Support push_to_hub canonical datasets

[albertvillanova]

Support push_to_hub canonical datasets.

This is necessary in the Space to convert script-datasets to Parquet: https://huggingface.co/spaces/albertvillanova/convert-dataset-to-parquet

Note that before this PR, the repo_id "dataset_name" was transformed to "user/dataset_name". This behavior was introduced by:

• Reduce the number of commits in push_to_hub #6269

[HuggingFaceDocBuilderDev]

The docs for this PR live here. All of your documentation changes will be reflected on that endpoint. The docs are available until 30 days after the last update.

[julien-c]

nice catch @albertvillanova

[albertvillanova]

@huggingface/datasets this PR is ready for review.

[lhoestq]

thanks for the fix !

[github-actions]

Show benchmarks

PyArrow==8.0.0

Show updated benchmarks!

Benchmark: benchmark_array_xd.json

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
new / old (diff)	0.005306 / 0.011353 (-0.006047)	0.003454 / 0.011008 (-0.007555)	0.062157 / 0.038508 (0.023649)	0.051945 / 0.023109 (0.028835)	0.241834 / 0.275898 (-0.034064)	0.265590 / 0.323480 (-0.057890)	0.003149 / 0.007986 (-0.004837)	0.002695 / 0.004328 (-0.001633)	0.049197 / 0.004250 (0.044947)	0.045576 / 0.037052 (0.008524)	0.242866 / 0.258489 (-0.015623)	0.280963 / 0.293841 (-0.012878)	0.028466 / 0.128546 (-0.100080)	0.010670 / 0.075646 (-0.064976)	0.206501 / 0.419271 (-0.212771)	0.035314 / 0.043533 (-0.008219)	0.240893 / 0.255139 (-0.014246)	0.264762 / 0.283200 (-0.018438)	0.019988 / 0.141683 (-0.121695)	1.095222 / 1.452155 (-0.356933)	1.144051 / 1.492716 (-0.348666)

Benchmark: benchmark_getitem_100B.json

metric	get_batch_of_1024_random_rows	get_batch_of_1024_rows	get_first_row	get_last_row
new / old (diff)	0.098034 / 0.018006 (0.080028)	0.308541 / 0.000490 (0.308051)	0.000261 / 0.000200 (0.000061)	0.000059 / 0.000054 (0.000004)

Benchmark: benchmark_indices_mapping.json

metric	select	shard	shuffle	sort	train_test_split
new / old (diff)	0.018646 / 0.037411 (-0.018766)	0.062881 / 0.014526 (0.048355)	0.074062 / 0.176557 (-0.102494)	0.120860 / 0.737135 (-0.616276)	0.075388 / 0.296338 (-0.220951)

Benchmark: benchmark_iterating.json

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
new / old (diff)	0.282974 / 0.215209 (0.067765)	2.755589 / 2.077655 (0.677934)	1.459536 / 1.504120 (-0.044584)	1.364543 / 1.541195 (-0.176652)	1.429860 / 1.468490 (-0.038630)	0.573277 / 4.584777 (-4.011500)	2.422983 / 3.745712 (-1.322730)	3.257258 / 5.269862 (-2.012603)	1.930053 / 4.565676 (-2.635623)	0.067476 / 0.424275 (-0.356799)	0.005612 / 0.007607 (-0.001995)	0.351538 / 0.226044 (0.125494)	3.380356 / 2.268929 (1.111427)	1.837887 / 55.444624 (-53.606738)	1.537994 / 6.876477 (-5.338483)	1.623630 / 2.142072 (-0.518442)	0.662652 / 4.805227 (-4.142576)	0.127074 / 6.500664 (-6.373590)	0.049311 / 0.075469 (-0.026158)

Benchmark: benchmark_map_filter.json

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
new / old (diff)	1.151273 / 1.841788 (-0.690515)	12.766622 / 8.074308 (4.692314)	10.967610 / 10.191392 (0.776218)	0.131305 / 0.680424 (-0.549119)	0.014227 / 0.534201 (-0.519974)	0.292054 / 0.579283 (-0.287229)	0.262737 / 0.434364 (-0.171627)	0.334360 / 0.540337 (-0.205978)	0.446711 / 1.386936 (-0.940225)

PyArrow==latest

Show updated benchmarks!

Benchmark: benchmark_array_xd.json

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
new / old (diff)	0.005194 / 0.011353 (-0.006159)	0.003508 / 0.011008 (-0.007500)	0.049287 / 0.038508 (0.010779)	0.052109 / 0.023109 (0.029000)	0.271501 / 0.275898 (-0.004397)	0.290959 / 0.323480 (-0.032521)	0.004347 / 0.007986 (-0.003638)	0.002659 / 0.004328 (-0.001669)	0.048769 / 0.004250 (0.044518)	0.039388 / 0.037052 (0.002336)	0.272811 / 0.258489 (0.014322)	0.305632 / 0.293841 (0.011791)	0.028419 / 0.128546 (-0.100127)	0.010617 / 0.075646 (-0.065029)	0.057433 / 0.419271 (-0.361838)	0.032383 / 0.043533 (-0.011149)	0.266566 / 0.255139 (0.011427)	0.290993 / 0.283200 (0.007794)	0.019939 / 0.141683 (-0.121743)	1.157623 / 1.452155 (-0.294532)	1.183298 / 1.492716 (-0.309419)

Benchmark: benchmark_getitem_100B.json

metric	get_batch_of_1024_random_rows	get_batch_of_1024_rows	get_first_row	get_last_row
new / old (diff)	0.099074 / 0.018006 (0.081068)	0.315282 / 0.000490 (0.314792)	0.000235 / 0.000200 (0.000035)	0.000057 / 0.000054 (0.000003)

Benchmark: benchmark_indices_mapping.json

metric	select	shard	shuffle	sort	train_test_split
new / old (diff)	0.022692 / 0.037411 (-0.014719)	0.076455 / 0.014526 (0.061929)	0.089094 / 0.176557 (-0.087462)	0.126407 / 0.737135 (-0.610728)	0.089588 / 0.296338 (-0.206750)

Benchmark: benchmark_iterating.json

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
new / old (diff)	0.338853 / 0.215209 (0.123644)	2.809843 / 2.077655 (0.732188)	1.538262 / 1.504120 (0.034143)	1.418290 / 1.541195 (-0.122905)	1.435145 / 1.468490 (-0.033345)	0.565763 / 4.584777 (-4.019014)	2.491525 / 3.745712 (-1.254187)	2.944879 / 5.269862 (-2.324983)	1.835840 / 4.565676 (-2.729837)	0.065101 / 0.424275 (-0.359174)	0.005196 / 0.007607 (-0.002412)	0.345291 / 0.226044 (0.119247)	3.399658 / 2.268929 (1.130729)	1.892321 / 55.444624 (-53.552303)	1.608293 / 6.876477 (-5.268184)	1.651188 / 2.142072 (-0.490884)	0.647806 / 4.805227 (-4.157421)	0.119318 / 6.500664 (-6.381346)	0.043058 / 0.075469 (-0.032412)

Benchmark: benchmark_map_filter.json

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
new / old (diff)	0.983956 / 1.841788 (-0.857831)	13.516125 / 8.074308 (5.441817)	11.712571 / 10.191392 (1.521179)	0.134253 / 0.680424 (-0.546171)	0.015844 / 0.534201 (-0.518357)	0.292444 / 0.579283 (-0.286839)	0.282182 / 0.434364 (-0.152182)	0.329327 / 0.540337 (-0.211010)	0.419960 / 1.386936 (-0.966976)