Skip to content

Commit

Permalink
Another attempt
Browse files Browse the repository at this point in the history
  • Loading branch information
@mariosasko
mariosasko committed Jul 26, 2022
1 parent feebd90 commit 59a1e3d
Showing 1 changed file with 2 additions and 2 deletions.
4 changes: 2 additions & 2 deletions setup.py
View file
Expand Up @@ -88,7 +88,7 @@
"huggingface-hub>=0.1.0,<1.0.0",
# Utilities from PyPA to e.g., compare versions
"packaging",
"responses<0.19",
"responses==0.16",
]

AUDIO_REQUIRE = [
Expand Down Expand Up @@ -120,7 +120,7 @@
"botocore>=1.22.8", # to be compatible with aiobotocore and boto3
"faiss-cpu>=1.6.4",
"fsspec[s3]",
"moto[s3,server]>=3.0.0",
"moto[s3,server]==2.0.4",
"rarfile>=4.0",
"s3fs>=2021.11.1", # aligned with fsspec[http]>=2021.11.1
"tensorflow>=2.3,!=2.6.0,!=2.6.1",
Expand Down

1 comment on commit 59a1e3d

@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==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) 0.007535 / 0.011353 (-0.003818) 0.003665 / 0.011008 (-0.007343) 0.028594 / 0.038508 (-0.009914) 0.029445 / 0.023109 (0.006336) 0.298883 / 0.275898 (0.022985) 0.368058 / 0.323480 (0.044579) 0.005409 / 0.007986 (-0.002577) 0.004204 / 0.004328 (-0.000124) 0.006600 / 0.004250 (0.002349) 0.040898 / 0.037052 (0.003846) 0.308968 / 0.258489 (0.050479) 0.354314 / 0.293841 (0.060473) 0.028681 / 0.128546 (-0.099866) 0.009288 / 0.075646 (-0.066358) 0.247136 / 0.419271 (-0.172135) 0.044598 / 0.043533 (0.001066) 0.303880 / 0.255139 (0.048741) 0.333203 / 0.283200 (0.050003) 0.084409 / 0.141683 (-0.057273) 1.467193 / 1.452155 (0.015038) 1.515633 / 1.492716 (0.022917)

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) 0.203804 / 0.018006 (0.185798) 0.493181 / 0.000490 (0.492691) 0.002171 / 0.000200 (0.001971) 0.000072 / 0.000054 (0.000017)

Benchmark: benchmark_indices_mapping.json

metric select shard shuffle sort train_test_split
new / old (diff) 0.021751 / 0.037411 (-0.015661) 0.092797 / 0.014526 (0.078271) 0.107650 / 0.176557 (-0.068907) 0.155552 / 0.737135 (-0.581584) 0.106798 / 0.296338 (-0.189541)

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) 0.413940 / 0.215209 (0.198730) 4.131958 / 2.077655 (2.054304) 1.854494 / 1.504120 (0.350374) 1.661644 / 1.541195 (0.120449) 1.704289 / 1.468490 (0.235799) 0.447288 / 4.584777 (-4.137489) 3.357257 / 3.745712 (-0.388455) 1.827672 / 5.269862 (-3.442190) 1.085227 / 4.565676 (-3.480449) 0.052913 / 0.424275 (-0.371362) 0.010796 / 0.007607 (0.003189) 0.520605 / 0.226044 (0.294561) 5.223097 / 2.268929 (2.954169) 2.288166 / 55.444624 (-53.156458) 1.939923 / 6.876477 (-4.936553) 2.064894 / 2.142072 (-0.077178) 0.560665 / 4.805227 (-4.244562) 0.118270 / 6.500664 (-6.382394) 0.064323 / 0.075469 (-0.011146)

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) 1.507612 / 1.841788 (-0.334176) 12.879615 / 8.074308 (4.805307) 26.370694 / 10.191392 (16.179302) 0.869058 / 0.680424 (0.188634) 0.594183 / 0.534201 (0.059982) 0.343480 / 0.579283 (-0.235803) 0.389761 / 0.434364 (-0.044603) 0.235614 / 0.540337 (-0.304724) 0.242229 / 1.386936 (-1.144707)
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) 0.005653 / 0.011353 (-0.005700) 0.003652 / 0.011008 (-0.007357) 0.026981 / 0.038508 (-0.011527) 0.027879 / 0.023109 (0.004769) 0.379276 / 0.275898 (0.103378) 0.438471 / 0.323480 (0.114991) 0.003597 / 0.007986 (-0.004389) 0.003102 / 0.004328 (-0.001227) 0.004647 / 0.004250 (0.000396) 0.037841 / 0.037052 (0.000789) 0.388531 / 0.258489 (0.130042) 0.428446 / 0.293841 (0.134605) 0.027115 / 0.128546 (-0.101431) 0.009499 / 0.075646 (-0.066147) 0.249363 / 0.419271 (-0.169909) 0.054827 / 0.043533 (0.011294) 0.385343 / 0.255139 (0.130204) 0.408659 / 0.283200 (0.125459) 0.097165 / 0.141683 (-0.044518) 1.493510 / 1.452155 (0.041356) 1.521340 / 1.492716 (0.028624)

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) 0.213770 / 0.018006 (0.195764) 0.401115 / 0.000490 (0.400625) 0.003718 / 0.000200 (0.003518) 0.000078 / 0.000054 (0.000024)

Benchmark: benchmark_indices_mapping.json

metric select shard shuffle sort train_test_split
new / old (diff) 0.022817 / 0.037411 (-0.014594) 0.095728 / 0.014526 (0.081202) 0.107219 / 0.176557 (-0.069338) 0.152265 / 0.737135 (-0.584871) 0.109304 / 0.296338 (-0.187034)

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) 0.429977 / 0.215209 (0.214768) 4.259470 / 2.077655 (2.181815) 2.021834 / 1.504120 (0.517714) 1.754753 / 1.541195 (0.213558) 1.804017 / 1.468490 (0.335526) 0.450778 / 4.584777 (-4.133999) 3.346255 / 3.745712 (-0.399458) 1.847810 / 5.269862 (-3.422051) 1.094879 / 4.565676 (-3.470798) 0.053510 / 0.424275 (-0.370765) 0.010905 / 0.007607 (0.003298) 0.524914 / 0.226044 (0.298869) 5.281856 / 2.268929 (3.012927) 2.352873 / 55.444624 (-53.091751) 1.990855 / 6.876477 (-4.885622) 2.110004 / 2.142072 (-0.032068) 0.559424 / 4.805227 (-4.245803) 0.119537 / 6.500664 (-6.381127) 0.065166 / 0.075469 (-0.010303)

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) 1.474225 / 1.841788 (-0.367563) 12.632846 / 8.074308 (4.558538) 26.038333 / 10.191392 (15.846941) 0.807930 / 0.680424 (0.127507) 0.541147 / 0.534201 (0.006946) 0.344311 / 0.579283 (-0.234972) 0.396425 / 0.434364 (-0.037939) 0.229250 / 0.540337 (-0.311087) 0.242172 / 1.386936 (-1.144764)

CML watermark

Please sign in to comment.