Suggest scikit-learn instead of sklearn

[osbm]

This is kinda unimportant fix but, the suggested pip install sklearn does not work.

The current error message if sklearn is not installed:

ImportError: To be able to use [dataset name], you need to install the following dependency: sklearn.
Please install it using 'pip install sklearn' for instance.

[julien-c]

good catch!

[lhoestq]

indeed thanks !

[HuggingFaceDocBuilderDev]

The documentation is not available anymore as the PR was closed or merged.

[lhoestq]

The test fail is unrelated to this PR and fixed on main - merging :)

[github-actions]

Show benchmarks

PyArrow==6.0.0

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)	0.008942 / 0.011353 (-0.002411)	0.004617 / 0.011008 (-0.006391)	0.101310 / 0.038508 (0.062802)	0.030997 / 0.023109 (0.007888)	0.306292 / 0.275898 (0.030394)	0.370533 / 0.323480 (0.047053)	0.007318 / 0.007986 (-0.000667)	0.003473 / 0.004328 (-0.000856)	0.078557 / 0.004250 (0.074307)	0.036312 / 0.037052 (-0.000740)	0.308993 / 0.258489 (0.050504)	0.344411 / 0.293841 (0.050570)	0.034384 / 0.128546 (-0.094162)	0.011631 / 0.075646 (-0.064016)	0.323948 / 0.419271 (-0.095324)	0.041176 / 0.043533 (-0.002357)	0.302512 / 0.255139 (0.047373)	0.322439 / 0.283200 (0.039239)	0.088955 / 0.141683 (-0.052728)	1.534918 / 1.452155 (0.082763)	1.555803 / 1.492716 (0.063087)

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)	0.195639 / 0.018006 (0.177633)	0.423068 / 0.000490 (0.422579)	0.004101 / 0.000200 (0.003901)	0.000079 / 0.000054 (0.000025)

Benchmark: benchmark_indices_mapping.json

metric	select	shard	shuffle	sort	train_test_split
new / old (diff)	0.023691 / 0.037411 (-0.013721)	0.100536 / 0.014526 (0.086011)	0.108399 / 0.176557 (-0.068157)	0.143515 / 0.737135 (-0.593620)	0.111886 / 0.296338 (-0.184452)

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)	0.417519 / 0.215209 (0.202310)	4.180463 / 2.077655 (2.102808)	1.862511 / 1.504120 (0.358391)	1.658724 / 1.541195 (0.117529)	1.735847 / 1.468490 (0.267357)	0.688257 / 4.584777 (-3.896520)	3.447976 / 3.745712 (-0.297737)	1.877939 / 5.269862 (-3.391922)	1.157385 / 4.565676 (-3.408292)	0.081418 / 0.424275 (-0.342857)	0.012395 / 0.007607 (0.004788)	0.518935 / 0.226044 (0.292891)	5.220355 / 2.268929 (2.951427)	2.308355 / 55.444624 (-53.136269)	1.960026 / 6.876477 (-4.916450)	2.013179 / 2.142072 (-0.128893)	0.802850 / 4.805227 (-4.002377)	0.146941 / 6.500664 (-6.353723)	0.064080 / 0.075469 (-0.011389)

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)	1.284443 / 1.841788 (-0.557344)	13.903755 / 8.074308 (5.829447)	14.467101 / 10.191392 (4.275709)	0.156813 / 0.680424 (-0.523611)	0.028583 / 0.534201 (-0.505618)	0.406349 / 0.579283 (-0.172934)	0.413178 / 0.434364 (-0.021186)	0.491283 / 0.540337 (-0.049055)	0.571171 / 1.386936 (-0.815765)

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)	0.006868 / 0.011353 (-0.004484)	0.004593 / 0.011008 (-0.006416)	0.077574 / 0.038508 (0.039066)	0.027703 / 0.023109 (0.004593)	0.342096 / 0.275898 (0.066198)	0.378500 / 0.323480 (0.055020)	0.005785 / 0.007986 (-0.002201)	0.003342 / 0.004328 (-0.000986)	0.076105 / 0.004250 (0.071855)	0.040369 / 0.037052 (0.003317)	0.343611 / 0.258489 (0.085122)	0.391859 / 0.293841 (0.098018)	0.032675 / 0.128546 (-0.095871)	0.011623 / 0.075646 (-0.064023)	0.086623 / 0.419271 (-0.332648)	0.051955 / 0.043533 (0.008423)	0.343425 / 0.255139 (0.088286)	0.368887 / 0.283200 (0.085688)	0.097117 / 0.141683 (-0.044566)	1.499546 / 1.452155 (0.047391)	1.593100 / 1.492716 (0.100383)

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)	0.193568 / 0.018006 (0.175562)	0.409211 / 0.000490 (0.408722)	0.003797 / 0.000200 (0.003597)	0.000083 / 0.000054 (0.000029)

Benchmark: benchmark_indices_mapping.json

metric	select	shard	shuffle	sort	train_test_split
new / old (diff)	0.024982 / 0.037411 (-0.012430)	0.101367 / 0.014526 (0.086841)	0.108546 / 0.176557 (-0.068010)	0.144402 / 0.737135 (-0.592733)	0.112233 / 0.296338 (-0.184105)

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)	0.432820 / 0.215209 (0.217611)	4.341045 / 2.077655 (2.263391)	2.058326 / 1.504120 (0.554207)	1.853913 / 1.541195 (0.312718)	1.942436 / 1.468490 (0.473946)	0.699130 / 4.584777 (-3.885647)	3.392879 / 3.745712 (-0.352833)	1.908277 / 5.269862 (-3.361585)	1.177998 / 4.565676 (-3.387678)	0.082700 / 0.424275 (-0.341576)	0.012505 / 0.007607 (0.004898)	0.526286 / 0.226044 (0.300242)	5.279599 / 2.268929 (3.010670)	2.505771 / 55.444624 (-52.938854)	2.158460 / 6.876477 (-4.718016)	2.211437 / 2.142072 (0.069365)	0.802065 / 4.805227 (-4.003163)	0.150766 / 6.500664 (-6.349898)	0.067639 / 0.075469 (-0.007830)

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)	1.286595 / 1.841788 (-0.555192)	13.961894 / 8.074308 (5.887586)	14.021865 / 10.191392 (3.830473)	0.164590 / 0.680424 (-0.515834)	0.016909 / 0.534201 (-0.517292)	0.392215 / 0.579283 (-0.187069)	0.408080 / 0.434364 (-0.026284)	0.488247 / 0.540337 (-0.052090)	0.575524 / 1.386936 (-0.811412)