Fix offline mode with single config

[lhoestq]

Reported in #4760

The cache was not able to reload a dataset with a single config form the cache if the config name is not specificed

For example

from datasets import load_dataset, config

config.HF_DATASETS_OFFLINE = True

load_dataset("openai_humaneval")

This was due to a regression in #6632

[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.

[mariosasko]

One nit. Looks good otherwise.

[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)	0.005093 / 0.011353 (-0.006260)	0.003317 / 0.011008 (-0.007692)	0.064795 / 0.038508 (0.026287)	0.030373 / 0.023109 (0.007263)	0.258776 / 0.275898 (-0.017122)	0.269768 / 0.323480 (-0.053711)	0.004186 / 0.007986 (-0.003799)	0.002630 / 0.004328 (-0.001699)	0.048643 / 0.004250 (0.044392)	0.044220 / 0.037052 (0.007168)	0.265113 / 0.258489 (0.006624)	0.292202 / 0.293841 (-0.001639)	0.027468 / 0.128546 (-0.101079)	0.010123 / 0.075646 (-0.065523)	0.226869 / 0.419271 (-0.192402)	0.035739 / 0.043533 (-0.007794)	0.253193 / 0.255139 (-0.001946)	0.271002 / 0.283200 (-0.012198)	0.017201 / 0.141683 (-0.124482)	1.105836 / 1.452155 (-0.346318)	1.161559 / 1.492716 (-0.331158)

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)	0.090481 / 0.018006 (0.072475)	0.299013 / 0.000490 (0.298524)	0.000220 / 0.000200 (0.000020)	0.000047 / 0.000054 (-0.000007)

Benchmark: benchmark_indices_mapping.json

metric	select	shard	shuffle	sort	train_test_split
new / old (diff)	0.017684 / 0.037411 (-0.019727)	0.061580 / 0.014526 (0.047054)	0.074370 / 0.176557 (-0.102186)	0.119468 / 0.737135 (-0.617667)	0.074671 / 0.296338 (-0.221668)

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)	0.284778 / 0.215209 (0.069569)	2.780241 / 2.077655 (0.702586)	1.504025 / 1.504120 (-0.000095)	1.386644 / 1.541195 (-0.154550)	1.402038 / 1.468490 (-0.066452)	0.555180 / 4.584777 (-4.029597)	2.410973 / 3.745712 (-1.334740)	2.773252 / 5.269862 (-2.496610)	1.722784 / 4.565676 (-2.842892)	0.062773 / 0.424275 (-0.361502)	0.004959 / 0.007607 (-0.002648)	0.337163 / 0.226044 (0.111119)	3.356947 / 2.268929 (1.088019)	1.880953 / 55.444624 (-53.563671)	1.556049 / 6.876477 (-5.320427)	1.578589 / 2.142072 (-0.563483)	0.641993 / 4.805227 (-4.163234)	0.118624 / 6.500664 (-6.382040)	0.042202 / 0.075469 (-0.033268)

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)	0.995321 / 1.841788 (-0.846467)	12.257597 / 8.074308 (4.183289)	9.646214 / 10.191392 (-0.545178)	0.131124 / 0.680424 (-0.549300)	0.014119 / 0.534201 (-0.520082)	0.287597 / 0.579283 (-0.291686)	0.266983 / 0.434364 (-0.167381)	0.328165 / 0.540337 (-0.212173)	0.422405 / 1.386936 (-0.964531)

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)	0.005091 / 0.011353 (-0.006262)	0.003358 / 0.011008 (-0.007650)	0.049136 / 0.038508 (0.010628)	0.031075 / 0.023109 (0.007966)	0.275047 / 0.275898 (-0.000851)	0.296845 / 0.323480 (-0.026635)	0.004949 / 0.007986 (-0.003037)	0.002586 / 0.004328 (-0.001743)	0.048164 / 0.004250 (0.043913)	0.040754 / 0.037052 (0.003702)	0.288715 / 0.258489 (0.030226)	0.312383 / 0.293841 (0.018542)	0.029372 / 0.128546 (-0.099174)	0.010097 / 0.075646 (-0.065549)	0.056752 / 0.419271 (-0.362520)	0.033128 / 0.043533 (-0.010405)	0.274986 / 0.255139 (0.019847)	0.292692 / 0.283200 (0.009493)	0.018309 / 0.141683 (-0.123374)	1.190320 / 1.452155 (-0.261834)	1.222529 / 1.492716 (-0.270188)

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)	0.091717 / 0.018006 (0.073711)	0.300278 / 0.000490 (0.299788)	0.000217 / 0.000200 (0.000017)	0.000065 / 0.000054 (0.000011)

Benchmark: benchmark_indices_mapping.json

metric	select	shard	shuffle	sort	train_test_split
new / old (diff)	0.021394 / 0.037411 (-0.016018)	0.074918 / 0.014526 (0.060392)	0.087461 / 0.176557 (-0.089095)	0.125499 / 0.737135 (-0.611636)	0.087484 / 0.296338 (-0.208854)

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)	0.296557 / 0.215209 (0.081348)	2.905527 / 2.077655 (0.827872)	1.624640 / 1.504120 (0.120520)	1.505495 / 1.541195 (-0.035700)	1.514066 / 1.468490 (0.045576)	0.569376 / 4.584777 (-4.015401)	2.448575 / 3.745712 (-1.297137)	2.772805 / 5.269862 (-2.497057)	1.757287 / 4.565676 (-2.808390)	0.064209 / 0.424275 (-0.360066)	0.005688 / 0.007607 (-0.001919)	0.353175 / 0.226044 (0.127131)	3.481591 / 2.268929 (1.212662)	1.995384 / 55.444624 (-53.449240)	1.684623 / 6.876477 (-5.191854)	1.675750 / 2.142072 (-0.466323)	0.644463 / 4.805227 (-4.160764)	0.115393 / 6.500664 (-6.385271)	0.040671 / 0.075469 (-0.034799)

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)	1.037487 / 1.841788 (-0.804301)	11.902194 / 8.074308 (3.827886)	10.148579 / 10.191392 (-0.042813)	0.150261 / 0.680424 (-0.530163)	0.015001 / 0.534201 (-0.519200)	0.291008 / 0.579283 (-0.288275)	0.278758 / 0.434364 (-0.155606)	0.334037 / 0.540337 (-0.206301)	0.419942 / 1.386936 (-0.966994)