Support standalone yaml

[lhoestq]

see (internal) https://huggingface.slack.com/archives/C02V51Q3800/p1703885853581679

[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.

[d710055071]

@lhoestq
hello
I think it should be defined in config.py
DATASET_ README_ FILENAME="README. md"
This can replace all "README. md"

[lhoestq]

Thanks for the feedback :) merging now

[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)	0.004890 / 0.011353 (-0.006463)	0.003535 / 0.011008 (-0.007473)	0.062894 / 0.038508 (0.024386)	0.029133 / 0.023109 (0.006024)	0.242387 / 0.275898 (-0.033511)	0.262720 / 0.323480 (-0.060760)	0.002880 / 0.007986 (-0.005106)	0.002674 / 0.004328 (-0.001655)	0.048932 / 0.004250 (0.044682)	0.041669 / 0.037052 (0.004617)	0.255922 / 0.258489 (-0.002567)	0.282106 / 0.293841 (-0.011734)	0.028137 / 0.128546 (-0.100409)	0.010620 / 0.075646 (-0.065026)	0.207799 / 0.419271 (-0.211473)	0.035499 / 0.043533 (-0.008034)	0.246158 / 0.255139 (-0.008981)	0.262671 / 0.283200 (-0.020528)	0.017297 / 0.141683 (-0.124386)	1.118681 / 1.452155 (-0.333474)	1.156732 / 1.492716 (-0.335985)

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)	0.091670 / 0.018006 (0.073664)	0.300327 / 0.000490 (0.299837)	0.000212 / 0.000200 (0.000012)	0.000042 / 0.000054 (-0.000012)

Benchmark: benchmark_indices_mapping.json

metric	select	shard	shuffle	sort	train_test_split
new / old (diff)	0.018080 / 0.037411 (-0.019332)	0.060357 / 0.014526 (0.045831)	0.072221 / 0.176557 (-0.104336)	0.119281 / 0.737135 (-0.617855)	0.073861 / 0.296338 (-0.222477)

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)	0.289848 / 0.215209 (0.074639)	2.845203 / 2.077655 (0.767549)	1.531271 / 1.504120 (0.027152)	1.366110 / 1.541195 (-0.175085)	1.395041 / 1.468490 (-0.073449)	0.563353 / 4.584777 (-4.021424)	2.389074 / 3.745712 (-1.356638)	2.752960 / 5.269862 (-2.516901)	1.715508 / 4.565676 (-2.850168)	0.063063 / 0.424275 (-0.361212)	0.004967 / 0.007607 (-0.002640)	0.340757 / 0.226044 (0.114713)	3.387667 / 2.268929 (1.118739)	1.845182 / 55.444624 (-53.599442)	1.569616 / 6.876477 (-5.306861)	1.571393 / 2.142072 (-0.570679)	0.643455 / 4.805227 (-4.161772)	0.116919 / 6.500664 (-6.383745)	0.042551 / 0.075469 (-0.032918)

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)	0.943761 / 1.841788 (-0.898027)	11.481068 / 8.074308 (3.406760)	10.422180 / 10.191392 (0.230788)	0.132015 / 0.680424 (-0.548408)	0.013932 / 0.534201 (-0.520268)	0.288340 / 0.579283 (-0.290943)	0.263695 / 0.434364 (-0.170669)	0.324459 / 0.540337 (-0.215878)	0.415204 / 1.386936 (-0.971732)

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)	0.005042 / 0.011353 (-0.006310)	0.003465 / 0.011008 (-0.007543)	0.050107 / 0.038508 (0.011599)	0.029542 / 0.023109 (0.006433)	0.273645 / 0.275898 (-0.002253)	0.293661 / 0.323480 (-0.029818)	0.004099 / 0.007986 (-0.003887)	0.002667 / 0.004328 (-0.001661)	0.048281 / 0.004250 (0.044030)	0.044406 / 0.037052 (0.007353)	0.284245 / 0.258489 (0.025756)	0.312303 / 0.293841 (0.018462)	0.030057 / 0.128546 (-0.098489)	0.010675 / 0.075646 (-0.064971)	0.058404 / 0.419271 (-0.360868)	0.051874 / 0.043533 (0.008342)	0.273308 / 0.255139 (0.018169)	0.289356 / 0.283200 (0.006157)	0.018628 / 0.141683 (-0.123055)	1.148764 / 1.452155 (-0.303391)	1.194181 / 1.492716 (-0.298535)

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)	0.091383 / 0.018006 (0.073376)	0.300221 / 0.000490 (0.299731)	0.000232 / 0.000200 (0.000032)	0.000044 / 0.000054 (-0.000011)

Benchmark: benchmark_indices_mapping.json

metric	select	shard	shuffle	sort	train_test_split
new / old (diff)	0.021814 / 0.037411 (-0.015597)	0.076420 / 0.014526 (0.061894)	0.087404 / 0.176557 (-0.089152)	0.126184 / 0.737135 (-0.610951)	0.089738 / 0.296338 (-0.206600)

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)	0.299839 / 0.215209 (0.084630)	2.929260 / 2.077655 (0.851605)	1.608327 / 1.504120 (0.104207)	1.479757 / 1.541195 (-0.061437)	1.494768 / 1.468490 (0.026278)	0.563873 / 4.584777 (-4.020904)	2.434442 / 3.745712 (-1.311270)	2.641384 / 5.269862 (-2.628478)	1.724222 / 4.565676 (-2.841454)	0.062125 / 0.424275 (-0.362150)	0.004994 / 0.007607 (-0.002613)	0.350895 / 0.226044 (0.124851)	3.448550 / 2.268929 (1.179621)	1.928910 / 55.444624 (-53.515714)	1.669887 / 6.876477 (-5.206590)	1.781304 / 2.142072 (-0.360768)	0.649301 / 4.805227 (-4.155926)	0.116255 / 6.500664 (-6.384409)	0.040947 / 0.075469 (-0.034522)

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)	0.977537 / 1.841788 (-0.864251)	12.119913 / 8.074308 (4.045605)	10.874078 / 10.191392 (0.682686)	0.130174 / 0.680424 (-0.550250)	0.016176 / 0.534201 (-0.518025)	0.287967 / 0.579283 (-0.291316)	0.280591 / 0.434364 (-0.153773)	0.324332 / 0.540337 (-0.216005)	0.419479 / 1.386936 (-0.967457)