Remove default trust_remote_code=True

[lhoestq]

TODO:

• fix tests

[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]

yay! 🎉

[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)	0.004881 / 0.011353 (-0.006472)	0.003246 / 0.011008 (-0.007762)	0.062496 / 0.038508 (0.023988)	0.030760 / 0.023109 (0.007651)	0.241500 / 0.275898 (-0.034398)	0.272073 / 0.323480 (-0.051407)	0.004123 / 0.007986 (-0.003863)	0.002796 / 0.004328 (-0.001533)	0.049015 / 0.004250 (0.044764)	0.047095 / 0.037052 (0.010043)	0.257002 / 0.258489 (-0.001487)	0.287602 / 0.293841 (-0.006239)	0.027281 / 0.128546 (-0.101265)	0.010132 / 0.075646 (-0.065514)	0.203699 / 0.419271 (-0.215572)	0.036553 / 0.043533 (-0.006980)	0.246221 / 0.255139 (-0.008918)	0.268137 / 0.283200 (-0.015062)	0.017260 / 0.141683 (-0.124423)	1.100677 / 1.452155 (-0.351478)	1.148367 / 1.492716 (-0.344349)

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)	0.102519 / 0.018006 (0.084513)	0.301929 / 0.000490 (0.301439)	0.000223 / 0.000200 (0.000023)	0.000046 / 0.000054 (-0.000009)

Benchmark: benchmark_indices_mapping.json

metric	select	shard	shuffle	sort	train_test_split
new / old (diff)	0.018590 / 0.037411 (-0.018821)	0.061615 / 0.014526 (0.047089)	0.074579 / 0.176557 (-0.101978)	0.121415 / 0.737135 (-0.615720)	0.075696 / 0.296338 (-0.220642)

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)	0.283842 / 0.215209 (0.068633)	2.788321 / 2.077655 (0.710666)	1.481376 / 1.504120 (-0.022743)	1.356064 / 1.541195 (-0.185131)	1.380592 / 1.468490 (-0.087898)	0.575577 / 4.584777 (-4.009199)	2.471858 / 3.745712 (-1.273854)	2.760769 / 5.269862 (-2.509093)	1.808638 / 4.565676 (-2.757038)	0.064930 / 0.424275 (-0.359345)	0.005056 / 0.007607 (-0.002551)	0.337794 / 0.226044 (0.111750)	3.359444 / 2.268929 (1.090515)	1.829540 / 55.444624 (-53.615084)	1.518660 / 6.876477 (-5.357817)	1.671612 / 2.142072 (-0.470460)	0.664286 / 4.805227 (-4.140941)	0.119593 / 6.500664 (-6.381071)	0.042519 / 0.075469 (-0.032950)

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)	0.993152 / 1.841788 (-0.848636)	11.733054 / 8.074308 (3.658746)	9.746734 / 10.191392 (-0.444658)	0.143026 / 0.680424 (-0.537398)	0.014900 / 0.534201 (-0.519301)	0.292243 / 0.579283 (-0.287040)	0.261301 / 0.434364 (-0.173063)	0.330838 / 0.540337 (-0.209500)	0.523719 / 1.386936 (-0.863217)

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)	0.005707 / 0.011353 (-0.005646)	0.003523 / 0.011008 (-0.007485)	0.052265 / 0.038508 (0.013757)	0.034296 / 0.023109 (0.011187)	0.266589 / 0.275898 (-0.009309)	0.288441 / 0.323480 (-0.035039)	0.004507 / 0.007986 (-0.003478)	0.002745 / 0.004328 (-0.001583)	0.049417 / 0.004250 (0.045167)	0.042679 / 0.037052 (0.005627)	0.278518 / 0.258489 (0.020029)	0.328751 / 0.293841 (0.034911)	0.029530 / 0.128546 (-0.099016)	0.010373 / 0.075646 (-0.065274)	0.058207 / 0.419271 (-0.361064)	0.033434 / 0.043533 (-0.010099)	0.267902 / 0.255139 (0.012763)	0.288192 / 0.283200 (0.004993)	0.018866 / 0.141683 (-0.122817)	1.132734 / 1.452155 (-0.319421)	1.172879 / 1.492716 (-0.319837)

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)	0.097787 / 0.018006 (0.079780)	0.305509 / 0.000490 (0.305019)	0.000268 / 0.000200 (0.000068)	0.000060 / 0.000054 (0.000006)

Benchmark: benchmark_indices_mapping.json

metric	select	shard	shuffle	sort	train_test_split
new / old (diff)	0.023230 / 0.037411 (-0.014181)	0.076637 / 0.014526 (0.062111)	0.088386 / 0.176557 (-0.088171)	0.131079 / 0.737135 (-0.606057)	0.091142 / 0.296338 (-0.205197)

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)	0.295586 / 0.215209 (0.080377)	2.872090 / 2.077655 (0.794435)	1.538152 / 1.504120 (0.034032)	1.405695 / 1.541195 (-0.135500)	1.421058 / 1.468490 (-0.047432)	0.561179 / 4.584777 (-4.023598)	0.943954 / 3.745712 (-2.801758)	2.684381 / 5.269862 (-2.585481)	1.757457 / 4.565676 (-2.808220)	0.062903 / 0.424275 (-0.361372)	0.004998 / 0.007607 (-0.002610)	0.370290 / 0.226044 (0.144245)	3.374988 / 2.268929 (1.106059)	1.899282 / 55.444624 (-53.545342)	1.598787 / 6.876477 (-5.277690)	1.735371 / 2.142072 (-0.406702)	0.647367 / 4.805227 (-4.157860)	0.116975 / 6.500664 (-6.383689)	0.040811 / 0.075469 (-0.034658)

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)	0.996380 / 1.841788 (-0.845408)	12.225657 / 8.074308 (4.151349)	10.291221 / 10.191392 (0.099829)	0.142791 / 0.680424 (-0.537633)	0.016087 / 0.534201 (-0.518114)	0.299978 / 0.579283 (-0.279305)	0.149444 / 0.434364 (-0.284920)	0.321354 / 0.540337 (-0.218984)	0.414492 / 1.386936 (-0.972444)

[henilp105]

@lhoestq Thanks for the PR, Is there a way to detect if trust_remote_code=True will be required for loading the dataset, without loading it? It would be great if you could please point me to the relevant documentation.

[lhoestq]

You can check the presence of a python loading script in the repository.

If there is a .py file named after the repository name, then it requires trust_remote_code.

[henilp105]

Thanks @lhoestq for the reference.