Skip to content

Commit

Permalink
Use scikit-learn package rather than sklearn (#2525)
Browse files Browse the repository at this point in the history
  • Loading branch information
@lesteve
lesteve authored Jun 21, 2021
1 parent a0b42ae commit 445a247
Showing 1 changed file with 1 addition and 1 deletion.
2 changes: 1 addition & 1 deletion setup.py
View file
Original file line number Diff line number Diff line change
Expand Up @@ -143,7 +143,7 @@
"sacrebleu",
"scipy",
"seqeval",
"sklearn",
"scikit-learn",
"jiwer",
"sentencepiece", # for bleurt
# to speed up pip backtracking
Expand Down

1 comment on commit 445a247

@github-actions
Copy link

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Show benchmarks

PyArrow==1.0.0

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) 0.023751 / 0.011353 (0.012398) 0.016401 / 0.011008 (0.005393) 0.053577 / 0.038508 (0.015069) 0.128966 / 0.023109 (0.105857) 0.410613 / 0.275898 (0.134715) 0.460098 / 0.323480 (0.136618) 0.012769 / 0.007986 (0.004783) 0.005268 / 0.004328 (0.000940) 0.012365 / 0.004250 (0.008115) 0.058822 / 0.037052 (0.021769) 0.413957 / 0.258489 (0.155468) 0.457767 / 0.293841 (0.163926) 0.158759 / 0.128546 (0.030213) 0.123252 / 0.075646 (0.047606) 0.466792 / 0.419271 (0.047521) 0.439973 / 0.043533 (0.396440) 0.413161 / 0.255139 (0.158022) 0.451439 / 0.283200 (0.168239) 3.931899 / 0.141683 (3.790216) 1.988839 / 1.452155 (0.536685) 2.153834 / 1.492716 (0.661117)

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) 0.017959 / 0.018006 (-0.000047) 0.491901 / 0.000490 (0.491412) 0.010932 / 0.000200 (0.010732) 0.000264 / 0.000054 (0.000209)

Benchmark: benchmark_indices_mapping.json

metric select shard shuffle sort train_test_split
new / old (diff) 0.047599 / 0.037411 (0.010188) 0.028040 / 0.014526 (0.013514) 0.029531 / 0.176557 (-0.147026) 0.053212 / 0.737135 (-0.683923) 0.031021 / 0.296338 (-0.265318)

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) 0.452980 / 0.215209 (0.237771) 4.335218 / 2.077655 (2.257564) 2.205294 / 1.504120 (0.701174) 1.978496 / 1.541195 (0.437302) 2.019904 / 1.468490 (0.551414) 6.780109 / 4.584777 (2.195332) 6.090471 / 3.745712 (2.344759) 8.509737 / 5.269862 (3.239876) 7.515629 / 4.565676 (2.949952) 0.660017 / 0.424275 (0.235742) 0.011062 / 0.007607 (0.003455) 0.562494 / 0.226044 (0.336449) 5.669889 / 2.268929 (3.400961) 2.729604 / 55.444624 (-52.715020) 2.328334 / 6.876477 (-4.548143) 2.348594 / 2.142072 (0.206522) 6.920262 / 4.805227 (2.115035) 4.039199 / 6.500664 (-2.461465) 10.302771 / 0.075469 (10.227302)

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) 12.986801 / 1.841788 (11.145013) 15.388461 / 8.074308 (7.314152) 31.669372 / 10.191392 (21.477980) 1.014943 / 0.680424 (0.334519) 0.670237 / 0.534201 (0.136036) 0.780410 / 0.579283 (0.201127) 0.605094 / 0.434364 (0.170730) 0.709878 / 0.540337 (0.169541) 1.594906 / 1.386936 (0.207969)
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) 0.023502 / 0.011353 (0.012149) 0.016703 / 0.011008 (0.005695) 0.056550 / 0.038508 (0.018042) 0.045181 / 0.023109 (0.022072) 0.364072 / 0.275898 (0.088173) 0.400476 / 0.323480 (0.076996) 0.012127 / 0.007986 (0.004141) 0.005513 / 0.004328 (0.001184) 0.013899 / 0.004250 (0.009649) 0.056474 / 0.037052 (0.019421) 0.358348 / 0.258489 (0.099858) 0.410368 / 0.293841 (0.116527) 0.154354 / 0.128546 (0.025807) 0.129130 / 0.075646 (0.053483) 0.474623 / 0.419271 (0.055352) 0.448248 / 0.043533 (0.404715) 0.357547 / 0.255139 (0.102408) 0.385331 / 0.283200 (0.102131) 1.803773 / 0.141683 (1.662090) 1.935320 / 1.452155 (0.483165) 1.999285 / 1.492716 (0.506569)

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) 0.078274 / 0.018006 (0.060268) 0.492878 / 0.000490 (0.492389) 0.030984 / 0.000200 (0.030784) 0.000287 / 0.000054 (0.000232)

Benchmark: benchmark_indices_mapping.json

metric select shard shuffle sort train_test_split
new / old (diff) 0.045521 / 0.037411 (0.008110) 0.028962 / 0.014526 (0.014436) 0.031314 / 0.176557 (-0.145243) 0.051099 / 0.737135 (-0.686037) 0.034080 / 0.296338 (-0.262259)

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) 0.431235 / 0.215209 (0.216026) 4.324208 / 2.077655 (2.246553) 2.157733 / 1.504120 (0.653613) 1.933757 / 1.541195 (0.392562) 1.987467 / 1.468490 (0.518977) 6.647840 / 4.584777 (2.063063) 5.877640 / 3.745712 (2.131928) 8.416458 / 5.269862 (3.146596) 7.391484 / 4.565676 (2.825808) 0.651010 / 0.424275 (0.226735) 0.010596 / 0.007607 (0.002988) 0.550287 / 0.226044 (0.324243) 5.478978 / 2.268929 (3.210050) 2.659269 / 55.444624 (-52.785355) 2.255290 / 6.876477 (-4.621187) 2.301055 / 2.142072 (0.158983) 6.898066 / 4.805227 (2.092838) 4.889392 / 6.500664 (-1.611272) 7.573463 / 0.075469 (7.497994)

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) 13.130306 / 1.841788 (11.288519) 15.909811 / 8.074308 (7.835503) 30.779850 / 10.191392 (20.588458) 0.821065 / 0.680424 (0.140641) 0.628765 / 0.534201 (0.094564) 0.781869 / 0.579283 (0.202586) 0.622963 / 0.434364 (0.188599) 0.714142 / 0.540337 (0.173804) 1.627937 / 1.386936 (0.241001)

CML watermark

Please sign in to comment.