 0.434364 (-0.157699) | 0.350068 / 0.540337 (-0.190269) | 0.440824 / 1.386936 (-0.946112) |\n\n</details>\n</details>\n\n![](https://cml.dev/watermark.png#a3bc89d8bfd47c2a175c3ce16d92b7307cdeafd6 \"CML watermark\")\n" ]

[ "The docs for this PR live [here](https://moon-ci-docs.huggingface.co/docs/datasets/pr_6794). All of your documentation changes will be reflected on that endpoint. The docs are available until 30 days after the last update.", "CI is failing because of the missing parquet export of one test dataset, PR to fix this at https://github.com/huggingface/dataset-viewer/pull/2689", "I took your comments into account :) lmk what you think @mariosasko ", "<details>\n<summary>Show benchmarks</summary>\n\nPyArrow==8.0.0\n\n<details>\n<summary>Show updated benchmarks!</summary>\n\n### Benchmark: benchmark_array_xd.json\n\n| metric | read_batch_formatted_as_numpy after write_array2d | read_batch_formatted_as_numpy after write_flattened_sequence | read_batch_formatted_as_numpy after write_nested_sequence | read_batch_unformated after write_array2d | read_batch_unformated after write_flattened_sequence | read_batch_unformated after write_nested_sequence | read_col_formatted_as_numpy after write_array2d | read_col_formatted_as_numpy after write_flattened_sequence | read_col_formatted_as_numpy after write_nested_sequence | read_col_unformated after write_array2d | read_col_unformated after write_flattened_sequence | read_col_unformated after write_nested_sequence | read_formatted_as_numpy after write_array2d | read_formatted_as_numpy after write_flattened_sequence | read_formatted_as_numpy after write_nested_sequence | read_unformated after write_array2d | read_unformated after write_flattened_sequence | read_unformated after write_nested_sequence | write_array2d | write_flattened_sequence | write_nested_sequence |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.004956 / 0.011353 (-0.006397) | 0.003282 / 0.011008 (-0.007726) | 0.064028 / 0.038508 (0.025520) | 0.030420 / 0.023109 (0.007311) | 0.240097 / 0.275898 (-0.035801) | 0.266356 / 0.323480 (-0.057124) | 0.003116 / 0.007986 (-0.004869) | 0.002597 / 0.004328 (-0.001731) | 0.050230 / 0.004250 (0.045980) | 0.043864 / 0.037052 (0.006812) | 0.258711 / 0.258489 (0.000222) | 0.290816 / 0.293841 (-0.003025) | 0.027898 / 0.128546 (-0.100648) | 0.009941 / 0.075646 (-0.065705) | 0.208917 / 0.419271 (-0.210355) | 0.035891 / 0.043533 (-0.007642) | 0.253332 / 0.255139 (-0.001807) | 0.274300 / 0.283200 (-0.008900) | 0.019466 / 0.141683 (-0.122217) | 1.133896 / 1.452155 (-0.318259) | 1.178130 / 1.492716 (-0.314586) |\n\n### Benchmark: benchmark_getitem\\_100B.json\n\n| metric | get_batch_of\\_1024\\_random_rows | get_batch_of\\_1024\\_rows | get_first_row | get_last_row |\n|--------|---|---|---|---|\n| new / old (diff) | 0.091093 / 0.018006 (0.073087) | 0.293632 / 0.000490 (0.293142) | 0.000216 / 0.000200 (0.000016) | 0.000042 / 0.000054 (-0.000013) |\n\n### Benchmark: benchmark_indices_mapping.json\n\n| metric | select | shard | shuffle | sort | train_test_split |\n|--------|---|---|---|---|---|\n| new / old (diff) | 0.017722 / 0.037411 (-0.019689) | 0.060241 / 0.014526 (0.045715) | 0.072024 / 0.176557 (-0.104533) | 0.118521 / 0.737135 (-0.618615) | 0.071107 / 0.296338 (-0.225232) |\n\n### Benchmark: benchmark_iterating.json\n\n| metric | read 5000 | read 50000 | read_batch 50000 10 | read_batch 50000 100 | read_batch 50000 1000 | read_formatted numpy 5000 | read_formatted pandas 5000 | read_formatted tensorflow 5000 | read_formatted torch 5000 | read_formatted_batch numpy 5000 10 | read_formatted_batch numpy 5000 1000 | shuffled read 5000 | shuffled read 50000 | shuffled read_batch 50000 10 | shuffled read_batch 50000 100 | shuffled read_batch 50000 1000 | shuffled read_formatted numpy 5000 | shuffled read_formatted_batch numpy 5000 10 | shuffled read_formatted_batch numpy 5000 1000 |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.280950 / 0.215209 (0.065741) | 2.781361 / 2.077655 (0.703706) | 1.477949 / 1.504120 (-0.026171) | 1.356388 / 1.541195 (-0.184807) | 1.361808 / 1.468490 (-0.106682) | 0.565499 / 4.584777 (-4.019278) | 2.389206 / 3.745712 (-1.356506) | 2.712782 / 5.269862 (-2.557079) | 1.701402 / 4.565676 (-2.864274) | 0.063619 / 0.424275 (-0.360656) | 0.005321 / 0.007607 (-0.002286) | 0.336783 / 0.226044 (0.110739) | 3.299628 / 2.268929 (1.030699) | 1.794686 / 55.444624 (-53.649939) | 1.504207 / 6.876477 (-5.372270) | 1.524637 / 2.142072 (-0.617436) | 0.642833 / 4.805227 (-4.162395) | 0.117808 / 6.500664 (-6.382856) | 0.041539 / 0.075469 (-0.033930) |\n\n### Benchmark: benchmark_map_filter.json\n\n| metric | filter | map fast-tokenizer batched | map identity | map identity batched | map no-op batched | map no-op batched numpy | map no-op batched pandas | map no-op batched pytorch | map no-op batched tensorflow |\n|--------|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.960193 / 1.841788 (-0.881595) | 11.229147 / 8.074308 (3.154839) | 9.380653 / 10.191392 (-0.810739) | 0.137184 / 0.680424 (-0.543240) | 0.013399 / 0.534201 (-0.520802) | 0.314904 / 0.579283 (-0.264379) | 0.262539 / 0.434364 (-0.171825) | 0.354007 / 0.540337 (-0.186331) | 0.451698 / 1.386936 (-0.935238) |\n\n</details>\nPyArrow==latest\n\n<details>\n<summary>Show updated benchmarks!</summary>\n\n### Benchmark: benchmark_array_xd.json\n\n| metric | read_batch_formatted_as_numpy after write_array2d | read_batch_formatted_as_numpy after write_flattened_sequence | read_batch_formatted_as_numpy after write_nested_sequence | read_batch_unformated after write_array2d | read_batch_unformated after write_flattened_sequence | read_batch_unformated after write_nested_sequence | read_col_formatted_as_numpy after write_array2d | read_col_formatted_as_numpy after write_flattened_sequence | read_col_formatted_as_numpy after write_nested_sequence | read_col_unformated after write_array2d | read_col_unformated after write_flattened_sequence | read_col_unformated after write_nested_sequence | read_formatted_as_numpy after write_array2d | read_formatted_as_numpy after write_flattened_sequence | read_formatted_as_numpy after write_nested_sequence | read_unformated after write_array2d | read_unformated after write_flattened_sequence | read_unformated after write_nested_sequence | write_array2d | write_flattened_sequence | write_nested_sequence |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.005207 / 0.011353 (-0.006146) | 0.003660 / 0.011008 (-0.007348) | 0.049931 / 0.038508 (0.011423) | 0.030918 / 0.023109 (0.007809) | 0.271243 / 0.275898 (-0.004655) | 0.295706 / 0.323480 (-0.027774) | 0.004106 / 0.007986 (-0.003879) | 0.002750 / 0.004328 (-0.001578) | 0.048337 / 0.004250 (0.044086) | 0.039944 / 0.037052 (0.002892) | 0.284013 / 0.258489 (0.025524) | 0.306827 / 0.293841 (0.012987) | 0.029183 / 0.128546 (-0.099363) | 0.010033 / 0.075646 (-0.065613) | 0.058126 / 0.419271 (-0.361146) | 0.032427 / 0.043533 (-0.011106) | 0.276471 / 0.255139 (0.021332) | 0.288428 / 0.283200 (0.005229) | 0.017549 / 0.141683 (-0.124134) | 1.142361 / 1.452155 (-0.309793) | 1.184514 / 1.492716 (-0.308202) |\n\n### Benchmark: benchmark_getitem\\_100B.json\n\n| metric | get_batch_of\\_1024\\_random_rows | get_batch_of\\_1024\\_rows | get_first_row | get_last_row |\n|--------|---|---|---|---|\n| new / old (diff) | 0.090350 / 0.018006 (0.072344) | 0.292511 / 0.000490 (0.292021) | 0.000215 / 0.000200 (0.000015) | 0.000041 / 0.000054 (-0.000013) |\n\n### Benchmark: benchmark_indices_mapping.json\n\n| metric | select | shard | shuffle | sort | train_test_split |\n|--------|---|---|---|---|---|\n| new / old (diff) | 0.021572 / 0.037411 (-0.015840) | 0.074310 / 0.014526 (0.059784) | 0.086102 / 0.176557 (-0.090455) | 0.123507 / 0.737135 (-0.613629) | 0.087397 / 0.296338 (-0.208941) |\n\n### Benchmark: benchmark_iterating.json\n\n| metric | read 5000 | read 50000 | read_batch 50000 10 | read_batch 50000 100 | read_batch 50000 1000 | read_formatted numpy 5000 | read_formatted pandas 5000 | read_formatted tensorflow 5000 | read_formatted torch 5000 | read_formatted_batch numpy 5000 10 | read_formatted_batch numpy 5000 1000 | shuffled read 5000 | shuffled read 50000 | shuffled read_batch 50000 10 | shuffled read_batch 50000 100 | shuffled read_batch 50000 1000 | shuffled read_formatted numpy 5000 | shuffled read_formatted_batch numpy 5000 10 | shuffled read_formatted_batch numpy 5000 1000 |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.294038 / 0.215209 (0.078829) | 2.889662 / 2.077655 (0.812007) | 1.591775 / 1.504120 (0.087655) | 1.468815 / 1.541195 (-0.072379) | 1.470226 / 1.468490 (0.001736) | 0.574557 / 4.584777 (-4.010220) | 2.481377 / 3.745712 (-1.264335) | 2.763368 / 5.269862 (-2.506493) | 1.713707 / 4.565676 (-2.851969) | 0.064158 / 0.424275 (-0.360117) | 0.005553 / 0.007607 (-0.002054) | 0.353480 / 0.226044 (0.127436) | 3.447689 / 2.268929 (1.178760) | 1.975802 / 55.444624 (-53.468822) | 1.673561 / 6.876477 (-5.202915) | 1.637212 / 2.142072 (-0.504860) | 0.640667 / 4.805227 (-4.164560) | 0.114618 / 6.500664 (-6.386046) | 0.038912 / 0.075469 (-0.036557) |\n\n### Benchmark: benchmark_map_filter.json\n\n| metric | filter | map fast-tokenizer batched | map identity | map identity batched | map no-op batched | map no-op batched numpy | map no-op batched pandas | map no-op batched pytorch | map no-op batched tensorflow |\n|--------|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 1.007581 / 1.841788 (-0.834207) | 11.874250 / 8.074308 (3.799942) | 10.312692 / 10.191392 (0.121300) | 0.142705 / 0.680424 (-0.537719) | 0.015438 / 0.534201 (-0.518763) | 0.285919 / 0.579283 (-0.293364) | 0.278223 / 0.434364 (-0.156141) | 0.323806 / 0.540337 (-0.216531) | 0.415007 / 1.386936 (-0.971929) |\n\n</details>\n</details>\n\n![](https://cml.dev/watermark.png#0f1f27c69f6cc8d085b66a8a2ba0440a39bc5bce \"CML watermark\")\n" ]

[]

[ "The docs for this PR live [here](https://moon-ci-docs.huggingface.co/docs/datasets/pr_6792). All of your documentation changes will be reflected on that endpoint. The docs are available until 30 days after the last update.", "<details>\n<summary>Show benchmarks</summary>\n\nPyArrow==8.0.0\n\n<details>\n<summary>Show updated benchmarks!</summary>\n\n### Benchmark: benchmark_array_xd.json\n\n| metric | read_batch_formatted_as_numpy after write_array2d | read_batch_formatted_as_numpy after write_flattened_sequence | read_batch_formatted_as_numpy after write_nested_sequence | read_batch_unformated after write_array2d | read_batch_unformated after write_flattened_sequence | read_batch_unformated after write_nested_sequence | read_col_formatted_as_numpy after write_array2d | read_col_formatted_as_numpy after write_flattened_sequence | read_col_formatted_as_numpy after write_nested_sequence | read_col_unformated after write_array2d | read_col_unformated after write_flattened_sequence | read_col_unformated after write_nested_sequence | read_formatted_as_numpy after write_array2d | read_formatted_as_numpy after write_flattened_sequence | read_formatted_as_numpy after write_nested_sequence | read_unformated after write_array2d | read_unformated after write_flattened_sequence | read_unformated after write_nested_sequence | write_array2d | write_flattened_sequence | write_nested_sequence |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.005212 / 0.011353 (-0.006141) | 0.003536 / 0.011008 (-0.007472) | 0.063042 / 0.038508 (0.024534) | 0.032654 / 0.023109 (0.009545) | 0.242040 / 0.275898 (-0.033858) | 0.267735 / 0.323480 (-0.055745) | 0.003188 / 0.007986 (-0.004797) | 0.002697 / 0.004328 (-0.001631) | 0.050127 / 0.004250 (0.045877) | 0.045960 / 0.037052 (0.008908) | 0.260926 / 0.258489 (0.002437) | 0.293953 / 0.293841 (0.000112) | 0.028352 / 0.128546 (-0.100194) | 0.010558 / 0.075646 (-0.065088) | 0.208104 / 0.419271 (-0.211167) | 0.035889 / 0.043533 (-0.007644) | 0.246265 / 0.255139 (-0.008874) | 0.271819 / 0.283200 (-0.011381) | 0.018491 / 0.141683 (-0.123192) | 1.299274 / 1.452155 (-0.152881) | 1.205932 / 1.492716 (-0.286784) |\n\n### Benchmark: benchmark_getitem\\_100B.json\n\n| metric | get_batch_of\\_1024\\_random_rows | get_batch_of\\_1024\\_rows | get_first_row | get_last_row |\n|--------|---|---|---|---|\n| new / old (diff) | 0.095574 / 0.018006 (0.077568) | 0.306493 / 0.000490 (0.306003) | 0.000216 / 0.000200 (0.000016) | 0.000042 / 0.000054 (-0.000012) |\n\n### Benchmark: benchmark_indices_mapping.json\n\n| metric | select | shard | shuffle | sort | train_test_split |\n|--------|---|---|---|---|---|\n| new / old (diff) | 0.018304 / 0.037411 (-0.019107) | 0.061312 / 0.014526 (0.046786) | 0.074483 / 0.176557 (-0.102073) | 0.122231 / 0.737135 (-0.614905) | 0.075315 / 0.296338 (-0.221024) |\n\n### Benchmark: benchmark_iterating.json\n\n| metric | read 5000 | read 50000 | read_batch 50000 10 | read_batch 50000 100 | read_batch 50000 1000 | read_formatted numpy 5000 | read_formatted pandas 5000 | read_formatted tensorflow 5000 | read_formatted torch 5000 | read_formatted_batch numpy 5000 10 | read_formatted_batch numpy 5000 1000 | shuffled read 5000 | shuffled read 50000 | shuffled read_batch 50000 10 | shuffled read_batch 50000 100 | shuffled read_batch 50000 1000 | shuffled read_formatted numpy 5000 | shuffled read_formatted_batch numpy 5000 10 | shuffled read_formatted_batch numpy 5000 1000 |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.275632 / 0.215209 (0.060423) | 2.696402 / 2.077655 (0.618747) | 1.418657 / 1.504120 (-0.085463) | 1.300014 / 1.541195 (-0.241181) | 1.299148 / 1.468490 (-0.169342) | 0.561893 / 4.584777 (-4.022884) | 2.410710 / 3.745712 (-1.335002) | 2.749058 / 5.269862 (-2.520803) | 1.712835 / 4.565676 (-2.852841) | 0.062278 / 0.424275 (-0.361997) | 0.005040 / 0.007607 (-0.002567) | 0.330352 / 0.226044 (0.104308) | 3.291274 / 2.268929 (1.022345) | 1.780987 / 55.444624 (-53.663638) | 1.514764 / 6.876477 (-5.361713) | 1.533892 / 2.142072 (-0.608181) | 0.632307 / 4.805227 (-4.172921) | 0.116011 / 6.500664 (-6.384653) | 0.041964 / 0.075469 (-0.033505) |\n\n### Benchmark: benchmark_map_filter.json\n\n| metric | filter | map fast-tokenizer batched | map identity | map identity batched | map no-op batched | map no-op batched numpy | map no-op batched pandas | map no-op batched pytorch | map no-op batched tensorflow |\n|--------|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.982713 / 1.841788 (-0.859075) | 11.521597 / 8.074308 (3.447289) | 9.713063 / 10.191392 (-0.478329) | 0.132115 / 0.680424 (-0.548309) | 0.014564 / 0.534201 (-0.519637) | 0.294087 / 0.579283 (-0.285196) | 0.267399 / 0.434364 (-0.166965) | 0.327967 / 0.540337 (-0.212370) | 0.419279 / 1.386936 (-0.967657) |\n\n</details>\nPyArrow==latest\n\n<details>\n<summary>Show updated benchmarks!</summary>\n\n### Benchmark: benchmark_array_xd.json\n\n| metric | read_batch_formatted_as_numpy after write_array2d | read_batch_formatted_as_numpy after write_flattened_sequence | read_batch_formatted_as_numpy after write_nested_sequence | read_batch_unformated after write_array2d | read_batch_unformated after write_flattened_sequence | read_batch_unformated after write_nested_sequence | read_col_formatted_as_numpy after write_array2d | read_col_formatted_as_numpy after write_flattened_sequence | read_col_formatted_as_numpy after write_nested_sequence | read_col_unformated after write_array2d | read_col_unformated after write_flattened_sequence | read_col_unformated after write_nested_sequence | read_formatted_as_numpy after write_array2d | read_formatted_as_numpy after write_flattened_sequence | read_formatted_as_numpy after write_nested_sequence | read_unformated after write_array2d | read_unformated after write_flattened_sequence | read_unformated after write_nested_sequence | write_array2d | write_flattened_sequence | write_nested_sequence |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.005098 / 0.011353 (-0.006255) | 0.003513 / 0.011008 (-0.007495) | 0.050121 / 0.038508 (0.011613) | 0.030842 / 0.023109 (0.007732) | 0.271323 / 0.275898 (-0.004575) | 0.293592 / 0.323480 (-0.029887) | 0.004225 / 0.007986 (-0.003761) | 0.002802 / 0.004328 (-0.001527) | 0.049035 / 0.004250 (0.044785) | 0.040748 / 0.037052 (0.003696) | 0.282542 / 0.258489 (0.024053) | 0.303779 / 0.293841 (0.009938) | 0.029213 / 0.128546 (-0.099333) | 0.010578 / 0.075646 (-0.065068) | 0.058053 / 0.419271 (-0.361219) | 0.032830 / 0.043533 (-0.010703) | 0.272226 / 0.255139 (0.017087) | 0.290485 / 0.283200 (0.007285) | 0.017968 / 0.141683 (-0.123714) | 1.166998 / 1.452155 (-0.285156) | 1.256354 / 1.492716 (-0.236362) |\n\n### Benchmark: benchmark_getitem\\_100B.json\n\n| metric | get_batch_of\\_1024\\_random_rows | get_batch_of\\_1024\\_rows | get_first_row | get_last_row |\n|--------|---|---|---|---|\n| new / old (diff) | 0.096126 / 0.018006 (0.078120) | 0.306303 / 0.000490 (0.305813) | 0.000246 / 0.000200 (0.000047) | 0.000049 / 0.000054 (-0.000006) |\n\n### Benchmark: benchmark_indices_mapping.json\n\n| metric | select | shard | shuffle | sort | train_test_split |\n|--------|---|---|---|---|---|\n| new / old (diff) | 0.022413 / 0.037411 (-0.014998) | 0.075008 / 0.014526 (0.060482) | 0.087703 / 0.176557 (-0.088854) | 0.127358 / 0.737135 (-0.609777) | 0.088817 / 0.296338 (-0.207521) |\n\n### Benchmark: benchmark_iterating.json\n\n| metric | read 5000 | read 50000 | read_batch 50000 10 | read_batch 50000 100 | read_batch 50000 1000 | read_formatted numpy 5000 | read_formatted pandas 5000 | read_formatted tensorflow 5000 | read_formatted torch 5000 | read_formatted_batch numpy 5000 10 | read_formatted_batch numpy 5000 1000 | shuffled read 5000 | shuffled read 50000 | shuffled read_batch 50000 10 | shuffled read_batch 50000 100 | shuffled read_batch 50000 1000 | shuffled read_formatted numpy 5000 | shuffled read_formatted_batch numpy 5000 10 | shuffled read_formatted_batch numpy 5000 1000 |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.301103 / 0.215209 (0.085894) | 2.965441 / 2.077655 (0.887787) | 1.608075 / 1.504120 (0.103955) | 1.479214 / 1.541195 (-0.061981) | 1.492039 / 1.468490 (0.023549) | 0.574455 / 4.584777 (-4.010322) | 2.483234 / 3.745712 (-1.262478) | 2.795901 / 5.269862 (-2.473961) | 1.742034 / 4.565676 (-2.823642) | 0.064170 / 0.424275 (-0.360105) | 0.005572 / 0.007607 (-0.002035) | 0.349500 / 0.226044 (0.123456) | 3.482161 / 2.268929 (1.213232) | 1.950065 / 55.444624 (-53.494559) | 1.675270 / 6.876477 (-5.201207) | 1.674534 / 2.142072 (-0.467538) | 0.657478 / 4.805227 (-4.147749) | 0.117534 / 6.500664 (-6.383130) | 0.040880 / 0.075469 (-0.034589) |\n\n### Benchmark: benchmark_map_filter.json\n\n| metric | filter | map fast-tokenizer batched | map identity | map identity batched | map no-op batched | map no-op batched numpy | map no-op batched pandas | map no-op batched pytorch | map no-op batched tensorflow |\n|--------|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 1.035276 / 1.841788 (-0.806511) | 12.035581 / 8.074308 (3.961273) | 10.127778 / 10.191392 (-0.063614) | 0.142289 / 0.680424 (-0.538134) | 0.014702 / 0.534201 (-0.519499) | 0.288206 / 0.579283 (-0.291077) | 0.282251 / 0.434364 (-0.152113) | 0.323479 / 0.540337 (-0.216858) | 0.419019 / 1.386936 (-0.967917) |\n\n</details>\n</details>\n\n![](https://cml.dev/watermark.png#0f27d7b77c73412cfc50b24354bfd7a3e838202f \"CML watermark\")\n" ]

[ "I realized I was passing a string column to this instead of float. Is it possible to add a warning or error to prevent users from falsely believing there's a bug?", "Hello!\r\n\r\nI agree that we could add some safeguards around the type of `ds[column]`. At least for FAISS, we need the column to be made of embeddings as FAISS doesn't perform the embeddings itself.\r\n\r\nI can propose a PR sometime this week.", "@Dref360 thanks for the initiative!" ]

[]

[ "Default `writer_batch_size `is set to 1000 (see [map](https://huggingface.co/docs/datasets/v2.16.1/en/package_reference/main_classes#datasets.Dataset.map)).\r\nThe \"tmp1335llua\" is probably the temp file it creates while writing to disk.\r\nMaybe try lowering the `writer_batch_size`.\r\n\r\nFor multi-processing you should probably pass the `processor `as an argument (with e.g. partial) to the function or create it inside so that the sub-processes have access to it and maybe add `if __name__ == \"__main__\"` (not sure that's necessary?).\r\n", "Hi @Modexus,\r\n\r\nThank you very much for the help! Yep after playing around with map, I managed to get the parallel processing to work by implementing it like you suggested.\r\n\r\nRegarding the temp files, it seems like the temp files just keep growing in size as the map continues. Eventually, once map finishes, the temp files are deleted, but they are instead saved as cache .arrow files. These cache files are absolutely gigantic (~ 30-50x the size of the initial dataset!).\r\n\r\nAfter playing around with the `prepare_dataset()` function above, it seems this issue is caused by the following line in the function, where the log-Mel spectrogram of the audio is calculated:\r\n\r\n`# compute log-Mel input features from input audio array\r\n batch[\"input_features\"] = processor.feature_extractor(audio[\"array\"], \r\n sampling_rate=audio[\"sampling_rate\"]).input_features[0]\r\n`\r\n\r\nWhen I remove this line, the final cache files are approximately the same size as the initial dataset.\r\n\r\nCan I check whether this is expected behavior with the whisper feature extractor? I cant imagine the spectrograms are that large!\r\n\r\nThank you so much for the help!", "I'm having a similar issue with the spectrographs taking up an incredibly large amount of space. (i.e. 100GB for 3GB of audio). Is this really normal behavior?", "Upon taking a look at the hex contents of the mapped dataset files I found that the overwhelming majority of the data contained within them was duplicated junk similar to this. I'm not very familiar with the inner workings of AI but I have to assume this is an inefficient way of storing data at best and a bug at worst.\r\n\r\n", "Same problem, dataset.map takes long time to process 12GB raw audio data and create 200GB cache file. Is there any method can run process(map) during train, instead current run \r\nonce and save cache file ? " ]

[ "All data is saved in the arrow format on disk.\r\nIf you return a tensor it gets converted to arrow before saving to disk when using map.\r\n\r\nTo get a tensor when you access data elements you can use `dataset.set_format(\"pt\")`.\r\nNote that this just changes how the data is loaded, not how it is stored.", "> All data is saved in the arrow format on disk. If you return a tensor it gets converted to arrow before saving to disk when using map.\r\n> \r\n> To get a tensor when you access data elements you can use `dataset.set_format(\"pt\")`. Note that this just changes how the data is loaded, not how it is stored.\r\n\r\nThank you very much for your explanation, I understand what you mean now. So you're saying that when streaming=True, there's no need to convert it to the arrow format and save it to disk. But if we directly load all formats and then convert them into the arrow format after passing through the map function, it will convert torch.Tensor into a List. I see." ]

[ "From my current understanding, this timeout is only used when we need to get the results.\r\n\r\nOne of:\r\n1. All tasks are done\r\n2. One worker died\r\n\r\nYour function should work fine and it's definitely a bug if it doesn't.", "When one of the `map`'s worker processes crashes, the linked code re-raises an error from the crash and returns it to the caller.\r\n\r\nIf your question is how to limit the time of long-running tasks/worker processes, such functionality doesn't exist in `datasets` (yet), which means you need to implement it yourself.\r\n\r\nE.g., you can implement it using the built-in `signal` module like this:\r\n```python\r\nimport time\r\nimport signal\r\nfrom contextlib import contextmanager\r\n\r\nfrom datasets import Dataset\r\n\r\n\r\n@contextmanager\r\ndef max_exec_time(t):\r\n def raise_timeout_handler(signum, frame):\r\n raise TimeoutError\r\n \r\n orig_handler = signal.getsignal(signal.SIGALRM)\r\n signal.signal(signal.SIGALRM, raise_timeout_handler)\r\n try:\r\n signal.alarm(t)\r\n yield\r\n finally:\r\n signal.alarm(0)\r\n signal.signal(signal.SIGALRM, orig_handler)\r\n\r\n\r\ndef worker(example, rank):\r\n try:\r\n with max_exec_time(20): # 20 sec execution limit\r\n if rank % 2 == 0:\r\n time.sleep(50) # simulate a long-running task\r\n example[\"a\"] = 100\r\n except TimeoutError:\r\n example[\"a\"] = None # Or return empty batches here in the \"batched\" mode\r\n return example\r\n\r\ndata = Dataset.from_list([{\"a\": 1}, {\"a\": 2}])\r\ndata = data.map(worker, num_proc=2, with_rank=True)\r\nprint(data[0])\r\n```", "> From my current understanding, this timeout is only used when we need to get the results.\r\n> \r\n> One of:\r\n> \r\n> 1. All tasks are done\r\n> 2. One worker died\r\n> \r\n> Your function should work fine and it's definitely a bug if it doesn't.\r\n\r\nthanks for responding! can you reproduce the stuck with the above example code?", "> When one of the `map`'s worker processes crashes, the linked code re-raises an error from the crash and returns it to the caller.\r\n> \r\n> If your question is how to limit the time of long-running tasks/worker processes, such functionality doesn't exist in `datasets` (yet), which means you need to implement it yourself.\r\n> \r\n> E.g., you can implement it using the built-in `signal` module like this:\r\n> \r\n> ```python\r\n> import time\r\n> import signal\r\n> from contextlib import contextmanager\r\n> \r\n> from datasets import Dataset\r\n> \r\n> \r\n> @contextmanager\r\n> def max_exec_time(t):\r\n> def raise_timeout_handler(signum, frame):\r\n> raise TimeoutError\r\n> \r\n> orig_handler = signal.getsignal(signal.SIGALRM)\r\n> signal.signal(signal.SIGALRM, raise_timeout_handler)\r\n> try:\r\n> signal.alarm(t)\r\n> yield\r\n> finally:\r\n> signal.alarm(0)\r\n> signal.signal(signal.SIGALRM, orig_handler)\r\n> \r\n> \r\n> def worker(example, rank):\r\n> try:\r\n> with max_exec_time(20): # 20 sec execution limit\r\n> if rank % 2 == 0:\r\n> time.sleep(50) # simulate a long-running task\r\n> example[\"a\"] = 100\r\n> except TimeoutError:\r\n> example[\"a\"] = None # Or return empty batches here in the \"batched\" mode\r\n> return example\r\n> \r\n> data = Dataset.from_list([{\"a\": 1}, {\"a\": 2}])\r\n> data = data.map(worker, num_proc=2, with_rank=True)\r\n> print(data[0])\r\n> ```\r\n\r\nthanks for responding! However, I don't think we should use `signal` in the context of multiprocessing since sometimes it will crash one process and raise the following error\r\nhttps://github.com/huggingface/datasets/blob/c3ddb1ef00334a6f973679a51e783905fbc9ef0b/src/datasets/utils/py_utils.py#L664", "> thanks for responding! However, I don't think we should use signal in the context of multiprocessing since sometimes it will crash one process and raise the following error\r\n\r\nThe above code has `try/except` to catch the error from the handler. Or do you get an error other than `TimeoutError`?", "> > thanks for responding! However, I don't think we should use signal in the context of multiprocessing since sometimes it will crash one process and raise the following error\r\n> \r\n> The above code has `try/except` to catch the error from the handler. Or do you get an error other than `TimeoutError`?\r\n\r\nyup, it will raise the RuntimeError: https://github.com/huggingface/datasets/blob/c3ddb1ef00334a6f973679a51e783905fbc9ef0b/src/datasets/utils/py_utils.py#L667C19-L670C22\r\n\r\n```\r\n raise RuntimeError(\r\n \"One of the subprocesses has abruptly died during map operation.\"\r\n \"To debug the error, disable multiprocessing.\"\r\n )\r\n```" ]

[ "The docs for this PR live [here](https://moon-ci-docs.huggingface.co/docs/datasets/pr_6786). All of your documentation changes will be reflected on that endpoint. The docs are available until 30 days after the last update.", "Hi ! Thanks for diving into this, this conversion to python lists is indeed quite slow.\r\n\r\nArray2DExtensionType and Array3DExtensionType currently rely on pyarrow lists, but we will soon modify them to use FixedShapeTensorArray instead which is more efficient (e.g. doesn't need to store an offset for each value). So ideally it would be cool to speed this code up without using those extension types or it will be blocking to improve Array2DExtensionType and Array3DExtensionType.\r\n\r\nIf I understand correctly you just need the logic from ArrayExtensionArray.to_numpy ? If so feel free to make a separate function and ArrayExtensionArray.to_numpy can call it" ]

[ "The docs for this PR live [here](https://moon-ci-docs.huggingface.co/docs/datasets/pr_6785). All of your documentation changes will be reflected on that endpoint. The docs are available until 30 days after the last update.", "<details>\n<summary>Show benchmarks</summary>\n\nPyArrow==8.0.0\n\n<details>\n<summary>Show updated benchmarks!</summary>\n\n### Benchmark: benchmark_array_xd.json\n\n| metric | read_batch_formatted_as_numpy after write_array2d | read_batch_formatted_as_numpy after write_flattened_sequence | read_batch_formatted_as_numpy after write_nested_sequence | read_batch_unformated after write_array2d | read_batch_unformated after write_flattened_sequence | read_batch_unformated after write_nested_sequence | read_col_formatted_as_numpy after write_array2d | read_col_formatted_as_numpy after write_flattened_sequence | read_col_formatted_as_numpy after write_nested_sequence | read_col_unformated after write_array2d | read_col_unformated after write_flattened_sequence | read_col_unformated after write_nested_sequence | read_formatted_as_numpy after write_array2d | read_formatted_as_numpy after write_flattened_sequence | read_formatted_as_numpy after write_nested_sequence | read_unformated after write_array2d | read_unformated after write_flattened_sequence | read_unformated after write_nested_sequence | write_array2d | write_flattened_sequence | write_nested_sequence |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.005224 / 0.011353 (-0.006129) | 0.003938 / 0.011008 (-0.007070) | 0.063829 / 0.038508 (0.025321) | 0.030975 / 0.023109 (0.007865) | 0.265090 / 0.275898 (-0.010808) | 0.290994 / 0.323480 (-0.032486) | 0.003083 / 0.007986 (-0.004902) | 0.002810 / 0.004328 (-0.001518) | 0.048860 / 0.004250 (0.044609) | 0.044663 / 0.037052 (0.007611) | 0.272161 / 0.258489 (0.013672) | 0.306966 / 0.293841 (0.013125) | 0.028028 / 0.128546 (-0.100518) | 0.010616 / 0.075646 (-0.065031) | 0.211649 / 0.419271 (-0.207623) | 0.035906 / 0.043533 (-0.007626) | 0.251779 / 0.255139 (-0.003360) | 0.275543 / 0.283200 (-0.007657) | 0.017710 / 0.141683 (-0.123973) | 1.127015 / 1.452155 (-0.325139) | 1.173319 / 1.492716 (-0.319397) |\n\n### Benchmark: benchmark_getitem\\_100B.json\n\n| metric | get_batch_of\\_1024\\_random_rows | get_batch_of\\_1024\\_rows | get_first_row | get_last_row |\n|--------|---|---|---|---|\n| new / old (diff) | 0.090625 / 0.018006 (0.072619) | 0.301973 / 0.000490 (0.301483) | 0.000217 / 0.000200 (0.000017) | 0.000053 / 0.000054 (-0.000002) |\n\n### Benchmark: benchmark_indices_mapping.json\n\n| metric | select | shard | shuffle | sort | train_test_split |\n|--------|---|---|---|---|---|\n| new / old (diff) | 0.018868 / 0.037411 (-0.018543) | 0.062402 / 0.014526 (0.047876) | 0.074053 / 0.176557 (-0.102504) | 0.121484 / 0.737135 (-0.615652) | 0.078674 / 0.296338 (-0.217664) |\n\n### Benchmark: benchmark_iterating.json\n\n| metric | read 5000 | read 50000 | read_batch 50000 10 | read_batch 50000 100 | read_batch 50000 1000 | read_formatted numpy 5000 | read_formatted pandas 5000 | read_formatted tensorflow 5000 | read_formatted torch 5000 | read_formatted_batch numpy 5000 10 | read_formatted_batch numpy 5000 1000 | shuffled read 5000 | shuffled read 50000 | shuffled read_batch 50000 10 | shuffled read_batch 50000 100 | shuffled read_batch 50000 1000 | shuffled read_formatted numpy 5000 | shuffled read_formatted_batch numpy 5000 10 | shuffled read_formatted_batch numpy 5000 1000 |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.277821 / 0.215209 (0.062612) | 2.761642 / 2.077655 (0.683987) | 1.452735 / 1.504120 (-0.051385) | 1.336303 / 1.541195 (-0.204891) | 1.343045 / 1.468490 (-0.125445) | 0.560917 / 4.584777 (-4.023860) | 2.353427 / 3.745712 (-1.392286) | 2.699067 / 5.269862 (-2.570795) | 1.704752 / 4.565676 (-2.860925) | 0.062668 / 0.424275 (-0.361607) | 0.005120 / 0.007607 (-0.002487) | 0.330455 / 0.226044 (0.104410) | 3.264604 / 2.268929 (0.995675) | 1.791940 / 55.444624 (-53.652685) | 1.526083 / 6.876477 (-5.350394) | 1.541429 / 2.142072 (-0.600643) | 0.630343 / 4.805227 (-4.174884) | 0.115189 / 6.500664 (-6.385475) | 0.041716 / 0.075469 (-0.033753) |\n\n### Benchmark: benchmark_map_filter.json\n\n| metric | filter | map fast-tokenizer batched | map identity | map identity batched | map no-op batched | map no-op batched numpy | map no-op batched pandas | map no-op batched pytorch | map no-op batched tensorflow |\n|--------|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.975008 / 1.841788 (-0.866779) | 11.326924 / 8.074308 (3.252616) | 9.810300 / 10.191392 (-0.381092) | 0.141068 / 0.680424 (-0.539356) | 0.013950 / 0.534201 (-0.520251) | 0.285691 / 0.579283 (-0.293592) | 0.257968 / 0.434364 (-0.176396) | 0.322976 / 0.540337 (-0.217361) | 0.411114 / 1.386936 (-0.975822) |\n\n</details>\nPyArrow==latest\n\n<details>\n<summary>Show updated benchmarks!</summary>\n\n### Benchmark: benchmark_array_xd.json\n\n| metric | read_batch_formatted_as_numpy after write_array2d | read_batch_formatted_as_numpy after write_flattened_sequence | read_batch_formatted_as_numpy after write_nested_sequence | read_batch_unformated after write_array2d | read_batch_unformated after write_flattened_sequence | read_batch_unformated after write_nested_sequence | read_col_formatted_as_numpy after write_array2d | read_col_formatted_as_numpy after write_flattened_sequence | read_col_formatted_as_numpy after write_nested_sequence | read_col_unformated after write_array2d | read_col_unformated after write_flattened_sequence | read_col_unformated after write_nested_sequence | read_formatted_as_numpy after write_array2d | read_formatted_as_numpy after write_flattened_sequence | read_formatted_as_numpy after write_nested_sequence | read_unformated after write_array2d | read_unformated after write_flattened_sequence | read_unformated after write_nested_sequence | write_array2d | write_flattened_sequence | write_nested_sequence |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.005176 / 0.011353 (-0.006177) | 0.003631 / 0.011008 (-0.007377) | 0.050006 / 0.038508 (0.011498) | 0.030622 / 0.023109 (0.007513) | 0.277364 / 0.275898 (0.001466) | 0.299752 / 0.323480 (-0.023728) | 0.004110 / 0.007986 (-0.003876) | 0.002694 / 0.004328 (-0.001634) | 0.048966 / 0.004250 (0.044715) | 0.039634 / 0.037052 (0.002582) | 0.289959 / 0.258489 (0.031470) | 0.320689 / 0.293841 (0.026848) | 0.029285 / 0.128546 (-0.099261) | 0.010435 / 0.075646 (-0.065211) | 0.057432 / 0.419271 (-0.361840) | 0.032554 / 0.043533 (-0.010979) | 0.277354 / 0.255139 (0.022215) | 0.296872 / 0.283200 (0.013673) | 0.017338 / 0.141683 (-0.124344) | 1.134174 / 1.452155 (-0.317981) | 1.184695 / 1.492716 (-0.308021) |\n\n### Benchmark: benchmark_getitem\\_100B.json\n\n| metric | get_batch_of\\_1024\\_random_rows | get_batch_of\\_1024\\_rows | get_first_row | get_last_row |\n|--------|---|---|---|---|\n| new / old (diff) | 0.089953 / 0.018006 (0.071947) | 0.299372 / 0.000490 (0.298882) | 0.000212 / 0.000200 (0.000012) | 0.000043 / 0.000054 (-0.000012) |\n\n### Benchmark: benchmark_indices_mapping.json\n\n| metric | select | shard | shuffle | sort | train_test_split |\n|--------|---|---|---|---|---|\n| new / old (diff) | 0.021349 / 0.037411 (-0.016062) | 0.075167 / 0.014526 (0.060641) | 0.085910 / 0.176557 (-0.090647) | 0.124729 / 0.737135 (-0.612406) | 0.088313 / 0.296338 (-0.208025) |\n\n### Benchmark: benchmark_iterating.json\n\n| metric | read 5000 | read 50000 | read_batch 50000 10 | read_batch 50000 100 | read_batch 50000 1000 | read_formatted numpy 5000 | read_formatted pandas 5000 | read_formatted tensorflow 5000 | read_formatted torch 5000 | read_formatted_batch numpy 5000 10 | read_formatted_batch numpy 5000 1000 | shuffled read 5000 | shuffled read 50000 | shuffled read_batch 50000 10 | shuffled read_batch 50000 100 | shuffled read_batch 50000 1000 | shuffled read_formatted numpy 5000 | shuffled read_formatted_batch numpy 5000 10 | shuffled read_formatted_batch numpy 5000 1000 |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.291939 / 0.215209 (0.076730) | 2.851077 / 2.077655 (0.773423) | 1.609382 / 1.504120 (0.105262) | 1.469656 / 1.541195 (-0.071539) | 1.490469 / 1.468490 (0.021979) | 0.570421 / 4.584777 (-4.014356) | 2.441438 / 3.745712 (-1.304274) | 2.756514 / 5.269862 (-2.513347) | 1.714202 / 4.565676 (-2.851474) | 0.063656 / 0.424275 (-0.360619) | 0.005640 / 0.007607 (-0.001967) | 0.336240 / 0.226044 (0.110196) | 3.355434 / 2.268929 (1.086505) | 1.947553 / 55.444624 (-53.497072) | 1.672776 / 6.876477 (-5.203700) | 1.685316 / 2.142072 (-0.456757) | 0.638849 / 4.805227 (-4.166378) | 0.116304 / 6.500664 (-6.384360) | 0.041588 / 0.075469 (-0.033881) |\n\n### Benchmark: benchmark_map_filter.json\n\n| metric | filter | map fast-tokenizer batched | map identity | map identity batched | map no-op batched | map no-op batched numpy | map no-op batched pandas | map no-op batched pytorch | map no-op batched tensorflow |\n|--------|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 1.026700 / 1.841788 (-0.815088) | 12.044628 / 8.074308 (3.970319) | 10.464007 / 10.191392 (0.272615) | 0.156169 / 0.680424 (-0.524255) | 0.015624 / 0.534201 (-0.518577) | 0.287233 / 0.579283 (-0.292050) | 0.270374 / 0.434364 (-0.163990) | 0.325255 / 0.540337 (-0.215083) | 0.412021 / 1.386936 (-0.974915) |\n\n</details>\n</details>\n\n![](https://cml.dev/watermark.png#6f7f1718e3db54d7923ebe4383301fdd380c18b9 \"CML watermark\")\n" ]

[ "The docs for this PR live [here](https://moon-ci-docs.huggingface.co/docs/datasets/pr_6784). All of your documentation changes will be reflected on that endpoint. The docs are available until 30 days after the last update.", "CI failures are unrelated, so this is ready for the review", "<details>\n<summary>Show benchmarks</summary>\n\nPyArrow==8.0.0\n\n<details>\n<summary>Show updated benchmarks!</summary>\n\n### Benchmark: benchmark_array_xd.json\n\n| metric | read_batch_formatted_as_numpy after write_array2d | read_batch_formatted_as_numpy after write_flattened_sequence | read_batch_formatted_as_numpy after write_nested_sequence | read_batch_unformated after write_array2d | read_batch_unformated after write_flattened_sequence | read_batch_unformated after write_nested_sequence | read_col_formatted_as_numpy after write_array2d | read_col_formatted_as_numpy after write_flattened_sequence | read_col_formatted_as_numpy after write_nested_sequence | read_col_unformated after write_array2d | read_col_unformated after write_flattened_sequence | read_col_unformated after write_nested_sequence | read_formatted_as_numpy after write_array2d | read_formatted_as_numpy after write_flattened_sequence | read_formatted_as_numpy after write_nested_sequence | read_unformated after write_array2d | read_unformated after write_flattened_sequence | read_unformated after write_nested_sequence | write_array2d | write_flattened_sequence | write_nested_sequence |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.005130 / 0.011353 (-0.006223) | 0.003784 / 0.011008 (-0.007224) | 0.064899 / 0.038508 (0.026391) | 0.029456 / 0.023109 (0.006347) | 0.253384 / 0.275898 (-0.022514) | 0.273509 / 0.323480 (-0.049971) | 0.004116 / 0.007986 (-0.003870) | 0.002713 / 0.004328 (-0.001615) | 0.053984 / 0.004250 (0.049733) | 0.043538 / 0.037052 (0.006485) | 0.264696 / 0.258489 (0.006207) | 0.298321 / 0.293841 (0.004480) | 0.027916 / 0.128546 (-0.100630) | 0.010734 / 0.075646 (-0.064912) | 0.208284 / 0.419271 (-0.210988) | 0.035873 / 0.043533 (-0.007659) | 0.251028 / 0.255139 (-0.004111) | 0.270835 / 0.283200 (-0.012364) | 0.017475 / 0.141683 (-0.124208) | 1.130728 / 1.452155 (-0.321426) | 1.188672 / 1.492716 (-0.304044) |\n\n### Benchmark: benchmark_getitem\\_100B.json\n\n| metric | get_batch_of\\_1024\\_random_rows | get_batch_of\\_1024\\_rows | get_first_row | get_last_row |\n|--------|---|---|---|---|\n| new / old (diff) | 0.094191 / 0.018006 (0.076185) | 0.304064 / 0.000490 (0.303575) | 0.000251 / 0.000200 (0.000051) | 0.000058 / 0.000054 (0.000004) |\n\n### Benchmark: benchmark_indices_mapping.json\n\n| metric | select | shard | shuffle | sort | train_test_split |\n|--------|---|---|---|---|---|\n| new / old (diff) | 0.018414 / 0.037411 (-0.018998) | 0.061550 / 0.014526 (0.047024) | 0.074200 / 0.176557 (-0.102357) | 0.120250 / 0.737135 (-0.616885) | 0.076018 / 0.296338 (-0.220321) |\n\n### Benchmark: benchmark_iterating.json\n\n| metric | read 5000 | read 50000 | read_batch 50000 10 | read_batch 50000 100 | read_batch 50000 1000 | read_formatted numpy 5000 | read_formatted pandas 5000 | read_formatted tensorflow 5000 | read_formatted torch 5000 | read_formatted_batch numpy 5000 10 | read_formatted_batch numpy 5000 1000 | shuffled read 5000 | shuffled read 50000 | shuffled read_batch 50000 10 | shuffled read_batch 50000 100 | shuffled read_batch 50000 1000 | shuffled read_formatted numpy 5000 | shuffled read_formatted_batch numpy 5000 10 | shuffled read_formatted_batch numpy 5000 1000 |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.302517 / 0.215209 (0.087308) | 2.943936 / 2.077655 (0.866282) | 1.584847 / 1.504120 (0.080727) | 1.464501 / 1.541195 (-0.076694) | 1.472402 / 1.468490 (0.003912) | 0.570971 / 4.584777 (-4.013806) | 2.383207 / 3.745712 (-1.362505) | 2.811520 / 5.269862 (-2.458342) | 1.746997 / 4.565676 (-2.818680) | 0.063391 / 0.424275 (-0.360884) | 0.005296 / 0.007607 (-0.002311) | 0.358948 / 0.226044 (0.132903) | 3.604704 / 2.268929 (1.335776) | 1.935813 / 55.444624 (-53.508812) | 1.659944 / 6.876477 (-5.216533) | 1.687151 / 2.142072 (-0.454922) | 0.658044 / 4.805227 (-4.147183) | 0.120425 / 6.500664 (-6.380240) | 0.042694 / 0.075469 (-0.032775) |\n\n### Benchmark: benchmark_map_filter.json\n\n| metric | filter | map fast-tokenizer batched | map identity | map identity batched | map no-op batched | map no-op batched numpy | map no-op batched pandas | map no-op batched pytorch | map no-op batched tensorflow |\n|--------|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.986308 / 1.841788 (-0.855479) | 11.727945 / 8.074308 (3.653637) | 9.532785 / 10.191392 (-0.658607) | 0.140071 / 0.680424 (-0.540352) | 0.013472 / 0.534201 (-0.520729) | 0.285828 / 0.579283 (-0.293455) | 0.261571 / 0.434364 (-0.172793) | 0.323114 / 0.540337 (-0.217223) | 0.418132 / 1.386936 (-0.968804) |\n\n</details>\nPyArrow==latest\n\n<details>\n<summary>Show updated benchmarks!</summary>\n\n### Benchmark: benchmark_array_xd.json\n\n| metric | read_batch_formatted_as_numpy after write_array2d | read_batch_formatted_as_numpy after write_flattened_sequence | read_batch_formatted_as_numpy after write_nested_sequence | read_batch_unformated after write_array2d | read_batch_unformated after write_flattened_sequence | read_batch_unformated after write_nested_sequence | read_col_formatted_as_numpy after write_array2d | read_col_formatted_as_numpy after write_flattened_sequence | read_col_formatted_as_numpy after write_nested_sequence | read_col_unformated after write_array2d | read_col_unformated after write_flattened_sequence | read_col_unformated after write_nested_sequence | read_formatted_as_numpy after write_array2d | read_formatted_as_numpy after write_flattened_sequence | read_formatted_as_numpy after write_nested_sequence | read_unformated after write_array2d | read_unformated after write_flattened_sequence | read_unformated after write_nested_sequence | write_array2d | write_flattened_sequence | write_nested_sequence |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.005428 / 0.011353 (-0.005925) | 0.003954 / 0.011008 (-0.007054) | 0.050336 / 0.038508 (0.011828) | 0.029941 / 0.023109 (0.006831) | 0.281483 / 0.275898 (0.005585) | 0.304822 / 0.323480 (-0.018658) | 0.004151 / 0.007986 (-0.003835) | 0.002862 / 0.004328 (-0.001466) | 0.049196 / 0.004250 (0.044945) | 0.040266 / 0.037052 (0.003213) | 0.293515 / 0.258489 (0.035026) | 0.319165 / 0.293841 (0.025324) | 0.029186 / 0.128546 (-0.099360) | 0.010838 / 0.075646 (-0.064809) | 0.058789 / 0.419271 (-0.360483) | 0.032847 / 0.043533 (-0.010686) | 0.280164 / 0.255139 (0.025025) | 0.299609 / 0.283200 (0.016410) | 0.018291 / 0.141683 (-0.123392) | 1.153858 / 1.452155 (-0.298297) | 1.219108 / 1.492716 (-0.273608) |\n\n### Benchmark: benchmark_getitem\\_100B.json\n\n| metric | get_batch_of\\_1024\\_random_rows | get_batch_of\\_1024\\_rows | get_first_row | get_last_row |\n|--------|---|---|---|---|\n| new / old (diff) | 0.093783 / 0.018006 (0.075777) | 0.301526 / 0.000490 (0.301037) | 0.000211 / 0.000200 (0.000011) | 0.000055 / 0.000054 (0.000001) |\n\n### Benchmark: benchmark_indices_mapping.json\n\n| metric | select | shard | shuffle | sort | train_test_split |\n|--------|---|---|---|---|---|\n| new / old (diff) | 0.022105 / 0.037411 (-0.015306) | 0.074844 / 0.014526 (0.060318) | 0.087147 / 0.176557 (-0.089409) | 0.127678 / 0.737135 (-0.609457) | 0.088630 / 0.296338 (-0.207709) |\n\n### Benchmark: benchmark_iterating.json\n\n| metric | read 5000 | read 50000 | read_batch 50000 10 | read_batch 50000 100 | read_batch 50000 1000 | read_formatted numpy 5000 | read_formatted pandas 5000 | read_formatted tensorflow 5000 | read_formatted torch 5000 | read_formatted_batch numpy 5000 10 | read_formatted_batch numpy 5000 1000 | shuffled read 5000 | shuffled read 50000 | shuffled read_batch 50000 10 | shuffled read_batch 50000 100 | shuffled read_batch 50000 1000 | shuffled read_formatted numpy 5000 | shuffled read_formatted_batch numpy 5000 10 | shuffled read_formatted_batch numpy 5000 1000 |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.286805 / 0.215209 (0.071596) | 2.828664 / 2.077655 (0.751009) | 1.579771 / 1.504120 (0.075651) | 1.463137 / 1.541195 (-0.078058) | 1.509238 / 1.468490 (0.040748) | 0.583425 / 4.584777 (-4.001352) | 2.424905 / 3.745712 (-1.320807) | 2.819354 / 5.269862 (-2.450508) | 1.784695 / 4.565676 (-2.780981) | 0.063374 / 0.424275 (-0.360901) | 0.005337 / 0.007607 (-0.002270) | 0.342291 / 0.226044 (0.116247) | 3.404319 / 2.268929 (1.135390) | 1.956909 / 55.444624 (-53.487716) | 1.694317 / 6.876477 (-5.182160) | 1.696256 / 2.142072 (-0.445817) | 0.655748 / 4.805227 (-4.149480) | 0.116785 / 6.500664 (-6.383879) | 0.040930 / 0.075469 (-0.034539) |\n\n### Benchmark: benchmark_map_filter.json\n\n| metric | filter | map fast-tokenizer batched | map identity | map identity batched | map no-op batched | map no-op batched numpy | map no-op batched pandas | map no-op batched pytorch | map no-op batched tensorflow |\n|--------|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 1.034463 / 1.841788 (-0.807325) | 12.252041 / 8.074308 (4.177733) | 10.593960 / 10.191392 (0.402568) | 0.139311 / 0.680424 (-0.541112) | 0.016177 / 0.534201 (-0.518023) | 0.288910 / 0.579283 (-0.290373) | 0.281588 / 0.434364 (-0.152776) | 0.323066 / 0.540337 (-0.217272) | 0.427604 / 1.386936 (-0.959332) |\n\n</details>\n</details>\n\n![](https://cml.dev/watermark.png#a188022dc43a76a119d90c03832d51d6e4a94d91 \"CML watermark\")\n" ]

[ "Hi! You can fix this by updating the `datasets` package with `pip install -U datasets` and restarting the notebook.\r\n", "Kaggle removed the problematic `datasets==2.1.0` pin last week, so I'm closing this issue (now it pre-installs the latest version)." ]

[ "This may be a solution that only changes `cast_storage` of `Image`.\r\nHowever, I'm not totally sure that the assumptions hold that are made about the `ListArray`.\r\n\r\n```python\r\nelif pa.types.is_list(storage.type):\r\n from .features import Array3DExtensionType\r\n\r\n def get_shapes(arr):\r\n shape = ()\r\n while isinstance(arr, pa.ListArray):\r\n len_curr = len(arr)\r\n arr = arr.flatten()\r\n len_new = len(arr)\r\n shape = shape + (len_new // len_curr,)\r\n return shape\r\n\r\n def get_dtypes(arr):\r\n dtype = storage.type\r\n while hasattr(dtype, \"value_type\"):\r\n dtype = dtype.value_type\r\n return dtype\r\n\r\n arrays = []\r\n for i, is_null in enumerate(storage.is_null()):\r\n if not is_null.as_py():\r\n storage_part = storage.take([i])\r\n shape = get_shapes(storage_part)\r\n dtype = get_dtypes(storage_part)\r\n\r\n extension_type = Array3DExtensionType(shape=shape, dtype=str(dtype))\r\n array = pa.ExtensionArray.from_storage(extension_type, storage_part)\r\n arrays.append(array.to_numpy().squeeze(0))\r\n else:\r\n arrays.append(None)\r\n\r\n bytes_array = pa.array(\r\n [encode_np_array(arr)[\"bytes\"] if arr is not None else None for arr in arrays],\r\n type=pa.binary(),\r\n )\r\n path_array = pa.array([None] * len(storage), type=pa.string())\r\n storage = pa.StructArray.from_arrays(\r\n [bytes_array, path_array], [\"bytes\", \"path\"], mask=bytes_array.is_null()\r\n )\r\n```\r\n(Edited): to handle nulls\r\n\r\nNotably this doesn't change anything about the passing through of data or other things, just in the `Image` class.\r\nSeems quite fast:\r\n```bash\r\nFri Apr 5 17:55:51 2024 restats\r\n\r\n 63818 function calls (61995 primitive calls) in 0.812 seconds\r\n\r\n Ordered by: cumulative time\r\n List reduced from 1051 to 20 due to restriction <20>\r\n\r\n ncalls tottime percall cumtime percall filename:lineno(function)\r\n 47/1 0.000 0.000 0.810 0.810 {built-in method builtins.exec}\r\n 2/1 0.000 0.000 0.810 0.810 <string>:1(<module>)\r\n 2/1 0.000 0.000 0.809 0.809 arrow_dataset.py:594(wrapper)\r\n 2/1 0.000 0.000 0.809 0.809 arrow_dataset.py:551(wrapper)\r\n 2/1 0.000 0.000 0.809 0.809 arrow_dataset.py:2916(map)\r\n 3 0.000 0.000 0.807 0.269 arrow_dataset.py:3277(_map_single)\r\n 1 0.000 0.000 0.760 0.760 arrow_writer.py:589(finalize)\r\n 1 0.000 0.000 0.760 0.760 arrow_writer.py:423(write_examples_on_file)\r\n 1 0.000 0.000 0.759 0.759 arrow_writer.py:527(write_batch)\r\n 1 0.001 0.001 0.754 0.754 arrow_writer.py:161(__arrow_array__)\r\n 2/1 0.000 0.000 0.719 0.719 table.py:1800(wrapper)\r\n 1 0.000 0.000 0.719 0.719 table.py:1950(cast_array_to_feature)\r\n 1 0.006 0.006 0.718 0.718 image.py:209(cast_storage)\r\n 1 0.000 0.000 0.451 0.451 image.py:361(encode_np_array)\r\n 1 0.000 0.000 0.444 0.444 image.py:343(image_to_bytes)\r\n 1 0.000 0.000 0.413 0.413 Image.py:2376(save)\r\n 1 0.000 0.000 0.413 0.413 PngImagePlugin.py:1233(_save)\r\n 1 0.000 0.000 0.413 0.413 ImageFile.py:517(_save)\r\n 1 0.000 0.000 0.413 0.413 ImageFile.py:545(_encode_tile)\r\n 397 0.409 0.001 0.409 0.001 {method 'encode' of 'ImagingEncoder' objects}\r\n```", "Also encounter this problem. Has been strugging with it for a long time...", "This actually applies to all arrays (numpy or tensors like in torch), not only from external files.\r\n```python\r\nimport numpy as np\r\nimport datasets\r\n\r\nds = datasets.Dataset.from_dict(\r\n {\"image\": [np.random.randint(0, 255, (2048, 2048, 3), dtype=np.uint8)]},\r\n features=datasets.Features({\"image\": datasets.Image(decode=True)}),\r\n)\r\nds.set_format(\"numpy\")\r\n\r\nds = ds.map(load_from_cache_file=False)\r\n```" ]