Skip to content

Commit

Permalink
Add more info to README
Browse files Browse the repository at this point in the history
  • Loading branch information
@albertvillanova
albertvillanova committed Nov 30, 2021
1 parent 71ff505 commit f7ba594
Showing 1 changed file with 35 additions and 5 deletions.
40 changes: 35 additions & 5 deletions datasets/the_pile/README.md
View file
Original file line number Diff line number Diff line change
@@ -1,6 +1,23 @@
---
YAML tags:
- copy-paste the tags obtained with the online tagging app: https://huggingface.co/spaces/huggingface/datasets-tagging
annotations_creators:
- no-annotation
language_creators:
- found
languages:
- en
licenses:
- other-
multilinguality:
- monolingual
pretty_name: The Pile
size_categories:
- unknown
source_datasets:
- original
task_categories:
- sequence-modeling
task_ids:
- language-modeling
---

# Dataset Card for The Pile
Expand Down Expand Up @@ -74,11 +91,15 @@ datasets combined together.

### Data Fields

[More Information Needed]
#### all

- `meta` (dict): Metadata of the data instance, with keys:
- pile_set_name: Name of the subset.
- `text` (str): Text.

### Data Splits

[More Information Needed]
The "all" configuration is composed of 3 splits: train, validation and test.

## Dataset Creation

Expand Down Expand Up @@ -137,7 +158,16 @@ Please refer to the specific license depending on the subset you use:

### Citation Information

[More Information Needed]
```
@misc{gao2020pile,
title={The Pile: An 800GB Dataset of Diverse Text for Language Modeling},
author={Leo Gao and Stella Biderman and Sid Black and Laurence Golding and Travis Hoppe and Charles Foster and Jason Phang and Horace He and Anish Thite and Noa Nabeshima and Shawn Presser and Connor Leahy},
year={2020},
eprint={2101.00027},
archivePrefix={arXiv},
primaryClass={cs.CL}
}
```

### Contributions

Expand Down

1 comment on commit f7ba594

@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==3.0.0

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) 0.063558 / 0.011353 (0.052205) 0.003954 / 0.011008 (-0.007054) 0.028067 / 0.038508 (-0.010441) 0.033104 / 0.023109 (0.009994) 0.272005 / 0.275898 (-0.003893) 0.308292 / 0.323480 (-0.015188) 0.079471 / 0.007986 (0.071485) 0.004344 / 0.004328 (0.000015) 0.008533 / 0.004250 (0.004283) 0.041604 / 0.037052 (0.004552) 0.274248 / 0.258489 (0.015759) 0.300725 / 0.293841 (0.006884) 0.076387 / 0.128546 (-0.052159) 0.008187 / 0.075646 (-0.067460) 0.223993 / 0.419271 (-0.195279) 0.042006 / 0.043533 (-0.001527) 0.277698 / 0.255139 (0.022559) 0.290232 / 0.283200 (0.007033) 0.082668 / 0.141683 (-0.059014) 1.530849 / 1.452155 (0.078694) 1.555617 / 1.492716 (0.062901)

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) 0.321567 / 0.018006 (0.303560) 0.556166 / 0.000490 (0.555677) 0.003024 / 0.000200 (0.002824) 0.000068 / 0.000054 (0.000013)

Benchmark: benchmark_indices_mapping.json

metric select shard shuffle sort train_test_split
new / old (diff) 0.036675 / 0.037411 (-0.000737) 0.021670 / 0.014526 (0.007144) 0.027620 / 0.176557 (-0.148936) 0.195409 / 0.737135 (-0.541726) 0.028233 / 0.296338 (-0.268106)

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) 0.419382 / 0.215209 (0.204173) 4.199591 / 2.077655 (2.121936) 1.802805 / 1.504120 (0.298685) 1.588245 / 1.541195 (0.047050) 1.699951 / 1.468490 (0.231461) 0.418840 / 4.584777 (-4.165937) 4.665442 / 3.745712 (0.919730) 2.207990 / 5.269862 (-3.061872) 0.857078 / 4.565676 (-3.708599) 0.050009 / 0.424275 (-0.374266) 0.010995 / 0.007607 (0.003387) 0.525118 / 0.226044 (0.299074) 5.249591 / 2.268929 (2.980663) 2.269069 / 55.444624 (-53.175555) 1.859147 / 6.876477 (-5.017330) 1.999940 / 2.142072 (-0.142133) 0.527286 / 4.805227 (-4.277941) 0.113717 / 6.500664 (-6.386947) 0.056548 / 0.075469 (-0.018921)

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) 1.345101 / 1.841788 (-0.496686) 11.655033 / 8.074308 (3.580725) 25.149308 / 10.191392 (14.957916) 0.613960 / 0.680424 (-0.066464) 0.463325 / 0.534201 (-0.070876) 0.327293 / 0.579283 (-0.251990) 0.454093 / 0.434364 (0.019729) 0.225763 / 0.540337 (-0.314574) 0.263231 / 1.386936 (-1.123705)
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) 0.061758 / 0.011353 (0.050405) 0.003755 / 0.011008 (-0.007253) 0.026174 / 0.038508 (-0.012335) 0.030350 / 0.023109 (0.007241) 0.260146 / 0.275898 (-0.015752) 0.309394 / 0.323480 (-0.014085) 0.084207 / 0.007986 (0.076221) 0.004928 / 0.004328 (0.000599) 0.006625 / 0.004250 (0.002374) 0.036563 / 0.037052 (-0.000489) 0.270537 / 0.258489 (0.012048) 0.319441 / 0.293841 (0.025600) 0.075403 / 0.128546 (-0.053143) 0.008054 / 0.075646 (-0.067593) 0.221797 / 0.419271 (-0.197474) 0.040748 / 0.043533 (-0.002785) 0.272521 / 0.255139 (0.017382) 0.301020 / 0.283200 (0.017821) 0.078236 / 0.141683 (-0.063447) 1.439758 / 1.452155 (-0.012397) 1.517004 / 1.492716 (0.024288)

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) 0.332442 / 0.018006 (0.314435) 0.556924 / 0.000490 (0.556434) 0.002014 / 0.000200 (0.001814) 0.000077 / 0.000054 (0.000023)

Benchmark: benchmark_indices_mapping.json

metric select shard shuffle sort train_test_split
new / old (diff) 0.034776 / 0.037411 (-0.002635) 0.020864 / 0.014526 (0.006339) 0.028089 / 0.176557 (-0.148468) 0.199649 / 0.737135 (-0.537486) 0.028232 / 0.296338 (-0.268106)

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) 0.436161 / 0.215209 (0.220952) 4.365875 / 2.077655 (2.288221) 1.942117 / 1.504120 (0.437997) 1.749678 / 1.541195 (0.208483) 1.850002 / 1.468490 (0.381512) 0.418371 / 4.584777 (-4.166406) 4.666293 / 3.745712 (0.920581) 3.462188 / 5.269862 (-1.807674) 0.868594 / 4.565676 (-3.697082) 0.050195 / 0.424275 (-0.374080) 0.011488 / 0.007607 (0.003881) 0.542630 / 0.226044 (0.316586) 5.434305 / 2.268929 (3.165376) 2.395807 / 55.444624 (-53.048817) 2.386206 / 6.876477 (-4.490270) 2.209340 / 2.142072 (0.067268) 0.533314 / 4.805227 (-4.271913) 0.114734 / 6.500664 (-6.385930) 0.057698 / 0.075469 (-0.017771)

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) 1.373281 / 1.841788 (-0.468506) 11.478102 / 8.074308 (3.403794) 23.346346 / 10.191392 (13.154954) 0.786431 / 0.680424 (0.106007) 0.467572 / 0.534201 (-0.066629) 0.327749 / 0.579283 (-0.251534) 0.501866 / 0.434364 (0.067502) 0.253749 / 0.540337 (-0.286588) 0.264613 / 1.386936 (-1.122323)

CML watermark

Please sign in to comment.