html_url
stringlengths 51
51
| comments
stringlengths 67
24.7k
| title
stringlengths 6
280
| body
stringlengths 51
36.2k
| comment_length
int64 16
1.45k
| text
stringlengths 190
38.3k
| embeddings
sequence |
|---|---|---|---|---|---|---|
https://github.com/huggingface/datasets/issues/6638 | Looks like it works with latest datasets repository
```
- `datasets` version: 2.16.2.dev0
- Platform: Linux-5.15.0-92-generic-x86_64-with-glibc2.29
- Python version: 3.8.10
- `huggingface_hub` version: 0.20.3
- PyArrow version: 15.0.0
- Pandas version: 2.0.1
- `fsspec` version: 2023.10.0
```
Could you explain which is the minimum version that fixes this?
Edit: Looks like that's 2.16.0, will close out issue | Cannot download wmt16 dataset | ### Describe the bug
As of this morning (PST) 2/1/2024, seeing the wmt16 dataset is missing from opus , could you suggest an alternative?
```
Downloading data files: 0%| | 0/4 [00:00<?, ?it/s]Traceback (most recent call last):
File "test.py", line 2, in <module>
raw_datasets = load_dataset("wmt16","ro-en",split="train")
File "/usr/local/lib/python3.8/dist-packages/datasets/load.py", line 2153, in load_dataset
builder_instance.download_and_prepare(
File "/usr/local/lib/python3.8/dist-packages/datasets/builder.py", line 954, in download_and_prepare
self._download_and_prepare(
File "/usr/local/lib/python3.8/dist-packages/datasets/builder.py", line 1717, in _download_and_prepare
super()._download_and_prepare(
File "/usr/local/lib/python3.8/dist-packages/datasets/builder.py", line 1027, in _download_and_prepare
split_generators = self._split_generators(dl_manager, **split_generators_kwargs)
File "/root/.cache/huggingface/modules/datasets_modules/datasets/wmt16/746749a11d25c02058042da7502d973ff410e73457f3d305fc1177dc0e8c4227/wmt_utils.py", line 754, in _split_generators
downloaded_files = dl_manager.download_and_extract(urls_to_download)
File "/usr/local/lib/python3.8/dist-packages/datasets/download/download_manager.py", line 565, in download_and_extract
return self.extract(self.download(url_or_urls))
File "/usr/local/lib/python3.8/dist-packages/datasets/download/download_manager.py", line 428, in download
downloaded_path_or_paths = map_nested(
File "/usr/local/lib/python3.8/dist-packages/datasets/utils/py_utils.py", line 464, in map_nested
mapped = [
File "/usr/local/lib/python3.8/dist-packages/datasets/utils/py_utils.py", line 465, in <listcomp>
_single_map_nested((function, obj, types, None, True, None))
File "/usr/local/lib/python3.8/dist-packages/datasets/utils/py_utils.py", line 384, in _single_map_nested
mapped = [_single_map_nested((function, v, types, None, True, None)) for v in pbar]
File "/usr/local/lib/python3.8/dist-packages/datasets/utils/py_utils.py", line 384, in <listcomp>
mapped = [_single_map_nested((function, v, types, None, True, None)) for v in pbar]
File "/usr/local/lib/python3.8/dist-packages/datasets/utils/py_utils.py", line 367, in _single_map_nested
return function(data_struct)
File "/usr/local/lib/python3.8/dist-packages/datasets/download/download_manager.py", line 454, in _download
return cached_path(url_or_filename, download_config=download_config)
File "/usr/local/lib/python3.8/dist-packages/datasets/utils/file_utils.py", line 182, in cached_path
output_path = get_from_cache(
File "/usr/local/lib/python3.8/dist-packages/datasets/utils/file_utils.py", line 596, in get_from_cache
raise FileNotFoundError(f"Couldn't find file at {url}")
FileNotFoundError: Couldn't find file at https://opus.nlpl.eu/download.php?f=SETIMES/v2/tmx/en-ro.tmx.gz
```
### Steps to reproduce the bug
```
from datasets import load_dataset
raw_datasets = load_dataset("wmt16","ro-en",split="train")
```
### Expected behavior
Expect the dataset to be downloaded/ at least a clean exit with error explaining dataset is missing and a suggestion for next steps
### Environment info
- `datasets` version: 2.14.7
- Platform: Linux-5.15.0-92-generic-x86_64-with-glibc2.29
- Python version: 3.8.10
- Huggingface_hub version: 0.17.3
- PyArrow version: 15.0.0
- Pandas version: 2.0.1
| 57 | Cannot download wmt16 dataset
### Describe the bug
As of this morning (PST) 2/1/2024, seeing the wmt16 dataset is missing from opus , could you suggest an alternative?
```
Downloading data files: 0%| | 0/4 [00:00<?, ?it/s]Traceback (most recent call last):
File "test.py", line 2, in <module>
raw_datasets = load_dataset("wmt16","ro-en",split="train")
File "/usr/local/lib/python3.8/dist-packages/datasets/load.py", line 2153, in load_dataset
builder_instance.download_and_prepare(
File "/usr/local/lib/python3.8/dist-packages/datasets/builder.py", line 954, in download_and_prepare
self._download_and_prepare(
File "/usr/local/lib/python3.8/dist-packages/datasets/builder.py", line 1717, in _download_and_prepare
super()._download_and_prepare(
File "/usr/local/lib/python3.8/dist-packages/datasets/builder.py", line 1027, in _download_and_prepare
split_generators = self._split_generators(dl_manager, **split_generators_kwargs)
File "/root/.cache/huggingface/modules/datasets_modules/datasets/wmt16/746749a11d25c02058042da7502d973ff410e73457f3d305fc1177dc0e8c4227/wmt_utils.py", line 754, in _split_generators
downloaded_files = dl_manager.download_and_extract(urls_to_download)
File "/usr/local/lib/python3.8/dist-packages/datasets/download/download_manager.py", line 565, in download_and_extract
return self.extract(self.download(url_or_urls))
File "/usr/local/lib/python3.8/dist-packages/datasets/download/download_manager.py", line 428, in download
downloaded_path_or_paths = map_nested(
File "/usr/local/lib/python3.8/dist-packages/datasets/utils/py_utils.py", line 464, in map_nested
mapped = [
File "/usr/local/lib/python3.8/dist-packages/datasets/utils/py_utils.py", line 465, in <listcomp>
_single_map_nested((function, obj, types, None, True, None))
File "/usr/local/lib/python3.8/dist-packages/datasets/utils/py_utils.py", line 384, in _single_map_nested
mapped = [_single_map_nested((function, v, types, None, True, None)) for v in pbar]
File "/usr/local/lib/python3.8/dist-packages/datasets/utils/py_utils.py", line 384, in <listcomp>
mapped = [_single_map_nested((function, v, types, None, True, None)) for v in pbar]
File "/usr/local/lib/python3.8/dist-packages/datasets/utils/py_utils.py", line 367, in _single_map_nested
return function(data_struct)
File "/usr/local/lib/python3.8/dist-packages/datasets/download/download_manager.py", line 454, in _download
return cached_path(url_or_filename, download_config=download_config)
File "/usr/local/lib/python3.8/dist-packages/datasets/utils/file_utils.py", line 182, in cached_path
output_path = get_from_cache(
File "/usr/local/lib/python3.8/dist-packages/datasets/utils/file_utils.py", line 596, in get_from_cache
raise FileNotFoundError(f"Couldn't find file at {url}")
FileNotFoundError: Couldn't find file at https://opus.nlpl.eu/download.php?f=SETIMES/v2/tmx/en-ro.tmx.gz
```
### Steps to reproduce the bug
```
from datasets import load_dataset
raw_datasets = load_dataset("wmt16","ro-en",split="train")
```
### Expected behavior
Expect the dataset to be downloaded/ at least a clean exit with error explaining dataset is missing and a suggestion for next steps
### Environment info
- `datasets` version: 2.14.7
- Platform: Linux-5.15.0-92-generic-x86_64-with-glibc2.29
- Python version: 3.8.10
- Huggingface_hub version: 0.17.3
- PyArrow version: 15.0.0
- Pandas version: 2.0.1
Looks like it works with latest datasets repository
```
- `datasets` version: 2.16.2.dev0
- Platform: Linux-5.15.0-92-generic-x86_64-with-glibc2.29
- Python version: 3.8.10
- `huggingface_hub` version: 0.20.3
- PyArrow version: 15.0.0
- Pandas version: 2.0.1
- `fsspec` version: 2023.10.0
```
Could you explain which is the minimum version that fixes this?
Edit: Looks like that's 2.16.0, will close out issue | [
-0.36052435636520386,
0.317926287651062,
-0.002050556242465973,
0.38708245754241943,
0.4038749039173126,
0.11385878920555115,
0.2532888948917389,
0.20784994959831238,
-0.027222130447626114,
0.21313440799713135,
0.026451997458934784,
0.03987692669034004,
-0.3029175102710724,
0.040763065218925476,
0.31888723373413086,
-0.1025252416729927,
-0.04760010167956352,
0.14544767141342163,
-0.681085467338562,
-0.021854877471923828,
-0.39371827244758606,
0.04465974122285843,
-0.19849680364131927,
-0.12580342590808868,
0.09383539855480194,
0.10055328905582428,
-0.16295811533927917,
0.15399499237537384,
-0.38727134466171265,
-0.36015573143959045,
0.27366510033607483,
-0.02034168690443039,
0.015924938023090363,
0.3857197165489197,
-0.0001130660020862706,
0.08519972860813141,
0.14452162384986877,
-0.10201624780893326,
-0.19559814035892487,
-0.3037537932395935,
-0.3311106562614441,
-0.1036369726061821,
-0.15824337303638458,
0.06927962601184845,
-0.24550971388816833,
-0.055901795625686646,
-0.18517820537090302,
-0.5401483774185181,
-0.016727354377508163,
0.42170268297195435,
0.23550038039684296,
0.484872430562973,
0.35655486583709717,
-0.2649945914745331,
0.2638974189758301,
0.08156862109899521,
-0.12339821457862854,
-0.1918424814939499,
0.27871453762054443,
-0.10150211304426193,
0.2279515117406845,
0.18325337767601013,
0.017423465847969055,
0.1436951756477356,
0.20806071162223816,
0.04344392195343971,
0.2050970494747162,
-0.28204482793807983,
0.2262723296880722,
0.29122257232666016,
0.4240016043186188,
0.09453783929347992,
-0.26550546288490295,
0.10446406900882721,
0.04568328335881233,
-0.8138516545295715,
0.16890345513820648,
0.3360201120376587,
0.045869600027799606,
0.160896435379982,
-0.09935765713453293,
0.12913590669631958,
-0.010949559509754181,
0.19396920502185822,
-0.08163818717002869,
0.29255780577659607,
0.04062768816947937,
0.05127301067113876,
0.14542195200920105,
-0.09675976634025574,
0.04531024023890495,
-0.08745452016592026,
0.035761479288339615,
0.3617963194847107,
-0.2636680006980896,
0.17469875514507294,
0.04791216552257538,
0.17806598544120789,
0.13792447745800018,
0.12443238496780396,
-0.2301606684923172,
-0.14308272302150726,
0.15823215246200562,
-0.24341677129268646,
0.37003225088119507,
-0.1312004178762436,
-0.400727242231369,
-0.20608839392662048,
0.2326180785894394,
0.3219103515148163,
0.23010516166687012,
-0.06361816823482513,
0.1047496497631073,
-0.2295609563589096,
0.04944990575313568,
-0.022846248000860214,
0.28195181488990784,
-0.05631207302212715,
-0.28039097785949707,
0.025094978511333466,
0.1288427859544754,
-0.14529109001159668,
0.018505088984966278,
0.18536239862442017,
-0.09285268187522888,
0.33846819400787354,
0.21600337326526642,
0.07964572310447693,
-0.20634523034095764,
-0.21399661898612976,
-0.227199524641037,
-0.13588941097259521,
-0.046290650963783264,
0.0036794431507587433,
0.3561675548553467,
-0.07296943664550781,
0.22682107985019684,
0.12110622227191925,
-0.14704562723636627,
-0.12097116559743881,
-0.09337835758924484,
-0.11293669044971466,
0.09954296052455902,
0.4110374450683594,
0.08147472888231277,
0.15208283066749573,
-0.050447236746549606,
0.14584259688854218,
-0.032392971217632294,
0.18502628803253174,
0.07762625068426132,
-0.3643757402896881,
-0.16982623934745789,
0.2770959436893463,
-0.14592251181602478,
0.02576773799955845,
0.11087282001972198,
-0.17097881436347961,
-0.2616032361984253,
0.12307314574718475,
-0.12111534923315048,
-0.14929582178592682,
-0.16370312869548798,
-0.10787604749202728,
0.45133689045906067,
0.4419046640396118,
0.0882568433880806,
-0.03239230066537857,
-0.6517634391784668,
-0.1498183012008667,
0.20000171661376953,
0.1622234284877777,
-0.2046852707862854,
0.02917708456516266,
-0.27312731742858887,
-0.12635564804077148,
0.37148454785346985,
-0.3860093951225281,
-0.5101290941238403,
-0.12777788937091827,
-0.21607929468154907,
0.0599537268280983,
-0.038483619689941406,
0.15600261092185974,
0.42113348841667175,
-0.05334126949310303,
0.33788198232650757,
0.1812024712562561,
-0.23486663401126862,
0.008521392941474915,
-0.14132314920425415,
-0.40417033433914185,
0.44959738850593567,
0.17204917967319489,
0.3796803951263428,
-0.09577739238739014,
0.014443082734942436,
0.05205095186829567,
0.4336634576320648,
0.23383212089538574,
0.1211891770362854,
-0.015320208854973316,
0.4003097414970398,
-0.06476525962352753,
-0.0482119657099247,
-0.06610175967216492,
-0.2763664722442627,
0.25427353382110596,
0.2915329933166504,
0.10316836833953857,
0.08507335186004639,
0.04363948851823807,
-0.17034514248371124,
0.047847338020801544,
-0.2892531752586365,
-0.3850303888320923,
0.11659523844718933,
-0.037858206778764725,
0.15110984444618225,
-0.06575346738100052,
-0.10380002856254578,
-0.3095017373561859,
-0.003438226878643036,
0.14934326708316803,
0.14895384013652802,
0.0438448041677475,
-0.11421184986829758,
0.012902714312076569,
0.212632954120636,
0.2532339096069336,
0.19013485312461853,
0.006144281476736069,
-0.09847218543291092,
0.18976247310638428,
0.12817595899105072,
0.30328336358070374,
-0.14390650391578674,
0.4107021987438202,
0.37855470180511475,
-0.5947980284690857,
0.09422146528959274,
0.11947578191757202,
0.15728560090065002,
-0.13131307065486908,
0.06479185074567795,
0.028962645679712296,
-0.18153975903987885,
0.0471712164580822,
0.15391767024993896,
0.2556743025779724,
0.32432880997657776,
-0.17625758051872253,
0.06807416677474976,
-0.01112145185470581,
0.32015782594680786,
0.001012967899441719,
0.25607725977897644,
0.04175519198179245,
-0.1714586317539215,
0.11697812378406525,
0.45377275347709656,
0.1449935883283615,
0.048586953431367874,
-0.13649502396583557,
-0.41772934794425964,
0.017252303659915924,
-0.17096488177776337,
0.09157201647758484,
0.6213274002075195,
0.17812103033065796,
0.08388590812683105,
0.028731243684887886,
0.22802215814590454,
-0.16103661060333252,
0.49282488226890564,
0.19856144487857819,
0.21067366003990173,
0.16732674837112427,
-0.0006827898323535919,
0.09504730999469757,
-0.13925334811210632,
0.33682429790496826,
-0.0108681321144104,
0.12647850811481476,
-0.07401365786790848,
-0.28268468379974365,
-0.3848985731601715,
0.09621943533420563,
-0.41386744379997253,
-0.5886368751525879,
-0.24780847132205963,
-0.20807000994682312,
-0.05177464336156845,
0.1690845489501953,
-0.1629844754934311,
0.09249505400657654,
-0.3313494324684143,
0.0354093462228775,
-0.07200625538825989,
0.06952768564224243,
0.04312581568956375,
0.0037755072116851807,
-0.13464048504829407,
0.06626476347446442,
0.23552942276000977,
-0.39568692445755005,
0.4484192728996277,
-0.1507643759250641,
0.051978811621665955,
-0.43171533942222595,
-0.20725281536579132,
0.21228492259979248,
0.2740793824195862,
0.30607539415359497,
0.24500878155231476,
0.14426690340042114,
0.48971378803253174,
-0.12678158283233643,
0.1766439974308014,
-0.3439158797264099,
0.04378797113895416,
-0.1002589762210846,
-0.19042833149433136,
0.042275022715330124,
-0.16191957890987396,
-0.5782846808433533,
-0.5531324148178101,
-0.30773472785949707,
-0.038355935364961624,
-0.02829645574092865,
-0.009400328621268272,
0.0894286036491394,
-0.148014634847641,
-0.05477452278137207,
0.3384121358394623,
-0.07681933045387268,
-0.36061644554138184,
-0.24838939309120178,
0.28311216831207275,
-0.2172355353832245,
-0.5080080032348633,
0.26720815896987915,
-0.08912443369626999,
0.3436751961708069,
0.13344410061836243,
-0.5720171332359314,
0.43301481008529663,
0.011155424639582634,
0.10506868362426758,
0.03589901700615883,
-0.07896053791046143,
0.33992642164230347,
-0.049576058983802795,
-0.03564808517694473,
-0.28697866201400757,
-0.21784263849258423,
-0.14639563858509064,
-0.007904430851340294,
0.49254581332206726,
-0.08607163280248642,
0.2768721282482147,
0.11648814380168915,
0.7780134081840515,
0.03807223588228226,
-0.1931980699300766,
0.1653410643339157,
-0.24980218708515167,
0.2553129196166992,
0.1219145655632019,
-0.36560913920402527,
0.16422569751739502,
0.029239267110824585,
0.1096314936876297,
0.29481279850006104,
-0.05719093978404999,
-0.08297140896320343,
-0.15030112862586975,
0.19410251080989838,
-0.3572814464569092,
-0.4683474898338318,
0.036911241710186005,
0.15571057796478271,
0.3734396994113922,
0.3533400893211365,
0.331561416387558,
0.06629091501235962,
-0.2881094217300415,
0.08159129321575165,
0.3436906039714813,
0.1712893694639206,
0.0770689994096756,
0.21058228611946106,
-0.030316421762108803,
-0.4444749057292938,
-0.04847370833158493,
-0.2085513323545456,
0.6952723860740662,
-0.06273025274276733,
-0.12352544814348221,
-0.1420067995786667,
0.005674149841070175,
0.537901759147644,
0.0867379829287529,
0.3406134247779846,
0.08308537304401398,
-0.08257171511650085,
-0.2757599651813507,
-0.2109440267086029,
-0.04342319071292877,
0.09999813139438629,
0.20812253654003143,
0.452755331993103,
-0.06034759432077408,
-0.23581549525260925,
0.47391316294670105,
0.16521205008029938,
-0.20506073534488678,
-0.020012158900499344,
-0.022933103144168854,
-0.33845970034599304,
-0.1528332531452179,
-0.0044874101877212524,
-0.1307780146598816,
0.27874764800071716,
-0.36341750621795654,
-0.03844665363430977,
0.08077071607112885,
-0.02705782651901245,
0.030399663373827934,
-0.144413024187088,
0.26408883929252625,
0.0890652984380722,
0.5655502676963806,
0.5011258125305176,
0.01182773057371378,
0.508297324180603,
0.43557918071746826,
-0.040138840675354004,
-0.23070625960826874,
0.06227971985936165,
-0.22289803624153137,
0.07919081300497055,
0.16722798347473145,
-0.03727051615715027,
0.07904645800590515,
0.3591538071632385,
-0.0577760748565197,
-0.6436673998832703,
-0.12749384343624115,
0.28138116002082825,
-0.08190998435020447,
-0.1553291380405426,
-0.1166425570845604,
0.0808902382850647,
0.0199669748544693,
0.0410199910402298,
0.3474075198173523,
-0.11876808106899261,
-0.17826226353645325,
-0.21849828958511353,
0.10563429445028305,
0.5729706287384033,
-0.026557892560958862,
0.018177561461925507,
-0.052091460675001144,
-0.31103116273880005,
0.21284133195877075,
0.1945478320121765,
0.09727941453456879,
-0.23407474160194397,
-0.1695728749036789,
-0.09488128125667572,
-0.18864880502223969,
0.025974251329898834,
0.0414356105029583,
-0.1443951576948166,
0.3288700580596924,
-0.05526939779520035,
0.5448484420776367,
0.09806280583143234,
-0.08349676430225372,
-0.14776156842708588,
-0.26685646176338196,
-0.28395140171051025,
0.13666626811027527,
-0.07351422309875488,
0.49000805616378784,
-0.266923725605011,
-0.11597450822591782,
0.10133473575115204,
-0.241456538438797,
-0.3830203115940094,
0.05076775699853897,
-0.7256552577018738,
0.07651980221271515,
-0.052903153002262115,
-0.038887158036231995,
0.09882457554340363,
0.03645174577832222,
-0.17865419387817383,
0.26721543073654175,
-0.30016040802001953,
0.27889949083328247,
-0.2279781997203827,
0.06538908183574677,
0.032425407320261,
0.3202238976955414,
0.10617600381374359,
-0.08352074772119522,
-0.1955881416797638,
-0.09261757135391235,
-0.12532487511634827,
-0.20072267949581146,
0.21147286891937256,
-0.11623896658420563,
0.31142741441726685,
0.20567254722118378,
-0.04505879804491997,
0.11859385669231415,
0.017674557864665985,
-0.294194757938385,
0.16874101758003235,
0.12739770114421844,
-0.09488394856452942,
0.3187659680843353,
0.07281996309757233,
-0.046811364591121674,
-0.06456238776445389,
0.39273884892463684,
-0.1651582568883896,
-0.036189004778862,
0.320424884557724,
0.34339088201522827,
-0.059234313666820526,
-0.3559420108795166,
0.025882430374622345,
-0.25705423951148987,
-0.09844903647899628,
0.08616027235984802,
-0.01751309260725975,
0.3950469493865967,
0.0682646781206131,
-0.136641263961792,
0.22155843675136566,
-0.38310733437538147,
-0.14845623075962067,
-0.5851325988769531,
-0.11352889239788055,
0.014731116592884064,
-0.036689531058073044,
0.03943762555718422,
-0.1223355308175087,
0.28017276525497437,
-0.031005047261714935,
-0.07552001625299454,
-0.31703001260757446,
0.2484564185142517,
-0.037944160401821136,
-0.021646149456501007,
0.3473944067955017,
-0.038095228374004364,
0.1913871020078659,
-0.12828560173511505,
0.08000761270523071,
-0.07252217829227448,
-0.30955779552459717,
-0.15939441323280334,
-0.28724026679992676,
0.14272207021713257,
-0.11383754760026932,
-0.3669804632663727,
0.18980658054351807,
-0.2682432532310486,
-0.17426136136054993,
0.13683532178401947,
0.06277816742658615,
0.0325561985373497,
-0.14726628363132477,
0.0414363257586956,
0.1274479329586029,
0.027276838198304176,
-0.2698461413383484,
0.18622362613677979,
-0.05823451280593872,
0.3914121687412262,
0.08282317221164703,
0.17690612375736237,
0.2801856994628906,
-0.18184828758239746,
-0.02527957409620285,
-0.21177387237548828,
-0.12113615870475769,
-0.01884823478758335,
0.4527779221534729,
-0.06837256252765656,
-0.006337597966194153,
0.46826857328414917,
0.13654622435569763,
0.18379011750221252,
-0.057814404368400574,
0.25063586235046387,
-0.10570621490478516,
0.2739763855934143,
-0.3630700409412384,
-0.14659619331359863,
-0.18300189077854156,
0.27004873752593994,
0.009349629282951355,
-0.05254227668046951,
0.22998085618019104,
-0.17594175040721893,
-0.17185935378074646,
0.24356907606124878,
0.41248467564582825,
-0.24650278687477112,
0.30622345209121704,
0.4636017382144928,
-0.17864735424518585,
0.17346854507923126,
-0.12051404267549515,
-0.2235334813594818,
-0.0150814950466156,
0.5325531959533691,
-0.28548258543014526,
0.30240386724472046,
0.2380632758140564,
-0.0947258248925209,
0.11155258119106293,
-0.2780929505825043,
-0.14543834328651428,
0.16129262745380402,
-0.14945101737976074,
0.2031397819519043,
0.08046239614486694,
0.29135578870773315,
-0.23676443099975586,
0.11971129477024078,
-0.09510527551174164,
0.2228306084871292,
-0.0438835546374321,
0.1192011907696724,
-0.1368272602558136,
-0.33010250329971313,
-0.17682723701000214,
0.16876119375228882,
0.10811750590801239,
-0.005369942635297775,
0.24525777995586395,
0.30208808183670044,
-0.20671223104000092,
-0.40553492307662964,
-0.3392612636089325,
0.2949604094028473,
-0.04116279259324074,
-0.31704720854759216,
-0.028049584478139877,
0.4528757631778717,
-0.05290853604674339,
0.31937742233276367,
0.4425782561302185,
0.21958811581134796,
0.1460931897163391,
-0.1434238851070404,
0.011185556650161743,
0.15677529573440552,
-0.11890611052513123,
-0.304232656955719,
0.1213318407535553,
0.0572688989341259,
-0.1279427409172058,
0.07817009091377258,
0.15360961854457855,
-0.12414270639419556,
-0.16661985218524933,
0.08166816085577011,
-0.05184332653880119,
-0.2465672492980957,
0.19029158353805542,
0.012636596336960793,
-0.07693181931972504,
-0.256583034992218,
0.2185429036617279,
0.09194208681583405,
-0.09174740314483643,
0.5764175057411194,
-0.004267897456884384,
0.25605499744415283,
-0.14251317083835602,
0.068544402718544,
0.08807793259620667,
0.35946157574653625,
0.1567237377166748,
0.002019362524151802,
-0.2514655292034149,
-0.37280961871147156,
-0.7758952975273132,
-0.09439551830291748,
0.11068403720855713,
0.2136768102645874,
0.1856192797422409,
0.08264632523059845,
0.08397354930639267,
0.12231191992759705,
0.42334872484207153,
-0.06338699162006378,
0.3240467309951782,
-0.06094535067677498,
-0.40068361163139343,
-0.2425767332315445,
0.06915760040283203,
0.07099275290966034,
0.053303658962249756,
-0.3861621022224426,
0.18144232034683228,
-0.19469872117042542,
0.053651534020900726,
-0.42677363753318787,
0.10464473068714142,
-0.11190256476402283,
0.19696661829948425,
0.3448432385921478,
0.22462688386440277,
0.28886160254478455,
-0.10796292126178741,
-0.13690753281116486,
-0.24022871255874634,
-0.07545136660337448,
-0.13072092831134796,
-0.02078089863061905,
-0.071778804063797,
0.017377108335494995,
0.06367309391498566,
-0.5198937654495239,
-0.3963315188884735,
0.0451851524412632,
-0.0016635879874229431,
0.19802206754684448,
-0.5658490061759949,
0.3297119438648224,
-0.23240765929222107,
0.07845732569694519,
0.057564347982406616,
0.14609257876873016,
-0.13916876912117004,
-0.10186471045017242,
-0.14781564474105835,
-0.26046714186668396,
0.3043631911277771,
-0.2848113477230072,
-0.06687533855438232,
-0.5626125931739807,
0.14843294024467468,
-0.19020769000053406,
0.01544412411749363,
-0.335675448179245,
0.08486308157444,
0.2441290020942688,
-0.09692934155464172,
-0.22004255652427673,
0.21125668287277222,
-0.1466057300567627,
-0.01342058926820755,
0.15410436689853668,
0.46332186460494995,
0.10206621885299683,
-0.24516652524471283,
-0.12079721689224243,
-0.19758552312850952
] |
https://github.com/huggingface/datasets/issues/6624 | Hi, this dataset has been disabled by the authors, so unfortunately it's no longer possible to download it. | How to download the laion-coco dataset | The laion coco dataset is not available now. How to download it
https://huggingface.co/datasets/laion/laion-coco | 18 | How to download the laion-coco dataset
The laion coco dataset is not available now. How to download it
https://huggingface.co/datasets/laion/laion-coco
Hi, this dataset has been disabled by the authors, so unfortunately it's no longer possible to download it. | [
-0.15687450766563416,
-0.23160290718078613,
-0.20348447561264038,
0.2225615680217743,
-0.014091560617089272,
-0.005569621920585632,
0.1532430648803711,
0.07347936183214188,
-0.3041417598724365,
0.11177970468997955,
-0.3511919379234314,
0.1720883846282959,
0.03486033156514168,
0.37087300419807434,
0.06044836714863777,
-0.1595807671546936,
-0.13234062492847443,
-0.05253537371754646,
0.14318905770778656,
-0.17197003960609436,
-0.026270393282175064,
0.25338324904441833,
-0.058486972004175186,
-0.05342555791139603,
0.18231168389320374,
-0.1428816169500351,
-0.33576905727386475,
0.02891378663480282,
-0.3386974036693573,
-0.31054407358169556,
0.5495043396949768,
0.2182805836200714,
0.32229405641555786,
-0.03667532652616501,
-0.0000932125622057356,
-0.026982568204402924,
0.31394314765930176,
-0.09625187516212463,
0.1038481742143631,
-0.22185254096984863,
-0.4870290756225586,
-0.21821656823158264,
-0.09716705977916718,
-0.16300322115421295,
-0.003260403871536255,
0.18102732300758362,
0.0538439005613327,
-0.14801561832427979,
0.33395665884017944,
0.13310323655605316,
0.38284575939178467,
0.2748640775680542,
0.17201487720012665,
-0.22167356312274933,
0.17835509777069092,
0.012524273246526718,
-0.23609769344329834,
0.13506227731704712,
0.2767125964164734,
0.013647176325321198,
0.06437777727842331,
0.17624381184577942,
0.15020586550235748,
-0.2877812087535858,
-0.12902742624282837,
-0.027867933735251427,
0.0617622546851635,
-0.4359237551689148,
0.15072041749954224,
0.5334852337837219,
0.5680654644966125,
0.20327922701835632,
-0.16639328002929688,
0.10983921587467194,
0.030090872198343277,
-0.3939836919307709,
-0.05485551431775093,
0.49280887842178345,
0.027657251805067062,
0.2033359706401825,
-0.15457278490066528,
-0.1631777286529541,
-0.20080798864364624,
0.09048429131507874,
-0.07753688097000122,
0.11750780791044235,
-0.20870088040828705,
-0.09450899064540863,
0.019739573821425438,
-0.05554508417844772,
0.15310901403427124,
0.22084572911262512,
-0.1776551902294159,
-0.04765487462282181,
-0.09828510880470276,
-0.09394078701734543,
-0.0441359207034111,
0.42182889580726624,
0.2996205687522888,
0.09979921579360962,
0.34184980392456055,
-0.11558917909860611,
-0.2664100229740143,
-0.026869554072618484,
0.14688405394554138,
0.11175146698951721,
0.16333261132240295,
-0.047029510140419006,
0.19597628712654114,
0.36639678478240967,
0.15900525450706482,
-0.1050625741481781,
-0.04760279506444931,
0.2706533968448639,
-0.3811447024345398,
-0.0208311527967453,
-0.05041810870170593,
-0.42154616117477417,
-0.33822908997535706,
-0.11242543160915375,
0.08581556379795074,
0.033439382910728455,
0.052665114402770996,
0.2720377445220947,
-0.102695032954216,
0.07731027901172638,
-0.07735377550125122,
0.13746985793113708,
0.1855856478214264,
-0.18847957253456116,
-0.30035972595214844,
0.04696236178278923,
-0.08117634057998657,
-0.046592842787504196,
0.2791740298271179,
-0.3209453821182251,
0.41417214274406433,
-0.07596952468156815,
0.12121206521987915,
0.2687426507472992,
-0.15521389245986938,
-0.027999792248010635,
-0.1346912682056427,
0.17363595962524414,
0.0006307587027549744,
-0.13038313388824463,
0.13546329736709595,
-0.11338049918413162,
-0.1310008019208908,
-0.07227777689695358,
-0.23871256411075592,
-0.1970406472682953,
-0.11818603426218033,
0.2586930990219116,
-0.26164722442626953,
-0.24542787671089172,
-0.04862331226468086,
0.23783312737941742,
-0.20519354939460754,
0.161880761384964,
-0.012451417744159698,
-0.11796589195728302,
-0.08728339523077011,
-0.22149226069450378,
0.1685953587293625,
0.1519203782081604,
-0.1654866486787796,
-0.05847717449069023,
-0.30033159255981445,
-0.16768783330917358,
0.20192168653011322,
0.10067648440599442,
-0.15839727222919464,
-0.17008161544799805,
-0.33540692925453186,
-0.06625397503376007,
0.07545598596334457,
-0.31694212555885315,
-0.4344373047351837,
-0.06534099578857422,
0.08215659856796265,
-0.1012595146894455,
0.007294096052646637,
0.09524331986904144,
-0.027224861085414886,
-0.12457671761512756,
-0.002909887582063675,
0.16868017613887787,
0.0771801546216011,
-0.26437056064605713,
-0.016614675521850586,
-0.29327312111854553,
0.4095262885093689,
0.12250398099422455,
0.04077473282814026,
0.1712186187505722,
0.12401191890239716,
0.2817156910896301,
0.40853351354599,
0.12457583099603653,
-0.013186225667595863,
0.22511550784111023,
0.33419209718704224,
0.08154977858066559,
0.0510907843708992,
0.13709837198257446,
-0.198804572224617,
0.13347488641738892,
0.19419734179973602,
0.19385138154029846,
-0.07386013865470886,
-0.23064802587032318,
-0.18678757548332214,
0.05656750500202179,
0.10246934741735458,
-0.13950088620185852,
0.3226067125797272,
-0.09245029836893082,
0.3811361789703369,
0.0828937441110611,
-0.04852103814482689,
0.4812946319580078,
-0.0755355954170227,
0.027042385190725327,
-0.2005452662706375,
0.24368645250797272,
-0.0495762899518013,
0.20453792810440063,
0.49730098247528076,
0.08329710364341736,
0.07769346237182617,
0.0007731840014457703,
0.08660620450973511,
0.10627201199531555,
-0.09203316271305084,
0.11527568101882935,
0.118752121925354,
-0.06925816833972931,
0.26847973465919495,
-0.46755069494247437,
0.03528682142496109,
0.016221582889556885,
-0.11035238206386566,
0.26780545711517334,
-0.2557731866836548,
0.2051742523908615,
-0.013685938902199268,
-0.12920643389225006,
0.24315808713436127,
0.21094341576099396,
0.34760627150535583,
-0.11730483919382095,
-0.14922720193862915,
-0.1913599967956543,
0.37951967120170593,
0.032955605536699295,
0.32334017753601074,
-0.17438097298145294,
-0.4678892493247986,
0.3608666658401489,
0.12651106715202332,
-0.13533291220664978,
-0.03173878788948059,
0.03948666527867317,
-0.16290885210037231,
0.018069609999656677,
0.3102705776691437,
0.10872717201709747,
0.01732771098613739,
0.3184717893600464,
0.18856148421764374,
-0.26252415776252747,
-0.2050977498292923,
-0.38675376772880554,
0.15637141466140747,
-0.0065836794674396515,
0.1176200807094574,
-0.20473138988018036,
0.4125067889690399,
-0.05817439779639244,
-0.660692036151886,
-0.16870996356010437,
-0.20111185312271118,
0.13063502311706543,
0.17251187562942505,
-0.16396889090538025,
-0.4131260812282562,
-0.11829802393913269,
0.28062665462493896,
-0.15652990341186523,
-0.24326682090759277,
-0.4202251732349396,
0.16668950021266937,
-0.05708573758602142,
0.2664520740509033,
-0.17494729161262512,
-0.6327065825462341,
0.3446277379989624,
-0.11445263028144836,
0.24555732309818268,
-0.33224934339523315,
-0.09942455589771271,
-0.29042285680770874,
0.3674173951148987,
-0.10424847900867462,
-0.1793196052312851,
0.2122320681810379,
-0.1342441737651825,
0.10599824786186218,
-0.14777661859989166,
-0.47224798798561096,
0.16698342561721802,
-0.1992427557706833,
0.04450278729200363,
0.07842689752578735,
0.20357993245124817,
-0.10594567656517029,
-0.08475633710622787,
0.10747427493333817,
-0.33415818214416504,
-0.22068357467651367,
-0.08381931483745575,
-0.024899665266275406,
-0.11644086241722107,
-0.3023720979690552,
-0.25689250230789185,
-0.3988383412361145,
-0.12664979696273804,
0.3955882489681244,
0.16853083670139313,
-0.07986585795879364,
0.1720578819513321,
-0.2342132329940796,
0.018316902220249176,
0.02696358412504196,
-0.08336731791496277,
-0.35628727078437805,
-0.48496213555336,
0.3665924668312073,
-0.37186872959136963,
-0.4627988040447235,
0.05524639040231705,
0.0895812064409256,
0.31556302309036255,
-0.08437713980674744,
-0.5092445015907288,
-0.00605083629488945,
-0.16782045364379883,
0.30681830644607544,
0.10823015868663788,
-0.20279589295387268,
0.12048979103565216,
0.059480637311935425,
-0.15830852091312408,
-0.21920642256736755,
0.13290739059448242,
-0.0010388046503067017,
0.10623624920845032,
0.353161484003067,
-0.18118548393249512,
0.05776483565568924,
0.3392638564109802,
0.15703298151493073,
0.07965037226676941,
-0.026759073138237,
0.4130978286266327,
-0.1214686706662178,
0.4155830144882202,
0.13634178042411804,
-0.14848867058753967,
0.16374099254608154,
-0.0701206773519516,
0.1244482696056366,
0.07054039090871811,
0.20618563890457153,
-0.29425546526908875,
-0.29688704013824463,
0.03805997967720032,
-0.15496709942817688,
-0.1378684639930725,
-0.3878674805164337,
0.6003235578536987,
0.2615259885787964,
0.1908310353755951,
-0.05371428281068802,
-0.054753996431827545,
-0.39702853560447693,
0.23123547434806824,
0.2556089162826538,
0.13550430536270142,
-0.012854712083935738,
-0.45491093397140503,
-0.17726017534732819,
-0.8023486733436584,
0.2160910964012146,
-0.13573619723320007,
0.03605491295456886,
-0.16181950271129608,
-0.007432952523231506,
-0.0969468504190445,
-0.0651036947965622,
0.4015886187553406,
-0.11845771223306656,
0.2968618869781494,
0.08548636734485626,
-0.06145349517464638,
-0.28907692432403564,
0.029335133731365204,
-0.33569827675819397,
0.03953779488801956,
0.3135862350463867,
0.4411716163158417,
-0.10331442952156067,
-0.004532720893621445,
0.062222376465797424,
0.07363580167293549,
-0.17703936994075775,
0.009894847869873047,
0.04949253797531128,
-0.28877151012420654,
-0.024466857314109802,
-0.10761827975511551,
-0.1120278462767601,
0.2908495366573334,
-0.23982754349708557,
-0.09669677913188934,
0.05825427174568176,
-0.000735938549041748,
0.08998933434486389,
0.19552768766880035,
-0.027128631249070168,
0.39240774512290955,
0.13417145609855652,
0.09284773468971252,
0.09858967363834381,
0.08915683627128601,
0.3634926974773407,
0.13337333500385284,
-0.2325505167245865,
-0.052244871854782104,
-0.10082642734050751,
0.13034850358963013,
0.18913668394088745,
-0.11983288824558258,
-0.12400969862937927,
-0.13492949306964874,
0.02369607239961624,
0.04725450277328491,
0.045849669724702835,
0.08839872479438782,
-0.19373390078544617,
-0.28889551758766174,
-0.10596586763858795,
0.2967252731323242,
0.06844553351402283,
0.07414200901985168,
0.0017079641111195087,
-0.11294624209403992,
-0.14659914374351501,
-0.053627125918865204,
0.09904923290014267,
0.7766706943511963,
-0.06686695665121078,
0.05097713693976402,
0.2302929013967514,
0.15251263976097107,
0.5579816102981567,
-0.12885326147079468,
0.21712400019168854,
0.0077590495347976685,
-0.43838778138160706,
-0.07383349537849426,
0.09852467477321625,
-0.02565247192978859,
0.03018607571721077,
-0.05882611498236656,
0.2380700409412384,
-0.05764703452587128,
0.07424985617399216,
-0.0324258878827095,
0.09915342926979065,
-0.07244497537612915,
-0.08503660559654236,
-0.21115712821483612,
0.3691771626472473,
0.14398586750030518,
0.29312533140182495,
-0.03795558959245682,
-0.11035922169685364,
-0.22234350442886353,
-0.05418713018298149,
-0.23816990852355957,
0.2839227318763733,
-0.08745826780796051,
0.24990171194076538,
-0.18558645248413086,
-0.09796259552240372,
-0.04919955134391785,
0.3401137888431549,
0.40545520186424255,
0.2386852353811264,
-0.35951393842697144,
0.2467879354953766,
-0.12016839534044266,
-0.4936966001987457,
0.0671314001083374,
0.23224148154258728,
-0.16891829669475555,
-0.26537442207336426,
-0.0949503481388092,
0.05447643995285034,
0.15915247797966003,
0.07088667899370193,
-0.13657666742801666,
-0.1471758484840393,
-0.07345905900001526,
0.053723499178886414,
0.20119109749794006,
0.221663236618042,
0.10369361937046051,
-0.15562155842781067,
0.27164655923843384,
0.20877228677272797,
0.059481095522642136,
-0.2226160764694214,
0.18116351962089539,
-0.10397904366254807,
-0.22318372130393982,
0.27438509464263916,
0.10678578168153763,
0.1514163613319397,
0.14980098605155945,
0.08537181466817856,
0.035240113735198975,
-0.264177143573761,
0.006038181483745575,
-0.15090049803256989,
-0.2150089293718338,
-0.017117859795689583,
-0.13396137952804565,
0.2159978300333023,
-0.12066733837127686,
0.4302305281162262,
0.07882826030254364,
-0.11126096546649933,
0.15460923314094543,
-0.4053763747215271,
-0.2391005903482437,
0.2700425982475281,
-0.20511111617088318,
0.10150030255317688,
-0.09515120834112167,
0.0672832578420639,
0.20334509015083313,
0.03891104459762573,
-0.47123417258262634,
-0.01694726198911667,
-0.17287681996822357,
-0.19376567006111145,
0.2992827594280243,
0.022345714271068573,
0.2877039313316345,
-0.09475082159042358,
0.26288309693336487,
0.052686069160699844,
-0.1539992392063141,
-0.3256533741950989,
-0.08072581887245178,
0.012752771377563477,
0.1733858287334442,
0.056254900991916656,
0.0035361051559448242,
-0.13174182176589966,
-0.20954112708568573,
0.11410121619701385,
-0.10149963200092316,
0.04564213752746582,
0.20708367228507996,
0.05743368715047836,
-0.05501923710107803,
0.06847625970840454,
0.028635337948799133,
0.19533903896808624,
0.33884313702583313,
0.2168169617652893,
0.16797301173210144,
0.18215787410736084,
0.09362512826919556,
-0.034654177725315094,
0.306715726852417,
-0.10441838204860687,
-0.26171571016311646,
0.06753164529800415,
0.2798410654067993,
0.06890755891799927,
0.03763963282108307,
-0.01141345128417015,
0.09636452794075012,
-0.029762856662273407,
0.12914706766605377,
0.35980215668678284,
0.08750754594802856,
0.3100531995296478,
-0.47532209753990173,
-0.23672881722450256,
0.12255635857582092,
0.15189895033836365,
-0.049161992967128754,
0.2514217793941498,
0.22029903531074524,
-0.017325110733509064,
0.16476306319236755,
-0.13505451381206512,
0.5191757082939148,
-0.027538320049643517,
0.1664898693561554,
0.2251565158367157,
-0.1361297369003296,
0.046081990003585815,
0.23502855002880096,
0.004768013022840023,
0.23717498779296875,
0.140306755900383,
0.20365101099014282,
-0.034218572080135345,
0.28725671768188477,
0.21707597374916077,
0.3708851933479309,
-0.10904471576213837,
-0.26081207394599915,
0.07392021268606186,
-0.021033570170402527,
0.23460066318511963,
-0.0018411390483379364,
0.08577077090740204,
-0.06733561307191849,
-0.05838874727487564,
-0.3515784442424774,
0.33687272667884827,
0.019778823480010033,
-0.22592398524284363,
-0.15359972417354584,
-0.2364211082458496,
-0.22967620193958282,
0.1547272503376007,
0.10229087620973587,
-0.36245647072792053,
0.2183464914560318,
0.007728248834609985,
0.009116102010011673,
-0.14434979856014252,
-0.23814764618873596,
0.48364460468292236,
0.13227300345897675,
-0.12226594984531403,
0.12679094076156616,
-0.0402698777616024,
-0.20351697504520416,
0.10351915657520294,
0.24845130741596222,
0.050059545785188675,
-0.24018633365631104,
0.20259785652160645,
0.00718998908996582,
0.0745437741279602,
-0.2592480778694153,
-0.13190309703350067,
0.22991108894348145,
0.39236244559288025,
0.04802317172288895,
0.29389914870262146,
0.29248830676078796,
-0.20978206396102905,
0.0978928804397583,
0.03028181381523609,
0.2548646032810211,
-0.1333400011062622,
-0.00914277508854866,
-0.004980900790542364,
-0.13508035242557526,
-0.08546395599842072,
-0.08916810154914856,
-0.3414038121700287,
0.21731768548488617,
-0.04645419120788574,
-0.03395267575979233,
0.09335625916719437,
-0.11549031734466553,
0.18351863324642181,
-0.14800837635993958,
0.26010215282440186,
0.4892317056655884,
-0.08944222331047058,
0.05857408791780472,
-0.6094845533370972,
-0.41800493001937866,
0.01447698101401329,
-0.1829916536808014,
-0.13300928473472595,
-0.07280509173870087,
0.07216663658618927,
0.2422393560409546,
0.28405681252479553,
0.00780988484621048,
0.08722425997257233,
0.12835639715194702,
-0.014774831011891365,
-0.25689437985420227,
0.002415392082184553,
0.24349664151668549,
0.16879603266716003,
-0.14597102999687195,
0.04134267568588257,
0.16709232330322266,
-0.11496052145957947,
0.18395984172821045,
0.046328261494636536,
-0.09194129705429077,
-0.2813456058502197,
-0.22277165949344635,
0.6259546279907227,
0.10156749188899994,
-0.012621186673641205,
-0.18413889408111572,
-0.2473951280117035,
-0.18968182802200317,
-0.3732616603374481,
0.05168456584215164,
0.23292966187000275,
0.06607568264007568,
0.04630768299102783,
-0.20332905650138855,
0.03557569161057472,
0.007576841861009598,
0.31934189796447754,
0.20627108216285706,
0.05201311036944389,
-0.1563493013381958,
0.13168326020240784,
-0.18178662657737732,
-0.18289315700531006,
0.04596850275993347,
0.2691040635108948,
-0.23441565036773682,
-0.3153870105743408,
0.07056167721748352,
-0.2960246503353119,
0.24583424627780914,
-0.6862848997116089,
-0.18998919427394867,
-0.0055399686098098755,
-0.030638106167316437,
0.37103360891342163,
-0.05261297523975372,
0.005458042025566101,
0.2557203769683838,
0.247173011302948,
-0.06171877682209015,
-0.19959470629692078,
0.0682690292596817,
-0.04233193397521973,
-0.07502470910549164,
-0.1373392939567566,
0.3212102949619293,
0.19977745413780212,
-0.41600558161735535,
0.024148546159267426,
-0.030944112688302994
] |
https://github.com/huggingface/datasets/issues/6623 | @mariosasko, @lhoestq, @albertvillanova
hey guys! can anyone help? or can you guys suggest who can help with this? | streaming datasets doesn't work properly with multi-node | ### Feature request
Let’s say I have a dataset with 5 samples with values [1, 2, 3, 4, 5], with 2 GPUs (for DDP) and batch size of 2. This dataset is an `IterableDataset` since I am streaming it.
Now I split the dataset using `split_dataset_by_node` to ensure it doesn’t get repeated. And since it’s already splitted, I don’t have to use `DistributedSampler` (also they don't work with iterable datasets anyway)?
But in this case I noticed that the:
First iteraton:
first GPU will get → [1, 2]
first GPU will get → [3, 4]
Second iteraton:
first GPU will get → [5]
first GPU will get → Nothing
which actually creates an issue since in case of `DistributedSampler`, the samples are repeated internally to ensure non of the GPUs at any iteration is missing any data for gradient sync.
So my questions are:
1. Here since splitting is happening before hand, how to make sure each GPU get’s a batch at each iteration to avoid gradient sync issues?
2. Do we need to use `DistributedSampler`? If yes, how?
3. in the docstrings of `split_dataset_by_node`, this is mentioned: *"If the dataset has a number of shards that is a factor of `world_size` (i.e. if `dataset.n_shards % world_size == 0`), then the shards are evenly assigned across the nodes, which is the most optimized. Otherwise, each node keeps 1 example out of `world_size`, skipping the other examples."* Can you explain the last part here?
4. If `dataset.n_shards % world_size != 0`, is it possible to shard the streaming dataset on the fly to avoid the case where data is missing?
### Motivation
Somehow streaming datasets should work with DDP since for big LLMs a lot of data is required and DDP/multi-node is mostly used to train such models and streaming can actually help solve the data part of it.
### Your contribution
Yes, I can help in submitting the PR once we get mutual understanding on how it should behave. | 18 | streaming datasets doesn't work properly with multi-node
### Feature request
Let’s say I have a dataset with 5 samples with values [1, 2, 3, 4, 5], with 2 GPUs (for DDP) and batch size of 2. This dataset is an `IterableDataset` since I am streaming it.
Now I split the dataset using `split_dataset_by_node` to ensure it doesn’t get repeated. And since it’s already splitted, I don’t have to use `DistributedSampler` (also they don't work with iterable datasets anyway)?
But in this case I noticed that the:
First iteraton:
first GPU will get → [1, 2]
first GPU will get → [3, 4]
Second iteraton:
first GPU will get → [5]
first GPU will get → Nothing
which actually creates an issue since in case of `DistributedSampler`, the samples are repeated internally to ensure non of the GPUs at any iteration is missing any data for gradient sync.
So my questions are:
1. Here since splitting is happening before hand, how to make sure each GPU get’s a batch at each iteration to avoid gradient sync issues?
2. Do we need to use `DistributedSampler`? If yes, how?
3. in the docstrings of `split_dataset_by_node`, this is mentioned: *"If the dataset has a number of shards that is a factor of `world_size` (i.e. if `dataset.n_shards % world_size == 0`), then the shards are evenly assigned across the nodes, which is the most optimized. Otherwise, each node keeps 1 example out of `world_size`, skipping the other examples."* Can you explain the last part here?
4. If `dataset.n_shards % world_size != 0`, is it possible to shard the streaming dataset on the fly to avoid the case where data is missing?
### Motivation
Somehow streaming datasets should work with DDP since for big LLMs a lot of data is required and DDP/multi-node is mostly used to train such models and streaming can actually help solve the data part of it.
### Your contribution
Yes, I can help in submitting the PR once we get mutual understanding on how it should behave.
@mariosasko, @lhoestq, @albertvillanova
hey guys! can anyone help? or can you guys suggest who can help with this? | [
-0.5885023474693298,
-0.2717367708683014,
-0.022129017859697342,
0.05329866707324982,
-0.05288459733128548,
-0.24468931555747986,
0.28833165764808655,
-0.16213496029376984,
-0.05456581711769104,
0.17409411072731018,
0.13106313347816467,
0.20888292789459229,
-0.21015878021717072,
0.1710551381111145,
-0.05287248641252518,
-0.2316754013299942,
0.07644330710172653,
0.01057659462094307,
0.04034364968538284,
-0.2694319486618042,
-0.07909917831420898,
-0.16820268332958221,
-0.22020316123962402,
-0.4776022434234619,
-0.23010557889938354,
-0.34483352303504944,
-0.1880214959383011,
0.17915897071361542,
0.04292941093444824,
-0.0007918775081634521,
-0.024187937378883362,
0.21599668264389038,
0.22667305171489716,
0.26554247736930847,
-0.00011949403415201232,
-0.17549662292003632,
0.14307358860969543,
-0.033987656235694885,
-0.2797178626060486,
-0.10833810269832611,
-0.4525478482246399,
0.12000128626823425,
-0.0856480672955513,
-0.15769720077514648,
-0.005377188324928284,
0.0469653457403183,
0.16009725630283356,
-0.5938501954078674,
0.4552592635154724,
-0.03413572907447815,
0.08512648195028305,
0.01408693939447403,
-0.2336328625679016,
0.11152130365371704,
-0.18026015162467957,
0.3543376922607422,
0.005941025912761688,
-0.023008089512586594,
0.3836512565612793,
-0.10668213665485382,
-0.2726716697216034,
0.3208426833152771,
-0.17977496981620789,
0.3480873107910156,
-0.009544376283884048,
-0.1916971206665039,
-0.13937494158744812,
-0.3502672016620636,
-0.16309091448783875,
0.19770018756389618,
0.24191229045391083,
-0.12461245805025101,
-0.39788925647735596,
-0.521531343460083,
-0.14842446148395538,
-0.3500770926475525,
-0.23801052570343018,
0.18176981806755066,
-0.2481730580329895,
-0.0703870952129364,
-0.4510071277618408,
-0.23377658426761627,
-0.039797209203243256,
0.11789821088314056,
0.15419495105743408,
0.7694438695907593,
0.16675050556659698,
0.23403680324554443,
-0.1394328773021698,
0.29044508934020996,
0.3079175055027008,
-0.2309572994709015,
0.09334338456392288,
-0.13752999901771545,
-0.4325198531150818,
0.018086589872837067,
0.05734461918473244,
0.04305759817361832,
0.16307535767555237,
0.47958433628082275,
0.30966776609420776,
0.039084240794181824,
0.2876565754413605,
0.16717436909675598,
0.39613404870033264,
-0.18216004967689514,
-0.07782210409641266,
0.13634347915649414,
0.2166709154844284,
-0.017559371888637543,
0.08888927102088928,
0.020971247926354408,
0.21538354456424713,
0.23815034329891205,
-0.24705252051353455,
0.18153789639472961,
0.21558597683906555,
0.06319627165794373,
-0.057585135102272034,
-0.33602985739707947,
-0.03557462990283966,
-0.1887940913438797,
0.13759249448776245,
0.04293625056743622,
0.03659946843981743,
0.24672368168830872,
-0.4028873145580292,
0.15768982470035553,
-0.16016718745231628,
-0.5503530502319336,
-0.21842163801193237,
-0.3020685017108917,
-0.15051349997520447,
0.36816567182540894,
0.14705823361873627,
-0.0330776646733284,
0.2690260112285614,
-0.02857164293527603,
0.10693713277578354,
0.1553390473127365,
-0.07142554968595505,
-0.27262142300605774,
0.024798860773444176,
0.23033371567726135,
0.18815961480140686,
0.09010680019855499,
0.036062948405742645,
0.08091837167739868,
-0.07104326784610748,
0.17591574788093567,
-0.01980457827448845,
-0.33053672313690186,
0.19016502797603607,
0.08777797967195511,
-0.019701950252056122,
-0.013415093533694744,
-0.5710844993591309,
0.4722989499568939,
0.330960214138031,
-0.22632598876953125,
0.00015896186232566833,
-0.056308429688215256,
-0.37576544284820557,
-0.22712896764278412,
0.32718145847320557,
0.4236840605735779,
-0.08281472325325012,
-0.11036442220211029,
-0.07954616844654083,
0.10516591370105743,
0.559253990650177,
0.4052501320838928,
-0.17483340203762054,
-0.07906554639339447,
0.020856671035289764,
0.15972900390625,
0.16132764518260956,
-0.2096678912639618,
-0.25260594487190247,
0.4820108711719513,
-0.24641740322113037,
0.2487398236989975,
0.35342103242874146,
0.1166631281375885,
0.617304801940918,
-0.17541232705116272,
-0.07058148086071014,
0.6404229402542114,
-0.4585624933242798,
0.08418096601963043,
-0.38688284158706665,
-0.005343355238437653,
-0.10611443221569061,
0.25615841150283813,
0.16894984245300293,
0.12541131675243378,
0.24003133177757263,
0.2963806092739105,
0.3235984146595001,
0.08058828860521317,
0.35313987731933594,
-0.06989148259162903,
0.2761307954788208,
-0.2106955647468567,
0.049497976899147034,
-0.11211924254894257,
-0.6091260313987732,
0.28869563341140747,
0.07801954448223114,
-0.27191469073295593,
0.0680549144744873,
-0.23662494122982025,
0.44753289222717285,
-0.15206006169319153,
-0.0822150856256485,
-0.23689866065979004,
-0.03141500800848007,
0.26023662090301514,
0.2944309711456299,
-0.10916776955127716,
-0.3123038113117218,
0.2191813886165619,
-0.42277610301971436,
-0.02126145362854004,
-0.26497286558151245,
0.5215786099433899,
0.21814259886741638,
-0.21846994757652283,
-0.0010435879230499268,
0.045710932463407516,
0.14224129915237427,
-0.15118201076984406,
0.015415862202644348,
0.33918988704681396,
0.2921460270881653,
0.18274448812007904,
-0.08908434212207794,
0.1525992751121521,
0.40113458037376404,
-0.08527597039937973,
0.06607906520366669,
0.24203136563301086,
0.003439906984567642,
-0.3257346451282501,
-0.27416977286338806,
0.18110360205173492,
-0.15961357951164246,
0.3226219117641449,
0.12815937399864197,
0.05069836974143982,
0.2503218650817871,
-0.04333620145916939,
-0.1160278394818306,
0.07843576371669769,
0.10563630610704422,
-0.3318949341773987,
0.23888325691223145,
-0.10024002194404602,
-0.5517885088920593,
0.38689136505126953,
-0.007618499919772148,
0.06681092083454132,
-0.025056421756744385,
-0.183919757604599,
-0.0696144700050354,
-0.13988539576530457,
0.47877833247184753,
0.5028660297393799,
0.47347626090049744,
0.17877982556819916,
0.3006388545036316,
-0.029582621529698372,
0.09126587957143784,
-0.3178114593029022,
-0.0831855833530426,
0.04881472885608673,
0.23611204326152802,
0.2955628037452698,
-0.11971695721149445,
-0.04904571920633316,
-0.1970173865556717,
-0.23455947637557983,
0.3719569742679596,
0.05490919202566147,
-0.1710328459739685,
0.06568951159715652,
-0.31892722845077515,
-0.4927380681037903,
-0.22923643887043,
-0.045837175101041794,
-0.05014248937368393,
-0.17293314635753632,
0.06627776473760605,
0.7270770072937012,
-0.06778944283723831,
0.28438833355903625,
0.47979599237442017,
0.4485873579978943,
-0.21059271693229675,
0.1654486060142517,
-0.14575327932834625,
-0.18223978579044342,
0.0032022353261709213,
0.0369364470243454,
0.22898724675178528,
0.09430357813835144,
0.33727672696113586,
0.28389203548431396,
-0.16278596222400665,
-0.2981051802635193,
0.10762080550193787,
-0.09899180382490158,
0.00272422656416893,
-0.0963461697101593,
-0.17546817660331726,
-0.25050976872444153,
-0.03215603902935982,
0.003959033638238907,
-0.07086239755153656,
-0.24277536571025848,
-0.02813585475087166,
-0.013979624956846237,
-0.05706073343753815,
-0.2292422503232956,
-0.10658164322376251,
-0.008825637400150299,
-0.4428950250148773,
-0.33722180128097534,
0.03422681242227554,
-0.023857880383729935,
0.21627111732959747,
-0.106937475502491,
0.09920388460159302,
-0.317745178937912,
0.13228967785835266,
-0.2115560919046402,
0.004646945744752884,
-0.2248537242412567,
0.10413463413715363,
-0.20951318740844727,
-0.3363834023475647,
-0.18087321519851685,
0.016225799918174744,
-0.032758019864559174,
0.7001858353614807,
-0.3680460453033447,
-0.022672176361083984,
0.22942234575748444,
-0.28300631046295166,
0.13002845644950867,
-0.01745235174894333,
0.1841202676296234,
-0.21101859211921692,
0.0608733668923378,
-0.09832748770713806,
0.30754420161247253,
0.09404551982879639,
0.11913974583148956,
0.3478055000305176,
0.24098807573318481,
0.2969389855861664,
0.18660897016525269,
0.9571035504341125,
0.4824317395687103,
-0.04224644973874092,
0.14983172714710236,
0.19642145931720734,
-0.056599751114845276,
0.012742290273308754,
-0.0643312931060791,
0.3703346848487854,
-0.0830574780702591,
-0.045186202973127365,
0.12593674659729004,
-0.21250998973846436,
-0.34287744760513306,
-0.2323751598596573,
0.0019959434866905212,
0.11636404693126678,
-0.08302852511405945,
0.09244918823242188,
-0.2529964745044708,
0.17085915803909302,
0.15478980541229248,
0.08479215204715729,
0.20761753618717194,
-0.03211395442485809,
-0.4283945858478546,
0.06317974627017975,
0.27598223090171814,
0.016005147248506546,
-0.16827118396759033,
0.08855253458023071,
-0.33251258730888367,
0.20825934410095215,
0.11089914292097092,
0.41010433435440063,
-0.09692040830850601,
0.0012383460998535156,
0.20487654209136963,
0.16055065393447876,
0.2482411116361618,
-0.4635753929615021,
-0.15831811726093292,
-0.19165050983428955,
-0.3073832094669342,
0.14557787775993347,
-0.06610328704118729,
-0.13106635212898254,
0.39359205961227417,
-0.03613927215337753,
0.06572027504444122,
-0.19327493011951447,
0.00100698322057724,
-0.2515191435813904,
-0.15094858407974243,
-0.19483473896980286,
0.015966925770044327,
-0.22142896056175232,
-0.05043540522456169,
-0.26836690306663513,
0.20394307374954224,
0.11528261750936508,
-0.03895575553178787,
-0.2019885629415512,
-0.12639975547790527,
-0.17505629360675812,
0.09587056189775467,
-0.061167456209659576,
0.06695056706666946,
0.04935375973582268,
0.040436554700136185,
0.21243011951446533,
0.20874544978141785,
0.3953591585159302,
0.5721637606620789,
0.3576987087726593,
-0.10423833131790161,
-0.20544858276844025,
0.21369221806526184,
-0.012208551168441772,
0.35178130865097046,
0.4592912197113037,
0.059458423405885696,
0.12396785616874695,
0.199274480342865,
0.22486010193824768,
-0.3814557194709778,
-0.3206067383289337,
-0.04009976610541344,
-0.01633504405617714,
-0.3836514949798584,
-0.6445526480674744,
0.19172415137290955,
-0.08655966073274612,
-0.09299278259277344,
0.46199873089790344,
-0.20214125514030457,
-0.18130603432655334,
0.3067103624343872,
0.11880967020988464,
0.7116556763648987,
0.2004757821559906,
-0.1280091106891632,
-0.08082973957061768,
0.14721162617206573,
0.6211409568786621,
-0.12598812580108643,
0.30969175696372986,
-0.09895922243595123,
-0.34365129470825195,
-0.09388183802366257,
0.02876719832420349,
0.218648761510849,
0.3569571375846863,
-0.09297424554824829,
0.3380039632320404,
0.2797406017780304,
0.05031014606356621,
0.024454642087221146,
0.38943755626678467,
0.2260688841342926,
-0.29189732670783997,
0.05191393569111824,
0.040939681231975555,
-0.26360800862312317,
0.1095682680606842,
-0.029113762080669403,
0.06142812594771385,
0.30432945489883423,
-0.017380662262439728,
-0.3451733887195587,
-0.001589260995388031,
-0.2504502534866333,
0.1799796223640442,
-0.2980843186378479,
-0.3518190383911133,
-0.014309242367744446,
0.22457793354988098,
-0.051153820008039474,
-0.4006478190422058,
-0.060093075037002563,
0.1267896294593811,
0.09874902665615082,
0.1411825716495514,
-0.23552784323692322,
-0.2179218977689743,
0.28058183193206787,
-0.06922587007284164,
-0.07415220141410828,
0.032699450850486755,
0.1559431254863739,
-0.4639070928096771,
-0.21319936215877533,
-0.03263254463672638,
0.19737188518047333,
0.26600706577301025,
0.28038570284843445,
0.13937042653560638,
-0.07972268015146255,
-0.2499895989894867,
-0.001124952919781208,
-0.0405939482152462,
-0.15223562717437744,
0.531467080116272,
-0.15184012055397034,
0.0904516875743866,
-0.20528092980384827,
0.3922710418701172,
0.1825878918170929,
0.1486111581325531,
-0.02756732702255249,
0.11244922876358032,
-0.1182924211025238,
-0.11030668020248413,
-0.09317553043365479,
-0.06052488088607788,
-0.11856216937303543,
0.008454520255327225,
-0.29406440258026123,
0.19962066411972046,
0.27630722522735596,
-0.016298582777380943,
0.07249584794044495,
0.3035910129547119,
-0.25033047795295715,
-0.457568883895874,
0.10085532814264297,
0.08082135021686554,
-0.07574636489152908,
-0.10017263889312744,
0.08798982203006744,
-0.11817143857479095,
0.15497618913650513,
0.6689578890800476,
-0.2593654990196228,
-0.22264665365219116,
0.009598851203918457,
0.2004629671573639,
0.08540656417608261,
0.1396985650062561,
-0.08338335156440735,
-0.07921189069747925,
-0.12510380148887634,
0.07765056192874908,
0.024542182683944702,
-0.1204695850610733,
-0.0428435355424881,
0.19635778665542603,
0.21678051352500916,
-0.09578903019428253,
-0.23582208156585693,
-0.324596643447876,
-0.3343942165374756,
-0.10725420713424683,
0.20262300968170166,
0.012907840311527252,
-0.0461440235376358,
-0.08299568295478821,
0.6750946640968323,
0.08769296854734421,
-0.3021237552165985,
0.24845552444458008,
-0.0822640061378479,
-0.050996020436286926,
-0.17750069499015808,
0.19915303587913513,
0.3842018246650696,
-0.12150600552558899,
-0.17338357865810394,
-0.23282815515995026,
0.37631648778915405,
-0.08606995642185211,
0.4022093415260315,
-0.08369158208370209,
0.15544922649860382,
-0.16203370690345764,
0.5856527090072632,
0.23155951499938965,
-0.04679866507649422,
-0.20633524656295776,
0.28424015641212463,
0.10000159591436386,
-0.20904818177223206,
-0.20226672291755676,
0.051618073135614395,
0.33272844552993774,
0.1875542551279068,
0.4447084665298462,
0.39823317527770996,
-0.12155617773532867,
-0.2642560303211212,
0.09424786269664764,
0.5052286386489868,
-0.19930879771709442,
0.20498985052108765,
0.0005274559371173382,
0.13664285838603973,
0.19577041268348694,
-0.1093682050704956,
0.22564071416854858,
0.03534727543592453,
0.304769366979599,
-0.3092609941959381,
0.7399809956550598,
0.1800457090139389,
-0.04088064283132553,
0.26147162914276123,
-0.4083422124385834,
-0.027955103665590286,
0.0511932373046875,
-0.27629613876342773,
0.21890032291412354,
-0.11413562297821045,
0.05269940197467804,
-0.2647417187690735,
-0.22853687405586243,
-0.014623750001192093,
0.2962864339351654,
-0.002693724352866411,
0.031127575784921646,
-0.5655739307403564,
-0.008411988615989685,
-0.20485199987888336,
0.0021113306283950806,
0.028320545330643654,
-0.03827006742358208,
0.5511711835861206,
0.07702211290597916,
-0.041578423231840134,
-0.21449188888072968,
-0.06528092920780182,
0.05460263416171074,
0.14385336637496948,
-0.28721684217453003,
0.2275797724723816,
0.3027360439300537,
-0.13196931779384613,
0.07006557285785675,
0.1666833758354187,
0.07416398823261261,
0.26832813024520874,
-0.053863245993852615,
0.2339264303445816,
0.3050244450569153,
-0.06516563147306442,
-0.09334377199411392,
0.40177035331726074,
0.03326905518770218,
-0.15578347444534302,
0.2595016062259674,
-0.00023682788014411926,
0.017757251858711243,
0.2346586138010025,
0.2817247807979584,
-0.06593585014343262,
0.025959346443414688,
0.459255188703537,
-0.06756645441055298,
0.14546972513198853,
0.13626807928085327,
0.005246706306934357,
-0.37799328565597534,
-0.01038924977183342,
0.350979745388031,
-0.45905059576034546,
0.05125473812222481,
-0.022730836644768715,
0.004870064556598663,
-0.1468215137720108,
0.5160329937934875,
0.16500578820705414,
0.03988877683877945,
-0.2612239420413971,
-0.29604411125183105,
-0.5132619738578796,
0.08484984189271927,
-0.4117299020290375,
-0.16564467549324036,
-0.02849666215479374,
0.43281829357147217,
0.07081688940525055,
0.19639864563941956,
-0.3687134385108948,
0.08428403735160828,
-0.1745309680700302,
0.5773721933364868,
-0.2791088819503784,
-0.08937016129493713,
-0.5378432273864746,
0.15635201334953308,
0.09824883192777634,
-0.3539040684700012,
0.45212867856025696,
0.26110634207725525,
-0.1391412615776062,
0.07427337765693665,
0.005071297287940979,
0.26286131143569946,
0.005235630087554455,
-0.3676958382129669,
0.09250601381063461,
0.33780398964881897,
0.0506775826215744,
0.09396832436323166,
-0.0835827887058258,
0.12262168526649475,
-0.4526345729827881,
0.19286896288394928,
0.09724687039852142,
0.3620455265045166,
-0.19569164514541626,
-0.15890656411647797,
-0.29978421330451965,
0.35243454575538635,
0.03415100276470184,
-0.20737066864967346,
-0.27364835143089294,
0.08183158934116364,
-0.1954088807106018,
0.2924795150756836,
0.3380967974662781,
0.14995163679122925,
-0.23145905137062073,
-0.12317950278520584,
-0.31143924593925476,
-0.1804192215204239,
0.3792820870876312,
-0.47693243622779846,
0.045383088290691376,
-0.4637560248374939,
0.23137111961841583,
0.2531643509864807,
0.1970571130514145,
-0.25030016899108887,
-0.3180619478225708,
0.22695736587047577,
-0.32072004675865173,
0.057510972023010254,
0.07607784122228622,
0.12044914066791534,
-0.2699720561504364,
-0.1574166715145111,
-0.40562427043914795,
0.032632339745759964,
-0.5015792846679688,
0.22198708355426788,
-0.30460089445114136
] |
https://github.com/huggingface/datasets/issues/6623 | Hi !
1. When the dataset is running of of examples, the last batches received by the GPU can be incomplete or empty/missing. We haven't implemented yet a way to ignore the last batch. It might require the datasets to provide the number of examples per shard though, so that we can know when to stop.
2. Samplers are not compatible with IterableDatasets in pytorch
3. if `dataset.n_shards % world_size != 0` then all the nodes will read/stream the full dataset in order (possibly reading/streaming the same data multiple times), BUT will only yield one example out of `world_size` so that each example goes to one exactly one GPU.
4. no, sharding should be down up-front and can take some time depending on the dataset size and format | streaming datasets doesn't work properly with multi-node | ### Feature request
Let’s say I have a dataset with 5 samples with values [1, 2, 3, 4, 5], with 2 GPUs (for DDP) and batch size of 2. This dataset is an `IterableDataset` since I am streaming it.
Now I split the dataset using `split_dataset_by_node` to ensure it doesn’t get repeated. And since it’s already splitted, I don’t have to use `DistributedSampler` (also they don't work with iterable datasets anyway)?
But in this case I noticed that the:
First iteraton:
first GPU will get → [1, 2]
first GPU will get → [3, 4]
Second iteraton:
first GPU will get → [5]
first GPU will get → Nothing
which actually creates an issue since in case of `DistributedSampler`, the samples are repeated internally to ensure non of the GPUs at any iteration is missing any data for gradient sync.
So my questions are:
1. Here since splitting is happening before hand, how to make sure each GPU get’s a batch at each iteration to avoid gradient sync issues?
2. Do we need to use `DistributedSampler`? If yes, how?
3. in the docstrings of `split_dataset_by_node`, this is mentioned: *"If the dataset has a number of shards that is a factor of `world_size` (i.e. if `dataset.n_shards % world_size == 0`), then the shards are evenly assigned across the nodes, which is the most optimized. Otherwise, each node keeps 1 example out of `world_size`, skipping the other examples."* Can you explain the last part here?
4. If `dataset.n_shards % world_size != 0`, is it possible to shard the streaming dataset on the fly to avoid the case where data is missing?
### Motivation
Somehow streaming datasets should work with DDP since for big LLMs a lot of data is required and DDP/multi-node is mostly used to train such models and streaming can actually help solve the data part of it.
### Your contribution
Yes, I can help in submitting the PR once we get mutual understanding on how it should behave. | 128 | streaming datasets doesn't work properly with multi-node
### Feature request
Let’s say I have a dataset with 5 samples with values [1, 2, 3, 4, 5], with 2 GPUs (for DDP) and batch size of 2. This dataset is an `IterableDataset` since I am streaming it.
Now I split the dataset using `split_dataset_by_node` to ensure it doesn’t get repeated. And since it’s already splitted, I don’t have to use `DistributedSampler` (also they don't work with iterable datasets anyway)?
But in this case I noticed that the:
First iteraton:
first GPU will get → [1, 2]
first GPU will get → [3, 4]
Second iteraton:
first GPU will get → [5]
first GPU will get → Nothing
which actually creates an issue since in case of `DistributedSampler`, the samples are repeated internally to ensure non of the GPUs at any iteration is missing any data for gradient sync.
So my questions are:
1. Here since splitting is happening before hand, how to make sure each GPU get’s a batch at each iteration to avoid gradient sync issues?
2. Do we need to use `DistributedSampler`? If yes, how?
3. in the docstrings of `split_dataset_by_node`, this is mentioned: *"If the dataset has a number of shards that is a factor of `world_size` (i.e. if `dataset.n_shards % world_size == 0`), then the shards are evenly assigned across the nodes, which is the most optimized. Otherwise, each node keeps 1 example out of `world_size`, skipping the other examples."* Can you explain the last part here?
4. If `dataset.n_shards % world_size != 0`, is it possible to shard the streaming dataset on the fly to avoid the case where data is missing?
### Motivation
Somehow streaming datasets should work with DDP since for big LLMs a lot of data is required and DDP/multi-node is mostly used to train such models and streaming can actually help solve the data part of it.
### Your contribution
Yes, I can help in submitting the PR once we get mutual understanding on how it should behave.
Hi !
1. When the dataset is running of of examples, the last batches received by the GPU can be incomplete or empty/missing. We haven't implemented yet a way to ignore the last batch. It might require the datasets to provide the number of examples per shard though, so that we can know when to stop.
2. Samplers are not compatible with IterableDatasets in pytorch
3. if `dataset.n_shards % world_size != 0` then all the nodes will read/stream the full dataset in order (possibly reading/streaming the same data multiple times), BUT will only yield one example out of `world_size` so that each example goes to one exactly one GPU.
4. no, sharding should be down up-front and can take some time depending on the dataset size and format | [
-0.5885023474693298,
-0.2717367708683014,
-0.022129017859697342,
0.05329866707324982,
-0.05288459733128548,
-0.24468931555747986,
0.28833165764808655,
-0.16213496029376984,
-0.05456581711769104,
0.17409411072731018,
0.13106313347816467,
0.20888292789459229,
-0.21015878021717072,
0.1710551381111145,
-0.05287248641252518,
-0.2316754013299942,
0.07644330710172653,
0.01057659462094307,
0.04034364968538284,
-0.2694319486618042,
-0.07909917831420898,
-0.16820268332958221,
-0.22020316123962402,
-0.4776022434234619,
-0.23010557889938354,
-0.34483352303504944,
-0.1880214959383011,
0.17915897071361542,
0.04292941093444824,
-0.0007918775081634521,
-0.024187937378883362,
0.21599668264389038,
0.22667305171489716,
0.26554247736930847,
-0.00011949403415201232,
-0.17549662292003632,
0.14307358860969543,
-0.033987656235694885,
-0.2797178626060486,
-0.10833810269832611,
-0.4525478482246399,
0.12000128626823425,
-0.0856480672955513,
-0.15769720077514648,
-0.005377188324928284,
0.0469653457403183,
0.16009725630283356,
-0.5938501954078674,
0.4552592635154724,
-0.03413572907447815,
0.08512648195028305,
0.01408693939447403,
-0.2336328625679016,
0.11152130365371704,
-0.18026015162467957,
0.3543376922607422,
0.005941025912761688,
-0.023008089512586594,
0.3836512565612793,
-0.10668213665485382,
-0.2726716697216034,
0.3208426833152771,
-0.17977496981620789,
0.3480873107910156,
-0.009544376283884048,
-0.1916971206665039,
-0.13937494158744812,
-0.3502672016620636,
-0.16309091448783875,
0.19770018756389618,
0.24191229045391083,
-0.12461245805025101,
-0.39788925647735596,
-0.521531343460083,
-0.14842446148395538,
-0.3500770926475525,
-0.23801052570343018,
0.18176981806755066,
-0.2481730580329895,
-0.0703870952129364,
-0.4510071277618408,
-0.23377658426761627,
-0.039797209203243256,
0.11789821088314056,
0.15419495105743408,
0.7694438695907593,
0.16675050556659698,
0.23403680324554443,
-0.1394328773021698,
0.29044508934020996,
0.3079175055027008,
-0.2309572994709015,
0.09334338456392288,
-0.13752999901771545,
-0.4325198531150818,
0.018086589872837067,
0.05734461918473244,
0.04305759817361832,
0.16307535767555237,
0.47958433628082275,
0.30966776609420776,
0.039084240794181824,
0.2876565754413605,
0.16717436909675598,
0.39613404870033264,
-0.18216004967689514,
-0.07782210409641266,
0.13634347915649414,
0.2166709154844284,
-0.017559371888637543,
0.08888927102088928,
0.020971247926354408,
0.21538354456424713,
0.23815034329891205,
-0.24705252051353455,
0.18153789639472961,
0.21558597683906555,
0.06319627165794373,
-0.057585135102272034,
-0.33602985739707947,
-0.03557462990283966,
-0.1887940913438797,
0.13759249448776245,
0.04293625056743622,
0.03659946843981743,
0.24672368168830872,
-0.4028873145580292,
0.15768982470035553,
-0.16016718745231628,
-0.5503530502319336,
-0.21842163801193237,
-0.3020685017108917,
-0.15051349997520447,
0.36816567182540894,
0.14705823361873627,
-0.0330776646733284,
0.2690260112285614,
-0.02857164293527603,
0.10693713277578354,
0.1553390473127365,
-0.07142554968595505,
-0.27262142300605774,
0.024798860773444176,
0.23033371567726135,
0.18815961480140686,
0.09010680019855499,
0.036062948405742645,
0.08091837167739868,
-0.07104326784610748,
0.17591574788093567,
-0.01980457827448845,
-0.33053672313690186,
0.19016502797603607,
0.08777797967195511,
-0.019701950252056122,
-0.013415093533694744,
-0.5710844993591309,
0.4722989499568939,
0.330960214138031,
-0.22632598876953125,
0.00015896186232566833,
-0.056308429688215256,
-0.37576544284820557,
-0.22712896764278412,
0.32718145847320557,
0.4236840605735779,
-0.08281472325325012,
-0.11036442220211029,
-0.07954616844654083,
0.10516591370105743,
0.559253990650177,
0.4052501320838928,
-0.17483340203762054,
-0.07906554639339447,
0.020856671035289764,
0.15972900390625,
0.16132764518260956,
-0.2096678912639618,
-0.25260594487190247,
0.4820108711719513,
-0.24641740322113037,
0.2487398236989975,
0.35342103242874146,
0.1166631281375885,
0.617304801940918,
-0.17541232705116272,
-0.07058148086071014,
0.6404229402542114,
-0.4585624933242798,
0.08418096601963043,
-0.38688284158706665,
-0.005343355238437653,
-0.10611443221569061,
0.25615841150283813,
0.16894984245300293,
0.12541131675243378,
0.24003133177757263,
0.2963806092739105,
0.3235984146595001,
0.08058828860521317,
0.35313987731933594,
-0.06989148259162903,
0.2761307954788208,
-0.2106955647468567,
0.049497976899147034,
-0.11211924254894257,
-0.6091260313987732,
0.28869563341140747,
0.07801954448223114,
-0.27191469073295593,
0.0680549144744873,
-0.23662494122982025,
0.44753289222717285,
-0.15206006169319153,
-0.0822150856256485,
-0.23689866065979004,
-0.03141500800848007,
0.26023662090301514,
0.2944309711456299,
-0.10916776955127716,
-0.3123038113117218,
0.2191813886165619,
-0.42277610301971436,
-0.02126145362854004,
-0.26497286558151245,
0.5215786099433899,
0.21814259886741638,
-0.21846994757652283,
-0.0010435879230499268,
0.045710932463407516,
0.14224129915237427,
-0.15118201076984406,
0.015415862202644348,
0.33918988704681396,
0.2921460270881653,
0.18274448812007904,
-0.08908434212207794,
0.1525992751121521,
0.40113458037376404,
-0.08527597039937973,
0.06607906520366669,
0.24203136563301086,
0.003439906984567642,
-0.3257346451282501,
-0.27416977286338806,
0.18110360205173492,
-0.15961357951164246,
0.3226219117641449,
0.12815937399864197,
0.05069836974143982,
0.2503218650817871,
-0.04333620145916939,
-0.1160278394818306,
0.07843576371669769,
0.10563630610704422,
-0.3318949341773987,
0.23888325691223145,
-0.10024002194404602,
-0.5517885088920593,
0.38689136505126953,
-0.007618499919772148,
0.06681092083454132,
-0.025056421756744385,
-0.183919757604599,
-0.0696144700050354,
-0.13988539576530457,
0.47877833247184753,
0.5028660297393799,
0.47347626090049744,
0.17877982556819916,
0.3006388545036316,
-0.029582621529698372,
0.09126587957143784,
-0.3178114593029022,
-0.0831855833530426,
0.04881472885608673,
0.23611204326152802,
0.2955628037452698,
-0.11971695721149445,
-0.04904571920633316,
-0.1970173865556717,
-0.23455947637557983,
0.3719569742679596,
0.05490919202566147,
-0.1710328459739685,
0.06568951159715652,
-0.31892722845077515,
-0.4927380681037903,
-0.22923643887043,
-0.045837175101041794,
-0.05014248937368393,
-0.17293314635753632,
0.06627776473760605,
0.7270770072937012,
-0.06778944283723831,
0.28438833355903625,
0.47979599237442017,
0.4485873579978943,
-0.21059271693229675,
0.1654486060142517,
-0.14575327932834625,
-0.18223978579044342,
0.0032022353261709213,
0.0369364470243454,
0.22898724675178528,
0.09430357813835144,
0.33727672696113586,
0.28389203548431396,
-0.16278596222400665,
-0.2981051802635193,
0.10762080550193787,
-0.09899180382490158,
0.00272422656416893,
-0.0963461697101593,
-0.17546817660331726,
-0.25050976872444153,
-0.03215603902935982,
0.003959033638238907,
-0.07086239755153656,
-0.24277536571025848,
-0.02813585475087166,
-0.013979624956846237,
-0.05706073343753815,
-0.2292422503232956,
-0.10658164322376251,
-0.008825637400150299,
-0.4428950250148773,
-0.33722180128097534,
0.03422681242227554,
-0.023857880383729935,
0.21627111732959747,
-0.106937475502491,
0.09920388460159302,
-0.317745178937912,
0.13228967785835266,
-0.2115560919046402,
0.004646945744752884,
-0.2248537242412567,
0.10413463413715363,
-0.20951318740844727,
-0.3363834023475647,
-0.18087321519851685,
0.016225799918174744,
-0.032758019864559174,
0.7001858353614807,
-0.3680460453033447,
-0.022672176361083984,
0.22942234575748444,
-0.28300631046295166,
0.13002845644950867,
-0.01745235174894333,
0.1841202676296234,
-0.21101859211921692,
0.0608733668923378,
-0.09832748770713806,
0.30754420161247253,
0.09404551982879639,
0.11913974583148956,
0.3478055000305176,
0.24098807573318481,
0.2969389855861664,
0.18660897016525269,
0.9571035504341125,
0.4824317395687103,
-0.04224644973874092,
0.14983172714710236,
0.19642145931720734,
-0.056599751114845276,
0.012742290273308754,
-0.0643312931060791,
0.3703346848487854,
-0.0830574780702591,
-0.045186202973127365,
0.12593674659729004,
-0.21250998973846436,
-0.34287744760513306,
-0.2323751598596573,
0.0019959434866905212,
0.11636404693126678,
-0.08302852511405945,
0.09244918823242188,
-0.2529964745044708,
0.17085915803909302,
0.15478980541229248,
0.08479215204715729,
0.20761753618717194,
-0.03211395442485809,
-0.4283945858478546,
0.06317974627017975,
0.27598223090171814,
0.016005147248506546,
-0.16827118396759033,
0.08855253458023071,
-0.33251258730888367,
0.20825934410095215,
0.11089914292097092,
0.41010433435440063,
-0.09692040830850601,
0.0012383460998535156,
0.20487654209136963,
0.16055065393447876,
0.2482411116361618,
-0.4635753929615021,
-0.15831811726093292,
-0.19165050983428955,
-0.3073832094669342,
0.14557787775993347,
-0.06610328704118729,
-0.13106635212898254,
0.39359205961227417,
-0.03613927215337753,
0.06572027504444122,
-0.19327493011951447,
0.00100698322057724,
-0.2515191435813904,
-0.15094858407974243,
-0.19483473896980286,
0.015966925770044327,
-0.22142896056175232,
-0.05043540522456169,
-0.26836690306663513,
0.20394307374954224,
0.11528261750936508,
-0.03895575553178787,
-0.2019885629415512,
-0.12639975547790527,
-0.17505629360675812,
0.09587056189775467,
-0.061167456209659576,
0.06695056706666946,
0.04935375973582268,
0.040436554700136185,
0.21243011951446533,
0.20874544978141785,
0.3953591585159302,
0.5721637606620789,
0.3576987087726593,
-0.10423833131790161,
-0.20544858276844025,
0.21369221806526184,
-0.012208551168441772,
0.35178130865097046,
0.4592912197113037,
0.059458423405885696,
0.12396785616874695,
0.199274480342865,
0.22486010193824768,
-0.3814557194709778,
-0.3206067383289337,
-0.04009976610541344,
-0.01633504405617714,
-0.3836514949798584,
-0.6445526480674744,
0.19172415137290955,
-0.08655966073274612,
-0.09299278259277344,
0.46199873089790344,
-0.20214125514030457,
-0.18130603432655334,
0.3067103624343872,
0.11880967020988464,
0.7116556763648987,
0.2004757821559906,
-0.1280091106891632,
-0.08082973957061768,
0.14721162617206573,
0.6211409568786621,
-0.12598812580108643,
0.30969175696372986,
-0.09895922243595123,
-0.34365129470825195,
-0.09388183802366257,
0.02876719832420349,
0.218648761510849,
0.3569571375846863,
-0.09297424554824829,
0.3380039632320404,
0.2797406017780304,
0.05031014606356621,
0.024454642087221146,
0.38943755626678467,
0.2260688841342926,
-0.29189732670783997,
0.05191393569111824,
0.040939681231975555,
-0.26360800862312317,
0.1095682680606842,
-0.029113762080669403,
0.06142812594771385,
0.30432945489883423,
-0.017380662262439728,
-0.3451733887195587,
-0.001589260995388031,
-0.2504502534866333,
0.1799796223640442,
-0.2980843186378479,
-0.3518190383911133,
-0.014309242367744446,
0.22457793354988098,
-0.051153820008039474,
-0.4006478190422058,
-0.060093075037002563,
0.1267896294593811,
0.09874902665615082,
0.1411825716495514,
-0.23552784323692322,
-0.2179218977689743,
0.28058183193206787,
-0.06922587007284164,
-0.07415220141410828,
0.032699450850486755,
0.1559431254863739,
-0.4639070928096771,
-0.21319936215877533,
-0.03263254463672638,
0.19737188518047333,
0.26600706577301025,
0.28038570284843445,
0.13937042653560638,
-0.07972268015146255,
-0.2499895989894867,
-0.001124952919781208,
-0.0405939482152462,
-0.15223562717437744,
0.531467080116272,
-0.15184012055397034,
0.0904516875743866,
-0.20528092980384827,
0.3922710418701172,
0.1825878918170929,
0.1486111581325531,
-0.02756732702255249,
0.11244922876358032,
-0.1182924211025238,
-0.11030668020248413,
-0.09317553043365479,
-0.06052488088607788,
-0.11856216937303543,
0.008454520255327225,
-0.29406440258026123,
0.19962066411972046,
0.27630722522735596,
-0.016298582777380943,
0.07249584794044495,
0.3035910129547119,
-0.25033047795295715,
-0.457568883895874,
0.10085532814264297,
0.08082135021686554,
-0.07574636489152908,
-0.10017263889312744,
0.08798982203006744,
-0.11817143857479095,
0.15497618913650513,
0.6689578890800476,
-0.2593654990196228,
-0.22264665365219116,
0.009598851203918457,
0.2004629671573639,
0.08540656417608261,
0.1396985650062561,
-0.08338335156440735,
-0.07921189069747925,
-0.12510380148887634,
0.07765056192874908,
0.024542182683944702,
-0.1204695850610733,
-0.0428435355424881,
0.19635778665542603,
0.21678051352500916,
-0.09578903019428253,
-0.23582208156585693,
-0.324596643447876,
-0.3343942165374756,
-0.10725420713424683,
0.20262300968170166,
0.012907840311527252,
-0.0461440235376358,
-0.08299568295478821,
0.6750946640968323,
0.08769296854734421,
-0.3021237552165985,
0.24845552444458008,
-0.0822640061378479,
-0.050996020436286926,
-0.17750069499015808,
0.19915303587913513,
0.3842018246650696,
-0.12150600552558899,
-0.17338357865810394,
-0.23282815515995026,
0.37631648778915405,
-0.08606995642185211,
0.4022093415260315,
-0.08369158208370209,
0.15544922649860382,
-0.16203370690345764,
0.5856527090072632,
0.23155951499938965,
-0.04679866507649422,
-0.20633524656295776,
0.28424015641212463,
0.10000159591436386,
-0.20904818177223206,
-0.20226672291755676,
0.051618073135614395,
0.33272844552993774,
0.1875542551279068,
0.4447084665298462,
0.39823317527770996,
-0.12155617773532867,
-0.2642560303211212,
0.09424786269664764,
0.5052286386489868,
-0.19930879771709442,
0.20498985052108765,
0.0005274559371173382,
0.13664285838603973,
0.19577041268348694,
-0.1093682050704956,
0.22564071416854858,
0.03534727543592453,
0.304769366979599,
-0.3092609941959381,
0.7399809956550598,
0.1800457090139389,
-0.04088064283132553,
0.26147162914276123,
-0.4083422124385834,
-0.027955103665590286,
0.0511932373046875,
-0.27629613876342773,
0.21890032291412354,
-0.11413562297821045,
0.05269940197467804,
-0.2647417187690735,
-0.22853687405586243,
-0.014623750001192093,
0.2962864339351654,
-0.002693724352866411,
0.031127575784921646,
-0.5655739307403564,
-0.008411988615989685,
-0.20485199987888336,
0.0021113306283950806,
0.028320545330643654,
-0.03827006742358208,
0.5511711835861206,
0.07702211290597916,
-0.041578423231840134,
-0.21449188888072968,
-0.06528092920780182,
0.05460263416171074,
0.14385336637496948,
-0.28721684217453003,
0.2275797724723816,
0.3027360439300537,
-0.13196931779384613,
0.07006557285785675,
0.1666833758354187,
0.07416398823261261,
0.26832813024520874,
-0.053863245993852615,
0.2339264303445816,
0.3050244450569153,
-0.06516563147306442,
-0.09334377199411392,
0.40177035331726074,
0.03326905518770218,
-0.15578347444534302,
0.2595016062259674,
-0.00023682788014411926,
0.017757251858711243,
0.2346586138010025,
0.2817247807979584,
-0.06593585014343262,
0.025959346443414688,
0.459255188703537,
-0.06756645441055298,
0.14546972513198853,
0.13626807928085327,
0.005246706306934357,
-0.37799328565597534,
-0.01038924977183342,
0.350979745388031,
-0.45905059576034546,
0.05125473812222481,
-0.022730836644768715,
0.004870064556598663,
-0.1468215137720108,
0.5160329937934875,
0.16500578820705414,
0.03988877683877945,
-0.2612239420413971,
-0.29604411125183105,
-0.5132619738578796,
0.08484984189271927,
-0.4117299020290375,
-0.16564467549324036,
-0.02849666215479374,
0.43281829357147217,
0.07081688940525055,
0.19639864563941956,
-0.3687134385108948,
0.08428403735160828,
-0.1745309680700302,
0.5773721933364868,
-0.2791088819503784,
-0.08937016129493713,
-0.5378432273864746,
0.15635201334953308,
0.09824883192777634,
-0.3539040684700012,
0.45212867856025696,
0.26110634207725525,
-0.1391412615776062,
0.07427337765693665,
0.005071297287940979,
0.26286131143569946,
0.005235630087554455,
-0.3676958382129669,
0.09250601381063461,
0.33780398964881897,
0.0506775826215744,
0.09396832436323166,
-0.0835827887058258,
0.12262168526649475,
-0.4526345729827881,
0.19286896288394928,
0.09724687039852142,
0.3620455265045166,
-0.19569164514541626,
-0.15890656411647797,
-0.29978421330451965,
0.35243454575538635,
0.03415100276470184,
-0.20737066864967346,
-0.27364835143089294,
0.08183158934116364,
-0.1954088807106018,
0.2924795150756836,
0.3380967974662781,
0.14995163679122925,
-0.23145905137062073,
-0.12317950278520584,
-0.31143924593925476,
-0.1804192215204239,
0.3792820870876312,
-0.47693243622779846,
0.045383088290691376,
-0.4637560248374939,
0.23137111961841583,
0.2531643509864807,
0.1970571130514145,
-0.25030016899108887,
-0.3180619478225708,
0.22695736587047577,
-0.32072004675865173,
0.057510972023010254,
0.07607784122228622,
0.12044914066791534,
-0.2699720561504364,
-0.1574166715145111,
-0.40562427043914795,
0.032632339745759964,
-0.5015792846679688,
0.22198708355426788,
-0.30460089445114136
] |
https://github.com/huggingface/datasets/issues/6623 | > if dataset.n_shards % world_size != 0 then all the nodes will read/stream the full dataset in order (possibly reading/streaming the same data multiple times), BUT will only yield one example out of world_size so that each example goes to one exactly one GPU.
considering there's just 1 shard and 2 worker nodes, do you mean each worker node will load the whole dataset but still receive half of that shard while streaming? | streaming datasets doesn't work properly with multi-node | ### Feature request
Let’s say I have a dataset with 5 samples with values [1, 2, 3, 4, 5], with 2 GPUs (for DDP) and batch size of 2. This dataset is an `IterableDataset` since I am streaming it.
Now I split the dataset using `split_dataset_by_node` to ensure it doesn’t get repeated. And since it’s already splitted, I don’t have to use `DistributedSampler` (also they don't work with iterable datasets anyway)?
But in this case I noticed that the:
First iteraton:
first GPU will get → [1, 2]
first GPU will get → [3, 4]
Second iteraton:
first GPU will get → [5]
first GPU will get → Nothing
which actually creates an issue since in case of `DistributedSampler`, the samples are repeated internally to ensure non of the GPUs at any iteration is missing any data for gradient sync.
So my questions are:
1. Here since splitting is happening before hand, how to make sure each GPU get’s a batch at each iteration to avoid gradient sync issues?
2. Do we need to use `DistributedSampler`? If yes, how?
3. in the docstrings of `split_dataset_by_node`, this is mentioned: *"If the dataset has a number of shards that is a factor of `world_size` (i.e. if `dataset.n_shards % world_size == 0`), then the shards are evenly assigned across the nodes, which is the most optimized. Otherwise, each node keeps 1 example out of `world_size`, skipping the other examples."* Can you explain the last part here?
4. If `dataset.n_shards % world_size != 0`, is it possible to shard the streaming dataset on the fly to avoid the case where data is missing?
### Motivation
Somehow streaming datasets should work with DDP since for big LLMs a lot of data is required and DDP/multi-node is mostly used to train such models and streaming can actually help solve the data part of it.
### Your contribution
Yes, I can help in submitting the PR once we get mutual understanding on how it should behave. | 73 | streaming datasets doesn't work properly with multi-node
### Feature request
Let’s say I have a dataset with 5 samples with values [1, 2, 3, 4, 5], with 2 GPUs (for DDP) and batch size of 2. This dataset is an `IterableDataset` since I am streaming it.
Now I split the dataset using `split_dataset_by_node` to ensure it doesn’t get repeated. And since it’s already splitted, I don’t have to use `DistributedSampler` (also they don't work with iterable datasets anyway)?
But in this case I noticed that the:
First iteraton:
first GPU will get → [1, 2]
first GPU will get → [3, 4]
Second iteraton:
first GPU will get → [5]
first GPU will get → Nothing
which actually creates an issue since in case of `DistributedSampler`, the samples are repeated internally to ensure non of the GPUs at any iteration is missing any data for gradient sync.
So my questions are:
1. Here since splitting is happening before hand, how to make sure each GPU get’s a batch at each iteration to avoid gradient sync issues?
2. Do we need to use `DistributedSampler`? If yes, how?
3. in the docstrings of `split_dataset_by_node`, this is mentioned: *"If the dataset has a number of shards that is a factor of `world_size` (i.e. if `dataset.n_shards % world_size == 0`), then the shards are evenly assigned across the nodes, which is the most optimized. Otherwise, each node keeps 1 example out of `world_size`, skipping the other examples."* Can you explain the last part here?
4. If `dataset.n_shards % world_size != 0`, is it possible to shard the streaming dataset on the fly to avoid the case where data is missing?
### Motivation
Somehow streaming datasets should work with DDP since for big LLMs a lot of data is required and DDP/multi-node is mostly used to train such models and streaming can actually help solve the data part of it.
### Your contribution
Yes, I can help in submitting the PR once we get mutual understanding on how it should behave.
> if dataset.n_shards % world_size != 0 then all the nodes will read/stream the full dataset in order (possibly reading/streaming the same data multiple times), BUT will only yield one example out of world_size so that each example goes to one exactly one GPU.
considering there's just 1 shard and 2 worker nodes, do you mean each worker node will load the whole dataset but still receive half of that shard while streaming? | [
-0.5885023474693298,
-0.2717367708683014,
-0.022129017859697342,
0.05329866707324982,
-0.05288459733128548,
-0.24468931555747986,
0.28833165764808655,
-0.16213496029376984,
-0.05456581711769104,
0.17409411072731018,
0.13106313347816467,
0.20888292789459229,
-0.21015878021717072,
0.1710551381111145,
-0.05287248641252518,
-0.2316754013299942,
0.07644330710172653,
0.01057659462094307,
0.04034364968538284,
-0.2694319486618042,
-0.07909917831420898,
-0.16820268332958221,
-0.22020316123962402,
-0.4776022434234619,
-0.23010557889938354,
-0.34483352303504944,
-0.1880214959383011,
0.17915897071361542,
0.04292941093444824,
-0.0007918775081634521,
-0.024187937378883362,
0.21599668264389038,
0.22667305171489716,
0.26554247736930847,
-0.00011949403415201232,
-0.17549662292003632,
0.14307358860969543,
-0.033987656235694885,
-0.2797178626060486,
-0.10833810269832611,
-0.4525478482246399,
0.12000128626823425,
-0.0856480672955513,
-0.15769720077514648,
-0.005377188324928284,
0.0469653457403183,
0.16009725630283356,
-0.5938501954078674,
0.4552592635154724,
-0.03413572907447815,
0.08512648195028305,
0.01408693939447403,
-0.2336328625679016,
0.11152130365371704,
-0.18026015162467957,
0.3543376922607422,
0.005941025912761688,
-0.023008089512586594,
0.3836512565612793,
-0.10668213665485382,
-0.2726716697216034,
0.3208426833152771,
-0.17977496981620789,
0.3480873107910156,
-0.009544376283884048,
-0.1916971206665039,
-0.13937494158744812,
-0.3502672016620636,
-0.16309091448783875,
0.19770018756389618,
0.24191229045391083,
-0.12461245805025101,
-0.39788925647735596,
-0.521531343460083,
-0.14842446148395538,
-0.3500770926475525,
-0.23801052570343018,
0.18176981806755066,
-0.2481730580329895,
-0.0703870952129364,
-0.4510071277618408,
-0.23377658426761627,
-0.039797209203243256,
0.11789821088314056,
0.15419495105743408,
0.7694438695907593,
0.16675050556659698,
0.23403680324554443,
-0.1394328773021698,
0.29044508934020996,
0.3079175055027008,
-0.2309572994709015,
0.09334338456392288,
-0.13752999901771545,
-0.4325198531150818,
0.018086589872837067,
0.05734461918473244,
0.04305759817361832,
0.16307535767555237,
0.47958433628082275,
0.30966776609420776,
0.039084240794181824,
0.2876565754413605,
0.16717436909675598,
0.39613404870033264,
-0.18216004967689514,
-0.07782210409641266,
0.13634347915649414,
0.2166709154844284,
-0.017559371888637543,
0.08888927102088928,
0.020971247926354408,
0.21538354456424713,
0.23815034329891205,
-0.24705252051353455,
0.18153789639472961,
0.21558597683906555,
0.06319627165794373,
-0.057585135102272034,
-0.33602985739707947,
-0.03557462990283966,
-0.1887940913438797,
0.13759249448776245,
0.04293625056743622,
0.03659946843981743,
0.24672368168830872,
-0.4028873145580292,
0.15768982470035553,
-0.16016718745231628,
-0.5503530502319336,
-0.21842163801193237,
-0.3020685017108917,
-0.15051349997520447,
0.36816567182540894,
0.14705823361873627,
-0.0330776646733284,
0.2690260112285614,
-0.02857164293527603,
0.10693713277578354,
0.1553390473127365,
-0.07142554968595505,
-0.27262142300605774,
0.024798860773444176,
0.23033371567726135,
0.18815961480140686,
0.09010680019855499,
0.036062948405742645,
0.08091837167739868,
-0.07104326784610748,
0.17591574788093567,
-0.01980457827448845,
-0.33053672313690186,
0.19016502797603607,
0.08777797967195511,
-0.019701950252056122,
-0.013415093533694744,
-0.5710844993591309,
0.4722989499568939,
0.330960214138031,
-0.22632598876953125,
0.00015896186232566833,
-0.056308429688215256,
-0.37576544284820557,
-0.22712896764278412,
0.32718145847320557,
0.4236840605735779,
-0.08281472325325012,
-0.11036442220211029,
-0.07954616844654083,
0.10516591370105743,
0.559253990650177,
0.4052501320838928,
-0.17483340203762054,
-0.07906554639339447,
0.020856671035289764,
0.15972900390625,
0.16132764518260956,
-0.2096678912639618,
-0.25260594487190247,
0.4820108711719513,
-0.24641740322113037,
0.2487398236989975,
0.35342103242874146,
0.1166631281375885,
0.617304801940918,
-0.17541232705116272,
-0.07058148086071014,
0.6404229402542114,
-0.4585624933242798,
0.08418096601963043,
-0.38688284158706665,
-0.005343355238437653,
-0.10611443221569061,
0.25615841150283813,
0.16894984245300293,
0.12541131675243378,
0.24003133177757263,
0.2963806092739105,
0.3235984146595001,
0.08058828860521317,
0.35313987731933594,
-0.06989148259162903,
0.2761307954788208,
-0.2106955647468567,
0.049497976899147034,
-0.11211924254894257,
-0.6091260313987732,
0.28869563341140747,
0.07801954448223114,
-0.27191469073295593,
0.0680549144744873,
-0.23662494122982025,
0.44753289222717285,
-0.15206006169319153,
-0.0822150856256485,
-0.23689866065979004,
-0.03141500800848007,
0.26023662090301514,
0.2944309711456299,
-0.10916776955127716,
-0.3123038113117218,
0.2191813886165619,
-0.42277610301971436,
-0.02126145362854004,
-0.26497286558151245,
0.5215786099433899,
0.21814259886741638,
-0.21846994757652283,
-0.0010435879230499268,
0.045710932463407516,
0.14224129915237427,
-0.15118201076984406,
0.015415862202644348,
0.33918988704681396,
0.2921460270881653,
0.18274448812007904,
-0.08908434212207794,
0.1525992751121521,
0.40113458037376404,
-0.08527597039937973,
0.06607906520366669,
0.24203136563301086,
0.003439906984567642,
-0.3257346451282501,
-0.27416977286338806,
0.18110360205173492,
-0.15961357951164246,
0.3226219117641449,
0.12815937399864197,
0.05069836974143982,
0.2503218650817871,
-0.04333620145916939,
-0.1160278394818306,
0.07843576371669769,
0.10563630610704422,
-0.3318949341773987,
0.23888325691223145,
-0.10024002194404602,
-0.5517885088920593,
0.38689136505126953,
-0.007618499919772148,
0.06681092083454132,
-0.025056421756744385,
-0.183919757604599,
-0.0696144700050354,
-0.13988539576530457,
0.47877833247184753,
0.5028660297393799,
0.47347626090049744,
0.17877982556819916,
0.3006388545036316,
-0.029582621529698372,
0.09126587957143784,
-0.3178114593029022,
-0.0831855833530426,
0.04881472885608673,
0.23611204326152802,
0.2955628037452698,
-0.11971695721149445,
-0.04904571920633316,
-0.1970173865556717,
-0.23455947637557983,
0.3719569742679596,
0.05490919202566147,
-0.1710328459739685,
0.06568951159715652,
-0.31892722845077515,
-0.4927380681037903,
-0.22923643887043,
-0.045837175101041794,
-0.05014248937368393,
-0.17293314635753632,
0.06627776473760605,
0.7270770072937012,
-0.06778944283723831,
0.28438833355903625,
0.47979599237442017,
0.4485873579978943,
-0.21059271693229675,
0.1654486060142517,
-0.14575327932834625,
-0.18223978579044342,
0.0032022353261709213,
0.0369364470243454,
0.22898724675178528,
0.09430357813835144,
0.33727672696113586,
0.28389203548431396,
-0.16278596222400665,
-0.2981051802635193,
0.10762080550193787,
-0.09899180382490158,
0.00272422656416893,
-0.0963461697101593,
-0.17546817660331726,
-0.25050976872444153,
-0.03215603902935982,
0.003959033638238907,
-0.07086239755153656,
-0.24277536571025848,
-0.02813585475087166,
-0.013979624956846237,
-0.05706073343753815,
-0.2292422503232956,
-0.10658164322376251,
-0.008825637400150299,
-0.4428950250148773,
-0.33722180128097534,
0.03422681242227554,
-0.023857880383729935,
0.21627111732959747,
-0.106937475502491,
0.09920388460159302,
-0.317745178937912,
0.13228967785835266,
-0.2115560919046402,
0.004646945744752884,
-0.2248537242412567,
0.10413463413715363,
-0.20951318740844727,
-0.3363834023475647,
-0.18087321519851685,
0.016225799918174744,
-0.032758019864559174,
0.7001858353614807,
-0.3680460453033447,
-0.022672176361083984,
0.22942234575748444,
-0.28300631046295166,
0.13002845644950867,
-0.01745235174894333,
0.1841202676296234,
-0.21101859211921692,
0.0608733668923378,
-0.09832748770713806,
0.30754420161247253,
0.09404551982879639,
0.11913974583148956,
0.3478055000305176,
0.24098807573318481,
0.2969389855861664,
0.18660897016525269,
0.9571035504341125,
0.4824317395687103,
-0.04224644973874092,
0.14983172714710236,
0.19642145931720734,
-0.056599751114845276,
0.012742290273308754,
-0.0643312931060791,
0.3703346848487854,
-0.0830574780702591,
-0.045186202973127365,
0.12593674659729004,
-0.21250998973846436,
-0.34287744760513306,
-0.2323751598596573,
0.0019959434866905212,
0.11636404693126678,
-0.08302852511405945,
0.09244918823242188,
-0.2529964745044708,
0.17085915803909302,
0.15478980541229248,
0.08479215204715729,
0.20761753618717194,
-0.03211395442485809,
-0.4283945858478546,
0.06317974627017975,
0.27598223090171814,
0.016005147248506546,
-0.16827118396759033,
0.08855253458023071,
-0.33251258730888367,
0.20825934410095215,
0.11089914292097092,
0.41010433435440063,
-0.09692040830850601,
0.0012383460998535156,
0.20487654209136963,
0.16055065393447876,
0.2482411116361618,
-0.4635753929615021,
-0.15831811726093292,
-0.19165050983428955,
-0.3073832094669342,
0.14557787775993347,
-0.06610328704118729,
-0.13106635212898254,
0.39359205961227417,
-0.03613927215337753,
0.06572027504444122,
-0.19327493011951447,
0.00100698322057724,
-0.2515191435813904,
-0.15094858407974243,
-0.19483473896980286,
0.015966925770044327,
-0.22142896056175232,
-0.05043540522456169,
-0.26836690306663513,
0.20394307374954224,
0.11528261750936508,
-0.03895575553178787,
-0.2019885629415512,
-0.12639975547790527,
-0.17505629360675812,
0.09587056189775467,
-0.061167456209659576,
0.06695056706666946,
0.04935375973582268,
0.040436554700136185,
0.21243011951446533,
0.20874544978141785,
0.3953591585159302,
0.5721637606620789,
0.3576987087726593,
-0.10423833131790161,
-0.20544858276844025,
0.21369221806526184,
-0.012208551168441772,
0.35178130865097046,
0.4592912197113037,
0.059458423405885696,
0.12396785616874695,
0.199274480342865,
0.22486010193824768,
-0.3814557194709778,
-0.3206067383289337,
-0.04009976610541344,
-0.01633504405617714,
-0.3836514949798584,
-0.6445526480674744,
0.19172415137290955,
-0.08655966073274612,
-0.09299278259277344,
0.46199873089790344,
-0.20214125514030457,
-0.18130603432655334,
0.3067103624343872,
0.11880967020988464,
0.7116556763648987,
0.2004757821559906,
-0.1280091106891632,
-0.08082973957061768,
0.14721162617206573,
0.6211409568786621,
-0.12598812580108643,
0.30969175696372986,
-0.09895922243595123,
-0.34365129470825195,
-0.09388183802366257,
0.02876719832420349,
0.218648761510849,
0.3569571375846863,
-0.09297424554824829,
0.3380039632320404,
0.2797406017780304,
0.05031014606356621,
0.024454642087221146,
0.38943755626678467,
0.2260688841342926,
-0.29189732670783997,
0.05191393569111824,
0.040939681231975555,
-0.26360800862312317,
0.1095682680606842,
-0.029113762080669403,
0.06142812594771385,
0.30432945489883423,
-0.017380662262439728,
-0.3451733887195587,
-0.001589260995388031,
-0.2504502534866333,
0.1799796223640442,
-0.2980843186378479,
-0.3518190383911133,
-0.014309242367744446,
0.22457793354988098,
-0.051153820008039474,
-0.4006478190422058,
-0.060093075037002563,
0.1267896294593811,
0.09874902665615082,
0.1411825716495514,
-0.23552784323692322,
-0.2179218977689743,
0.28058183193206787,
-0.06922587007284164,
-0.07415220141410828,
0.032699450850486755,
0.1559431254863739,
-0.4639070928096771,
-0.21319936215877533,
-0.03263254463672638,
0.19737188518047333,
0.26600706577301025,
0.28038570284843445,
0.13937042653560638,
-0.07972268015146255,
-0.2499895989894867,
-0.001124952919781208,
-0.0405939482152462,
-0.15223562717437744,
0.531467080116272,
-0.15184012055397034,
0.0904516875743866,
-0.20528092980384827,
0.3922710418701172,
0.1825878918170929,
0.1486111581325531,
-0.02756732702255249,
0.11244922876358032,
-0.1182924211025238,
-0.11030668020248413,
-0.09317553043365479,
-0.06052488088607788,
-0.11856216937303543,
0.008454520255327225,
-0.29406440258026123,
0.19962066411972046,
0.27630722522735596,
-0.016298582777380943,
0.07249584794044495,
0.3035910129547119,
-0.25033047795295715,
-0.457568883895874,
0.10085532814264297,
0.08082135021686554,
-0.07574636489152908,
-0.10017263889312744,
0.08798982203006744,
-0.11817143857479095,
0.15497618913650513,
0.6689578890800476,
-0.2593654990196228,
-0.22264665365219116,
0.009598851203918457,
0.2004629671573639,
0.08540656417608261,
0.1396985650062561,
-0.08338335156440735,
-0.07921189069747925,
-0.12510380148887634,
0.07765056192874908,
0.024542182683944702,
-0.1204695850610733,
-0.0428435355424881,
0.19635778665542603,
0.21678051352500916,
-0.09578903019428253,
-0.23582208156585693,
-0.324596643447876,
-0.3343942165374756,
-0.10725420713424683,
0.20262300968170166,
0.012907840311527252,
-0.0461440235376358,
-0.08299568295478821,
0.6750946640968323,
0.08769296854734421,
-0.3021237552165985,
0.24845552444458008,
-0.0822640061378479,
-0.050996020436286926,
-0.17750069499015808,
0.19915303587913513,
0.3842018246650696,
-0.12150600552558899,
-0.17338357865810394,
-0.23282815515995026,
0.37631648778915405,
-0.08606995642185211,
0.4022093415260315,
-0.08369158208370209,
0.15544922649860382,
-0.16203370690345764,
0.5856527090072632,
0.23155951499938965,
-0.04679866507649422,
-0.20633524656295776,
0.28424015641212463,
0.10000159591436386,
-0.20904818177223206,
-0.20226672291755676,
0.051618073135614395,
0.33272844552993774,
0.1875542551279068,
0.4447084665298462,
0.39823317527770996,
-0.12155617773532867,
-0.2642560303211212,
0.09424786269664764,
0.5052286386489868,
-0.19930879771709442,
0.20498985052108765,
0.0005274559371173382,
0.13664285838603973,
0.19577041268348694,
-0.1093682050704956,
0.22564071416854858,
0.03534727543592453,
0.304769366979599,
-0.3092609941959381,
0.7399809956550598,
0.1800457090139389,
-0.04088064283132553,
0.26147162914276123,
-0.4083422124385834,
-0.027955103665590286,
0.0511932373046875,
-0.27629613876342773,
0.21890032291412354,
-0.11413562297821045,
0.05269940197467804,
-0.2647417187690735,
-0.22853687405586243,
-0.014623750001192093,
0.2962864339351654,
-0.002693724352866411,
0.031127575784921646,
-0.5655739307403564,
-0.008411988615989685,
-0.20485199987888336,
0.0021113306283950806,
0.028320545330643654,
-0.03827006742358208,
0.5511711835861206,
0.07702211290597916,
-0.041578423231840134,
-0.21449188888072968,
-0.06528092920780182,
0.05460263416171074,
0.14385336637496948,
-0.28721684217453003,
0.2275797724723816,
0.3027360439300537,
-0.13196931779384613,
0.07006557285785675,
0.1666833758354187,
0.07416398823261261,
0.26832813024520874,
-0.053863245993852615,
0.2339264303445816,
0.3050244450569153,
-0.06516563147306442,
-0.09334377199411392,
0.40177035331726074,
0.03326905518770218,
-0.15578347444534302,
0.2595016062259674,
-0.00023682788014411926,
0.017757251858711243,
0.2346586138010025,
0.2817247807979584,
-0.06593585014343262,
0.025959346443414688,
0.459255188703537,
-0.06756645441055298,
0.14546972513198853,
0.13626807928085327,
0.005246706306934357,
-0.37799328565597534,
-0.01038924977183342,
0.350979745388031,
-0.45905059576034546,
0.05125473812222481,
-0.022730836644768715,
0.004870064556598663,
-0.1468215137720108,
0.5160329937934875,
0.16500578820705414,
0.03988877683877945,
-0.2612239420413971,
-0.29604411125183105,
-0.5132619738578796,
0.08484984189271927,
-0.4117299020290375,
-0.16564467549324036,
-0.02849666215479374,
0.43281829357147217,
0.07081688940525055,
0.19639864563941956,
-0.3687134385108948,
0.08428403735160828,
-0.1745309680700302,
0.5773721933364868,
-0.2791088819503784,
-0.08937016129493713,
-0.5378432273864746,
0.15635201334953308,
0.09824883192777634,
-0.3539040684700012,
0.45212867856025696,
0.26110634207725525,
-0.1391412615776062,
0.07427337765693665,
0.005071297287940979,
0.26286131143569946,
0.005235630087554455,
-0.3676958382129669,
0.09250601381063461,
0.33780398964881897,
0.0506775826215744,
0.09396832436323166,
-0.0835827887058258,
0.12262168526649475,
-0.4526345729827881,
0.19286896288394928,
0.09724687039852142,
0.3620455265045166,
-0.19569164514541626,
-0.15890656411647797,
-0.29978421330451965,
0.35243454575538635,
0.03415100276470184,
-0.20737066864967346,
-0.27364835143089294,
0.08183158934116364,
-0.1954088807106018,
0.2924795150756836,
0.3380967974662781,
0.14995163679122925,
-0.23145905137062073,
-0.12317950278520584,
-0.31143924593925476,
-0.1804192215204239,
0.3792820870876312,
-0.47693243622779846,
0.045383088290691376,
-0.4637560248374939,
0.23137111961841583,
0.2531643509864807,
0.1970571130514145,
-0.25030016899108887,
-0.3180619478225708,
0.22695736587047577,
-0.32072004675865173,
0.057510972023010254,
0.07607784122228622,
0.12044914066791534,
-0.2699720561504364,
-0.1574166715145111,
-0.40562427043914795,
0.032632339745759964,
-0.5015792846679688,
0.22198708355426788,
-0.30460089445114136
] |
https://github.com/huggingface/datasets/issues/6623 | Yes both nodes will stream from the 1 shard, but each node will skip half of the examples. This way in total each example is seen once and exactly once during you distributed training.
Though it terms of I/O, the dataset is effectively read/streamed twice. | streaming datasets doesn't work properly with multi-node | ### Feature request
Let’s say I have a dataset with 5 samples with values [1, 2, 3, 4, 5], with 2 GPUs (for DDP) and batch size of 2. This dataset is an `IterableDataset` since I am streaming it.
Now I split the dataset using `split_dataset_by_node` to ensure it doesn’t get repeated. And since it’s already splitted, I don’t have to use `DistributedSampler` (also they don't work with iterable datasets anyway)?
But in this case I noticed that the:
First iteraton:
first GPU will get → [1, 2]
first GPU will get → [3, 4]
Second iteraton:
first GPU will get → [5]
first GPU will get → Nothing
which actually creates an issue since in case of `DistributedSampler`, the samples are repeated internally to ensure non of the GPUs at any iteration is missing any data for gradient sync.
So my questions are:
1. Here since splitting is happening before hand, how to make sure each GPU get’s a batch at each iteration to avoid gradient sync issues?
2. Do we need to use `DistributedSampler`? If yes, how?
3. in the docstrings of `split_dataset_by_node`, this is mentioned: *"If the dataset has a number of shards that is a factor of `world_size` (i.e. if `dataset.n_shards % world_size == 0`), then the shards are evenly assigned across the nodes, which is the most optimized. Otherwise, each node keeps 1 example out of `world_size`, skipping the other examples."* Can you explain the last part here?
4. If `dataset.n_shards % world_size != 0`, is it possible to shard the streaming dataset on the fly to avoid the case where data is missing?
### Motivation
Somehow streaming datasets should work with DDP since for big LLMs a lot of data is required and DDP/multi-node is mostly used to train such models and streaming can actually help solve the data part of it.
### Your contribution
Yes, I can help in submitting the PR once we get mutual understanding on how it should behave. | 45 | streaming datasets doesn't work properly with multi-node
### Feature request
Let’s say I have a dataset with 5 samples with values [1, 2, 3, 4, 5], with 2 GPUs (for DDP) and batch size of 2. This dataset is an `IterableDataset` since I am streaming it.
Now I split the dataset using `split_dataset_by_node` to ensure it doesn’t get repeated. And since it’s already splitted, I don’t have to use `DistributedSampler` (also they don't work with iterable datasets anyway)?
But in this case I noticed that the:
First iteraton:
first GPU will get → [1, 2]
first GPU will get → [3, 4]
Second iteraton:
first GPU will get → [5]
first GPU will get → Nothing
which actually creates an issue since in case of `DistributedSampler`, the samples are repeated internally to ensure non of the GPUs at any iteration is missing any data for gradient sync.
So my questions are:
1. Here since splitting is happening before hand, how to make sure each GPU get’s a batch at each iteration to avoid gradient sync issues?
2. Do we need to use `DistributedSampler`? If yes, how?
3. in the docstrings of `split_dataset_by_node`, this is mentioned: *"If the dataset has a number of shards that is a factor of `world_size` (i.e. if `dataset.n_shards % world_size == 0`), then the shards are evenly assigned across the nodes, which is the most optimized. Otherwise, each node keeps 1 example out of `world_size`, skipping the other examples."* Can you explain the last part here?
4. If `dataset.n_shards % world_size != 0`, is it possible to shard the streaming dataset on the fly to avoid the case where data is missing?
### Motivation
Somehow streaming datasets should work with DDP since for big LLMs a lot of data is required and DDP/multi-node is mostly used to train such models and streaming can actually help solve the data part of it.
### Your contribution
Yes, I can help in submitting the PR once we get mutual understanding on how it should behave.
Yes both nodes will stream from the 1 shard, but each node will skip half of the examples. This way in total each example is seen once and exactly once during you distributed training.
Though it terms of I/O, the dataset is effectively read/streamed twice. | [
-0.5885023474693298,
-0.2717367708683014,
-0.022129017859697342,
0.05329866707324982,
-0.05288459733128548,
-0.24468931555747986,
0.28833165764808655,
-0.16213496029376984,
-0.05456581711769104,
0.17409411072731018,
0.13106313347816467,
0.20888292789459229,
-0.21015878021717072,
0.1710551381111145,
-0.05287248641252518,
-0.2316754013299942,
0.07644330710172653,
0.01057659462094307,
0.04034364968538284,
-0.2694319486618042,
-0.07909917831420898,
-0.16820268332958221,
-0.22020316123962402,
-0.4776022434234619,
-0.23010557889938354,
-0.34483352303504944,
-0.1880214959383011,
0.17915897071361542,
0.04292941093444824,
-0.0007918775081634521,
-0.024187937378883362,
0.21599668264389038,
0.22667305171489716,
0.26554247736930847,
-0.00011949403415201232,
-0.17549662292003632,
0.14307358860969543,
-0.033987656235694885,
-0.2797178626060486,
-0.10833810269832611,
-0.4525478482246399,
0.12000128626823425,
-0.0856480672955513,
-0.15769720077514648,
-0.005377188324928284,
0.0469653457403183,
0.16009725630283356,
-0.5938501954078674,
0.4552592635154724,
-0.03413572907447815,
0.08512648195028305,
0.01408693939447403,
-0.2336328625679016,
0.11152130365371704,
-0.18026015162467957,
0.3543376922607422,
0.005941025912761688,
-0.023008089512586594,
0.3836512565612793,
-0.10668213665485382,
-0.2726716697216034,
0.3208426833152771,
-0.17977496981620789,
0.3480873107910156,
-0.009544376283884048,
-0.1916971206665039,
-0.13937494158744812,
-0.3502672016620636,
-0.16309091448783875,
0.19770018756389618,
0.24191229045391083,
-0.12461245805025101,
-0.39788925647735596,
-0.521531343460083,
-0.14842446148395538,
-0.3500770926475525,
-0.23801052570343018,
0.18176981806755066,
-0.2481730580329895,
-0.0703870952129364,
-0.4510071277618408,
-0.23377658426761627,
-0.039797209203243256,
0.11789821088314056,
0.15419495105743408,
0.7694438695907593,
0.16675050556659698,
0.23403680324554443,
-0.1394328773021698,
0.29044508934020996,
0.3079175055027008,
-0.2309572994709015,
0.09334338456392288,
-0.13752999901771545,
-0.4325198531150818,
0.018086589872837067,
0.05734461918473244,
0.04305759817361832,
0.16307535767555237,
0.47958433628082275,
0.30966776609420776,
0.039084240794181824,
0.2876565754413605,
0.16717436909675598,
0.39613404870033264,
-0.18216004967689514,
-0.07782210409641266,
0.13634347915649414,
0.2166709154844284,
-0.017559371888637543,
0.08888927102088928,
0.020971247926354408,
0.21538354456424713,
0.23815034329891205,
-0.24705252051353455,
0.18153789639472961,
0.21558597683906555,
0.06319627165794373,
-0.057585135102272034,
-0.33602985739707947,
-0.03557462990283966,
-0.1887940913438797,
0.13759249448776245,
0.04293625056743622,
0.03659946843981743,
0.24672368168830872,
-0.4028873145580292,
0.15768982470035553,
-0.16016718745231628,
-0.5503530502319336,
-0.21842163801193237,
-0.3020685017108917,
-0.15051349997520447,
0.36816567182540894,
0.14705823361873627,
-0.0330776646733284,
0.2690260112285614,
-0.02857164293527603,
0.10693713277578354,
0.1553390473127365,
-0.07142554968595505,
-0.27262142300605774,
0.024798860773444176,
0.23033371567726135,
0.18815961480140686,
0.09010680019855499,
0.036062948405742645,
0.08091837167739868,
-0.07104326784610748,
0.17591574788093567,
-0.01980457827448845,
-0.33053672313690186,
0.19016502797603607,
0.08777797967195511,
-0.019701950252056122,
-0.013415093533694744,
-0.5710844993591309,
0.4722989499568939,
0.330960214138031,
-0.22632598876953125,
0.00015896186232566833,
-0.056308429688215256,
-0.37576544284820557,
-0.22712896764278412,
0.32718145847320557,
0.4236840605735779,
-0.08281472325325012,
-0.11036442220211029,
-0.07954616844654083,
0.10516591370105743,
0.559253990650177,
0.4052501320838928,
-0.17483340203762054,
-0.07906554639339447,
0.020856671035289764,
0.15972900390625,
0.16132764518260956,
-0.2096678912639618,
-0.25260594487190247,
0.4820108711719513,
-0.24641740322113037,
0.2487398236989975,
0.35342103242874146,
0.1166631281375885,
0.617304801940918,
-0.17541232705116272,
-0.07058148086071014,
0.6404229402542114,
-0.4585624933242798,
0.08418096601963043,
-0.38688284158706665,
-0.005343355238437653,
-0.10611443221569061,
0.25615841150283813,
0.16894984245300293,
0.12541131675243378,
0.24003133177757263,
0.2963806092739105,
0.3235984146595001,
0.08058828860521317,
0.35313987731933594,
-0.06989148259162903,
0.2761307954788208,
-0.2106955647468567,
0.049497976899147034,
-0.11211924254894257,
-0.6091260313987732,
0.28869563341140747,
0.07801954448223114,
-0.27191469073295593,
0.0680549144744873,
-0.23662494122982025,
0.44753289222717285,
-0.15206006169319153,
-0.0822150856256485,
-0.23689866065979004,
-0.03141500800848007,
0.26023662090301514,
0.2944309711456299,
-0.10916776955127716,
-0.3123038113117218,
0.2191813886165619,
-0.42277610301971436,
-0.02126145362854004,
-0.26497286558151245,
0.5215786099433899,
0.21814259886741638,
-0.21846994757652283,
-0.0010435879230499268,
0.045710932463407516,
0.14224129915237427,
-0.15118201076984406,
0.015415862202644348,
0.33918988704681396,
0.2921460270881653,
0.18274448812007904,
-0.08908434212207794,
0.1525992751121521,
0.40113458037376404,
-0.08527597039937973,
0.06607906520366669,
0.24203136563301086,
0.003439906984567642,
-0.3257346451282501,
-0.27416977286338806,
0.18110360205173492,
-0.15961357951164246,
0.3226219117641449,
0.12815937399864197,
0.05069836974143982,
0.2503218650817871,
-0.04333620145916939,
-0.1160278394818306,
0.07843576371669769,
0.10563630610704422,
-0.3318949341773987,
0.23888325691223145,
-0.10024002194404602,
-0.5517885088920593,
0.38689136505126953,
-0.007618499919772148,
0.06681092083454132,
-0.025056421756744385,
-0.183919757604599,
-0.0696144700050354,
-0.13988539576530457,
0.47877833247184753,
0.5028660297393799,
0.47347626090049744,
0.17877982556819916,
0.3006388545036316,
-0.029582621529698372,
0.09126587957143784,
-0.3178114593029022,
-0.0831855833530426,
0.04881472885608673,
0.23611204326152802,
0.2955628037452698,
-0.11971695721149445,
-0.04904571920633316,
-0.1970173865556717,
-0.23455947637557983,
0.3719569742679596,
0.05490919202566147,
-0.1710328459739685,
0.06568951159715652,
-0.31892722845077515,
-0.4927380681037903,
-0.22923643887043,
-0.045837175101041794,
-0.05014248937368393,
-0.17293314635753632,
0.06627776473760605,
0.7270770072937012,
-0.06778944283723831,
0.28438833355903625,
0.47979599237442017,
0.4485873579978943,
-0.21059271693229675,
0.1654486060142517,
-0.14575327932834625,
-0.18223978579044342,
0.0032022353261709213,
0.0369364470243454,
0.22898724675178528,
0.09430357813835144,
0.33727672696113586,
0.28389203548431396,
-0.16278596222400665,
-0.2981051802635193,
0.10762080550193787,
-0.09899180382490158,
0.00272422656416893,
-0.0963461697101593,
-0.17546817660331726,
-0.25050976872444153,
-0.03215603902935982,
0.003959033638238907,
-0.07086239755153656,
-0.24277536571025848,
-0.02813585475087166,
-0.013979624956846237,
-0.05706073343753815,
-0.2292422503232956,
-0.10658164322376251,
-0.008825637400150299,
-0.4428950250148773,
-0.33722180128097534,
0.03422681242227554,
-0.023857880383729935,
0.21627111732959747,
-0.106937475502491,
0.09920388460159302,
-0.317745178937912,
0.13228967785835266,
-0.2115560919046402,
0.004646945744752884,
-0.2248537242412567,
0.10413463413715363,
-0.20951318740844727,
-0.3363834023475647,
-0.18087321519851685,
0.016225799918174744,
-0.032758019864559174,
0.7001858353614807,
-0.3680460453033447,
-0.022672176361083984,
0.22942234575748444,
-0.28300631046295166,
0.13002845644950867,
-0.01745235174894333,
0.1841202676296234,
-0.21101859211921692,
0.0608733668923378,
-0.09832748770713806,
0.30754420161247253,
0.09404551982879639,
0.11913974583148956,
0.3478055000305176,
0.24098807573318481,
0.2969389855861664,
0.18660897016525269,
0.9571035504341125,
0.4824317395687103,
-0.04224644973874092,
0.14983172714710236,
0.19642145931720734,
-0.056599751114845276,
0.012742290273308754,
-0.0643312931060791,
0.3703346848487854,
-0.0830574780702591,
-0.045186202973127365,
0.12593674659729004,
-0.21250998973846436,
-0.34287744760513306,
-0.2323751598596573,
0.0019959434866905212,
0.11636404693126678,
-0.08302852511405945,
0.09244918823242188,
-0.2529964745044708,
0.17085915803909302,
0.15478980541229248,
0.08479215204715729,
0.20761753618717194,
-0.03211395442485809,
-0.4283945858478546,
0.06317974627017975,
0.27598223090171814,
0.016005147248506546,
-0.16827118396759033,
0.08855253458023071,
-0.33251258730888367,
0.20825934410095215,
0.11089914292097092,
0.41010433435440063,
-0.09692040830850601,
0.0012383460998535156,
0.20487654209136963,
0.16055065393447876,
0.2482411116361618,
-0.4635753929615021,
-0.15831811726093292,
-0.19165050983428955,
-0.3073832094669342,
0.14557787775993347,
-0.06610328704118729,
-0.13106635212898254,
0.39359205961227417,
-0.03613927215337753,
0.06572027504444122,
-0.19327493011951447,
0.00100698322057724,
-0.2515191435813904,
-0.15094858407974243,
-0.19483473896980286,
0.015966925770044327,
-0.22142896056175232,
-0.05043540522456169,
-0.26836690306663513,
0.20394307374954224,
0.11528261750936508,
-0.03895575553178787,
-0.2019885629415512,
-0.12639975547790527,
-0.17505629360675812,
0.09587056189775467,
-0.061167456209659576,
0.06695056706666946,
0.04935375973582268,
0.040436554700136185,
0.21243011951446533,
0.20874544978141785,
0.3953591585159302,
0.5721637606620789,
0.3576987087726593,
-0.10423833131790161,
-0.20544858276844025,
0.21369221806526184,
-0.012208551168441772,
0.35178130865097046,
0.4592912197113037,
0.059458423405885696,
0.12396785616874695,
0.199274480342865,
0.22486010193824768,
-0.3814557194709778,
-0.3206067383289337,
-0.04009976610541344,
-0.01633504405617714,
-0.3836514949798584,
-0.6445526480674744,
0.19172415137290955,
-0.08655966073274612,
-0.09299278259277344,
0.46199873089790344,
-0.20214125514030457,
-0.18130603432655334,
0.3067103624343872,
0.11880967020988464,
0.7116556763648987,
0.2004757821559906,
-0.1280091106891632,
-0.08082973957061768,
0.14721162617206573,
0.6211409568786621,
-0.12598812580108643,
0.30969175696372986,
-0.09895922243595123,
-0.34365129470825195,
-0.09388183802366257,
0.02876719832420349,
0.218648761510849,
0.3569571375846863,
-0.09297424554824829,
0.3380039632320404,
0.2797406017780304,
0.05031014606356621,
0.024454642087221146,
0.38943755626678467,
0.2260688841342926,
-0.29189732670783997,
0.05191393569111824,
0.040939681231975555,
-0.26360800862312317,
0.1095682680606842,
-0.029113762080669403,
0.06142812594771385,
0.30432945489883423,
-0.017380662262439728,
-0.3451733887195587,
-0.001589260995388031,
-0.2504502534866333,
0.1799796223640442,
-0.2980843186378479,
-0.3518190383911133,
-0.014309242367744446,
0.22457793354988098,
-0.051153820008039474,
-0.4006478190422058,
-0.060093075037002563,
0.1267896294593811,
0.09874902665615082,
0.1411825716495514,
-0.23552784323692322,
-0.2179218977689743,
0.28058183193206787,
-0.06922587007284164,
-0.07415220141410828,
0.032699450850486755,
0.1559431254863739,
-0.4639070928096771,
-0.21319936215877533,
-0.03263254463672638,
0.19737188518047333,
0.26600706577301025,
0.28038570284843445,
0.13937042653560638,
-0.07972268015146255,
-0.2499895989894867,
-0.001124952919781208,
-0.0405939482152462,
-0.15223562717437744,
0.531467080116272,
-0.15184012055397034,
0.0904516875743866,
-0.20528092980384827,
0.3922710418701172,
0.1825878918170929,
0.1486111581325531,
-0.02756732702255249,
0.11244922876358032,
-0.1182924211025238,
-0.11030668020248413,
-0.09317553043365479,
-0.06052488088607788,
-0.11856216937303543,
0.008454520255327225,
-0.29406440258026123,
0.19962066411972046,
0.27630722522735596,
-0.016298582777380943,
0.07249584794044495,
0.3035910129547119,
-0.25033047795295715,
-0.457568883895874,
0.10085532814264297,
0.08082135021686554,
-0.07574636489152908,
-0.10017263889312744,
0.08798982203006744,
-0.11817143857479095,
0.15497618913650513,
0.6689578890800476,
-0.2593654990196228,
-0.22264665365219116,
0.009598851203918457,
0.2004629671573639,
0.08540656417608261,
0.1396985650062561,
-0.08338335156440735,
-0.07921189069747925,
-0.12510380148887634,
0.07765056192874908,
0.024542182683944702,
-0.1204695850610733,
-0.0428435355424881,
0.19635778665542603,
0.21678051352500916,
-0.09578903019428253,
-0.23582208156585693,
-0.324596643447876,
-0.3343942165374756,
-0.10725420713424683,
0.20262300968170166,
0.012907840311527252,
-0.0461440235376358,
-0.08299568295478821,
0.6750946640968323,
0.08769296854734421,
-0.3021237552165985,
0.24845552444458008,
-0.0822640061378479,
-0.050996020436286926,
-0.17750069499015808,
0.19915303587913513,
0.3842018246650696,
-0.12150600552558899,
-0.17338357865810394,
-0.23282815515995026,
0.37631648778915405,
-0.08606995642185211,
0.4022093415260315,
-0.08369158208370209,
0.15544922649860382,
-0.16203370690345764,
0.5856527090072632,
0.23155951499938965,
-0.04679866507649422,
-0.20633524656295776,
0.28424015641212463,
0.10000159591436386,
-0.20904818177223206,
-0.20226672291755676,
0.051618073135614395,
0.33272844552993774,
0.1875542551279068,
0.4447084665298462,
0.39823317527770996,
-0.12155617773532867,
-0.2642560303211212,
0.09424786269664764,
0.5052286386489868,
-0.19930879771709442,
0.20498985052108765,
0.0005274559371173382,
0.13664285838603973,
0.19577041268348694,
-0.1093682050704956,
0.22564071416854858,
0.03534727543592453,
0.304769366979599,
-0.3092609941959381,
0.7399809956550598,
0.1800457090139389,
-0.04088064283132553,
0.26147162914276123,
-0.4083422124385834,
-0.027955103665590286,
0.0511932373046875,
-0.27629613876342773,
0.21890032291412354,
-0.11413562297821045,
0.05269940197467804,
-0.2647417187690735,
-0.22853687405586243,
-0.014623750001192093,
0.2962864339351654,
-0.002693724352866411,
0.031127575784921646,
-0.5655739307403564,
-0.008411988615989685,
-0.20485199987888336,
0.0021113306283950806,
0.028320545330643654,
-0.03827006742358208,
0.5511711835861206,
0.07702211290597916,
-0.041578423231840134,
-0.21449188888072968,
-0.06528092920780182,
0.05460263416171074,
0.14385336637496948,
-0.28721684217453003,
0.2275797724723816,
0.3027360439300537,
-0.13196931779384613,
0.07006557285785675,
0.1666833758354187,
0.07416398823261261,
0.26832813024520874,
-0.053863245993852615,
0.2339264303445816,
0.3050244450569153,
-0.06516563147306442,
-0.09334377199411392,
0.40177035331726074,
0.03326905518770218,
-0.15578347444534302,
0.2595016062259674,
-0.00023682788014411926,
0.017757251858711243,
0.2346586138010025,
0.2817247807979584,
-0.06593585014343262,
0.025959346443414688,
0.459255188703537,
-0.06756645441055298,
0.14546972513198853,
0.13626807928085327,
0.005246706306934357,
-0.37799328565597534,
-0.01038924977183342,
0.350979745388031,
-0.45905059576034546,
0.05125473812222481,
-0.022730836644768715,
0.004870064556598663,
-0.1468215137720108,
0.5160329937934875,
0.16500578820705414,
0.03988877683877945,
-0.2612239420413971,
-0.29604411125183105,
-0.5132619738578796,
0.08484984189271927,
-0.4117299020290375,
-0.16564467549324036,
-0.02849666215479374,
0.43281829357147217,
0.07081688940525055,
0.19639864563941956,
-0.3687134385108948,
0.08428403735160828,
-0.1745309680700302,
0.5773721933364868,
-0.2791088819503784,
-0.08937016129493713,
-0.5378432273864746,
0.15635201334953308,
0.09824883192777634,
-0.3539040684700012,
0.45212867856025696,
0.26110634207725525,
-0.1391412615776062,
0.07427337765693665,
0.005071297287940979,
0.26286131143569946,
0.005235630087554455,
-0.3676958382129669,
0.09250601381063461,
0.33780398964881897,
0.0506775826215744,
0.09396832436323166,
-0.0835827887058258,
0.12262168526649475,
-0.4526345729827881,
0.19286896288394928,
0.09724687039852142,
0.3620455265045166,
-0.19569164514541626,
-0.15890656411647797,
-0.29978421330451965,
0.35243454575538635,
0.03415100276470184,
-0.20737066864967346,
-0.27364835143089294,
0.08183158934116364,
-0.1954088807106018,
0.2924795150756836,
0.3380967974662781,
0.14995163679122925,
-0.23145905137062073,
-0.12317950278520584,
-0.31143924593925476,
-0.1804192215204239,
0.3792820870876312,
-0.47693243622779846,
0.045383088290691376,
-0.4637560248374939,
0.23137111961841583,
0.2531643509864807,
0.1970571130514145,
-0.25030016899108887,
-0.3180619478225708,
0.22695736587047577,
-0.32072004675865173,
0.057510972023010254,
0.07607784122228622,
0.12044914066791534,
-0.2699720561504364,
-0.1574166715145111,
-0.40562427043914795,
0.032632339745759964,
-0.5015792846679688,
0.22198708355426788,
-0.30460089445114136
] |
https://github.com/huggingface/datasets/issues/6623 | what if the number of samples in that shard % num_nodes != 0? it will break/get stuck? or is the data repeated in that case for gradient sync? | streaming datasets doesn't work properly with multi-node | ### Feature request
Let’s say I have a dataset with 5 samples with values [1, 2, 3, 4, 5], with 2 GPUs (for DDP) and batch size of 2. This dataset is an `IterableDataset` since I am streaming it.
Now I split the dataset using `split_dataset_by_node` to ensure it doesn’t get repeated. And since it’s already splitted, I don’t have to use `DistributedSampler` (also they don't work with iterable datasets anyway)?
But in this case I noticed that the:
First iteraton:
first GPU will get → [1, 2]
first GPU will get → [3, 4]
Second iteraton:
first GPU will get → [5]
first GPU will get → Nothing
which actually creates an issue since in case of `DistributedSampler`, the samples are repeated internally to ensure non of the GPUs at any iteration is missing any data for gradient sync.
So my questions are:
1. Here since splitting is happening before hand, how to make sure each GPU get’s a batch at each iteration to avoid gradient sync issues?
2. Do we need to use `DistributedSampler`? If yes, how?
3. in the docstrings of `split_dataset_by_node`, this is mentioned: *"If the dataset has a number of shards that is a factor of `world_size` (i.e. if `dataset.n_shards % world_size == 0`), then the shards are evenly assigned across the nodes, which is the most optimized. Otherwise, each node keeps 1 example out of `world_size`, skipping the other examples."* Can you explain the last part here?
4. If `dataset.n_shards % world_size != 0`, is it possible to shard the streaming dataset on the fly to avoid the case where data is missing?
### Motivation
Somehow streaming datasets should work with DDP since for big LLMs a lot of data is required and DDP/multi-node is mostly used to train such models and streaming can actually help solve the data part of it.
### Your contribution
Yes, I can help in submitting the PR once we get mutual understanding on how it should behave. | 28 | streaming datasets doesn't work properly with multi-node
### Feature request
Let’s say I have a dataset with 5 samples with values [1, 2, 3, 4, 5], with 2 GPUs (for DDP) and batch size of 2. This dataset is an `IterableDataset` since I am streaming it.
Now I split the dataset using `split_dataset_by_node` to ensure it doesn’t get repeated. And since it’s already splitted, I don’t have to use `DistributedSampler` (also they don't work with iterable datasets anyway)?
But in this case I noticed that the:
First iteraton:
first GPU will get → [1, 2]
first GPU will get → [3, 4]
Second iteraton:
first GPU will get → [5]
first GPU will get → Nothing
which actually creates an issue since in case of `DistributedSampler`, the samples are repeated internally to ensure non of the GPUs at any iteration is missing any data for gradient sync.
So my questions are:
1. Here since splitting is happening before hand, how to make sure each GPU get’s a batch at each iteration to avoid gradient sync issues?
2. Do we need to use `DistributedSampler`? If yes, how?
3. in the docstrings of `split_dataset_by_node`, this is mentioned: *"If the dataset has a number of shards that is a factor of `world_size` (i.e. if `dataset.n_shards % world_size == 0`), then the shards are evenly assigned across the nodes, which is the most optimized. Otherwise, each node keeps 1 example out of `world_size`, skipping the other examples."* Can you explain the last part here?
4. If `dataset.n_shards % world_size != 0`, is it possible to shard the streaming dataset on the fly to avoid the case where data is missing?
### Motivation
Somehow streaming datasets should work with DDP since for big LLMs a lot of data is required and DDP/multi-node is mostly used to train such models and streaming can actually help solve the data part of it.
### Your contribution
Yes, I can help in submitting the PR once we get mutual understanding on how it should behave.
what if the number of samples in that shard % num_nodes != 0? it will break/get stuck? or is the data repeated in that case for gradient sync? | [
-0.5885023474693298,
-0.2717367708683014,
-0.022129017859697342,
0.05329866707324982,
-0.05288459733128548,
-0.24468931555747986,
0.28833165764808655,
-0.16213496029376984,
-0.05456581711769104,
0.17409411072731018,
0.13106313347816467,
0.20888292789459229,
-0.21015878021717072,
0.1710551381111145,
-0.05287248641252518,
-0.2316754013299942,
0.07644330710172653,
0.01057659462094307,
0.04034364968538284,
-0.2694319486618042,
-0.07909917831420898,
-0.16820268332958221,
-0.22020316123962402,
-0.4776022434234619,
-0.23010557889938354,
-0.34483352303504944,
-0.1880214959383011,
0.17915897071361542,
0.04292941093444824,
-0.0007918775081634521,
-0.024187937378883362,
0.21599668264389038,
0.22667305171489716,
0.26554247736930847,
-0.00011949403415201232,
-0.17549662292003632,
0.14307358860969543,
-0.033987656235694885,
-0.2797178626060486,
-0.10833810269832611,
-0.4525478482246399,
0.12000128626823425,
-0.0856480672955513,
-0.15769720077514648,
-0.005377188324928284,
0.0469653457403183,
0.16009725630283356,
-0.5938501954078674,
0.4552592635154724,
-0.03413572907447815,
0.08512648195028305,
0.01408693939447403,
-0.2336328625679016,
0.11152130365371704,
-0.18026015162467957,
0.3543376922607422,
0.005941025912761688,
-0.023008089512586594,
0.3836512565612793,
-0.10668213665485382,
-0.2726716697216034,
0.3208426833152771,
-0.17977496981620789,
0.3480873107910156,
-0.009544376283884048,
-0.1916971206665039,
-0.13937494158744812,
-0.3502672016620636,
-0.16309091448783875,
0.19770018756389618,
0.24191229045391083,
-0.12461245805025101,
-0.39788925647735596,
-0.521531343460083,
-0.14842446148395538,
-0.3500770926475525,
-0.23801052570343018,
0.18176981806755066,
-0.2481730580329895,
-0.0703870952129364,
-0.4510071277618408,
-0.23377658426761627,
-0.039797209203243256,
0.11789821088314056,
0.15419495105743408,
0.7694438695907593,
0.16675050556659698,
0.23403680324554443,
-0.1394328773021698,
0.29044508934020996,
0.3079175055027008,
-0.2309572994709015,
0.09334338456392288,
-0.13752999901771545,
-0.4325198531150818,
0.018086589872837067,
0.05734461918473244,
0.04305759817361832,
0.16307535767555237,
0.47958433628082275,
0.30966776609420776,
0.039084240794181824,
0.2876565754413605,
0.16717436909675598,
0.39613404870033264,
-0.18216004967689514,
-0.07782210409641266,
0.13634347915649414,
0.2166709154844284,
-0.017559371888637543,
0.08888927102088928,
0.020971247926354408,
0.21538354456424713,
0.23815034329891205,
-0.24705252051353455,
0.18153789639472961,
0.21558597683906555,
0.06319627165794373,
-0.057585135102272034,
-0.33602985739707947,
-0.03557462990283966,
-0.1887940913438797,
0.13759249448776245,
0.04293625056743622,
0.03659946843981743,
0.24672368168830872,
-0.4028873145580292,
0.15768982470035553,
-0.16016718745231628,
-0.5503530502319336,
-0.21842163801193237,
-0.3020685017108917,
-0.15051349997520447,
0.36816567182540894,
0.14705823361873627,
-0.0330776646733284,
0.2690260112285614,
-0.02857164293527603,
0.10693713277578354,
0.1553390473127365,
-0.07142554968595505,
-0.27262142300605774,
0.024798860773444176,
0.23033371567726135,
0.18815961480140686,
0.09010680019855499,
0.036062948405742645,
0.08091837167739868,
-0.07104326784610748,
0.17591574788093567,
-0.01980457827448845,
-0.33053672313690186,
0.19016502797603607,
0.08777797967195511,
-0.019701950252056122,
-0.013415093533694744,
-0.5710844993591309,
0.4722989499568939,
0.330960214138031,
-0.22632598876953125,
0.00015896186232566833,
-0.056308429688215256,
-0.37576544284820557,
-0.22712896764278412,
0.32718145847320557,
0.4236840605735779,
-0.08281472325325012,
-0.11036442220211029,
-0.07954616844654083,
0.10516591370105743,
0.559253990650177,
0.4052501320838928,
-0.17483340203762054,
-0.07906554639339447,
0.020856671035289764,
0.15972900390625,
0.16132764518260956,
-0.2096678912639618,
-0.25260594487190247,
0.4820108711719513,
-0.24641740322113037,
0.2487398236989975,
0.35342103242874146,
0.1166631281375885,
0.617304801940918,
-0.17541232705116272,
-0.07058148086071014,
0.6404229402542114,
-0.4585624933242798,
0.08418096601963043,
-0.38688284158706665,
-0.005343355238437653,
-0.10611443221569061,
0.25615841150283813,
0.16894984245300293,
0.12541131675243378,
0.24003133177757263,
0.2963806092739105,
0.3235984146595001,
0.08058828860521317,
0.35313987731933594,
-0.06989148259162903,
0.2761307954788208,
-0.2106955647468567,
0.049497976899147034,
-0.11211924254894257,
-0.6091260313987732,
0.28869563341140747,
0.07801954448223114,
-0.27191469073295593,
0.0680549144744873,
-0.23662494122982025,
0.44753289222717285,
-0.15206006169319153,
-0.0822150856256485,
-0.23689866065979004,
-0.03141500800848007,
0.26023662090301514,
0.2944309711456299,
-0.10916776955127716,
-0.3123038113117218,
0.2191813886165619,
-0.42277610301971436,
-0.02126145362854004,
-0.26497286558151245,
0.5215786099433899,
0.21814259886741638,
-0.21846994757652283,
-0.0010435879230499268,
0.045710932463407516,
0.14224129915237427,
-0.15118201076984406,
0.015415862202644348,
0.33918988704681396,
0.2921460270881653,
0.18274448812007904,
-0.08908434212207794,
0.1525992751121521,
0.40113458037376404,
-0.08527597039937973,
0.06607906520366669,
0.24203136563301086,
0.003439906984567642,
-0.3257346451282501,
-0.27416977286338806,
0.18110360205173492,
-0.15961357951164246,
0.3226219117641449,
0.12815937399864197,
0.05069836974143982,
0.2503218650817871,
-0.04333620145916939,
-0.1160278394818306,
0.07843576371669769,
0.10563630610704422,
-0.3318949341773987,
0.23888325691223145,
-0.10024002194404602,
-0.5517885088920593,
0.38689136505126953,
-0.007618499919772148,
0.06681092083454132,
-0.025056421756744385,
-0.183919757604599,
-0.0696144700050354,
-0.13988539576530457,
0.47877833247184753,
0.5028660297393799,
0.47347626090049744,
0.17877982556819916,
0.3006388545036316,
-0.029582621529698372,
0.09126587957143784,
-0.3178114593029022,
-0.0831855833530426,
0.04881472885608673,
0.23611204326152802,
0.2955628037452698,
-0.11971695721149445,
-0.04904571920633316,
-0.1970173865556717,
-0.23455947637557983,
0.3719569742679596,
0.05490919202566147,
-0.1710328459739685,
0.06568951159715652,
-0.31892722845077515,
-0.4927380681037903,
-0.22923643887043,
-0.045837175101041794,
-0.05014248937368393,
-0.17293314635753632,
0.06627776473760605,
0.7270770072937012,
-0.06778944283723831,
0.28438833355903625,
0.47979599237442017,
0.4485873579978943,
-0.21059271693229675,
0.1654486060142517,
-0.14575327932834625,
-0.18223978579044342,
0.0032022353261709213,
0.0369364470243454,
0.22898724675178528,
0.09430357813835144,
0.33727672696113586,
0.28389203548431396,
-0.16278596222400665,
-0.2981051802635193,
0.10762080550193787,
-0.09899180382490158,
0.00272422656416893,
-0.0963461697101593,
-0.17546817660331726,
-0.25050976872444153,
-0.03215603902935982,
0.003959033638238907,
-0.07086239755153656,
-0.24277536571025848,
-0.02813585475087166,
-0.013979624956846237,
-0.05706073343753815,
-0.2292422503232956,
-0.10658164322376251,
-0.008825637400150299,
-0.4428950250148773,
-0.33722180128097534,
0.03422681242227554,
-0.023857880383729935,
0.21627111732959747,
-0.106937475502491,
0.09920388460159302,
-0.317745178937912,
0.13228967785835266,
-0.2115560919046402,
0.004646945744752884,
-0.2248537242412567,
0.10413463413715363,
-0.20951318740844727,
-0.3363834023475647,
-0.18087321519851685,
0.016225799918174744,
-0.032758019864559174,
0.7001858353614807,
-0.3680460453033447,
-0.022672176361083984,
0.22942234575748444,
-0.28300631046295166,
0.13002845644950867,
-0.01745235174894333,
0.1841202676296234,
-0.21101859211921692,
0.0608733668923378,
-0.09832748770713806,
0.30754420161247253,
0.09404551982879639,
0.11913974583148956,
0.3478055000305176,
0.24098807573318481,
0.2969389855861664,
0.18660897016525269,
0.9571035504341125,
0.4824317395687103,
-0.04224644973874092,
0.14983172714710236,
0.19642145931720734,
-0.056599751114845276,
0.012742290273308754,
-0.0643312931060791,
0.3703346848487854,
-0.0830574780702591,
-0.045186202973127365,
0.12593674659729004,
-0.21250998973846436,
-0.34287744760513306,
-0.2323751598596573,
0.0019959434866905212,
0.11636404693126678,
-0.08302852511405945,
0.09244918823242188,
-0.2529964745044708,
0.17085915803909302,
0.15478980541229248,
0.08479215204715729,
0.20761753618717194,
-0.03211395442485809,
-0.4283945858478546,
0.06317974627017975,
0.27598223090171814,
0.016005147248506546,
-0.16827118396759033,
0.08855253458023071,
-0.33251258730888367,
0.20825934410095215,
0.11089914292097092,
0.41010433435440063,
-0.09692040830850601,
0.0012383460998535156,
0.20487654209136963,
0.16055065393447876,
0.2482411116361618,
-0.4635753929615021,
-0.15831811726093292,
-0.19165050983428955,
-0.3073832094669342,
0.14557787775993347,
-0.06610328704118729,
-0.13106635212898254,
0.39359205961227417,
-0.03613927215337753,
0.06572027504444122,
-0.19327493011951447,
0.00100698322057724,
-0.2515191435813904,
-0.15094858407974243,
-0.19483473896980286,
0.015966925770044327,
-0.22142896056175232,
-0.05043540522456169,
-0.26836690306663513,
0.20394307374954224,
0.11528261750936508,
-0.03895575553178787,
-0.2019885629415512,
-0.12639975547790527,
-0.17505629360675812,
0.09587056189775467,
-0.061167456209659576,
0.06695056706666946,
0.04935375973582268,
0.040436554700136185,
0.21243011951446533,
0.20874544978141785,
0.3953591585159302,
0.5721637606620789,
0.3576987087726593,
-0.10423833131790161,
-0.20544858276844025,
0.21369221806526184,
-0.012208551168441772,
0.35178130865097046,
0.4592912197113037,
0.059458423405885696,
0.12396785616874695,
0.199274480342865,
0.22486010193824768,
-0.3814557194709778,
-0.3206067383289337,
-0.04009976610541344,
-0.01633504405617714,
-0.3836514949798584,
-0.6445526480674744,
0.19172415137290955,
-0.08655966073274612,
-0.09299278259277344,
0.46199873089790344,
-0.20214125514030457,
-0.18130603432655334,
0.3067103624343872,
0.11880967020988464,
0.7116556763648987,
0.2004757821559906,
-0.1280091106891632,
-0.08082973957061768,
0.14721162617206573,
0.6211409568786621,
-0.12598812580108643,
0.30969175696372986,
-0.09895922243595123,
-0.34365129470825195,
-0.09388183802366257,
0.02876719832420349,
0.218648761510849,
0.3569571375846863,
-0.09297424554824829,
0.3380039632320404,
0.2797406017780304,
0.05031014606356621,
0.024454642087221146,
0.38943755626678467,
0.2260688841342926,
-0.29189732670783997,
0.05191393569111824,
0.040939681231975555,
-0.26360800862312317,
0.1095682680606842,
-0.029113762080669403,
0.06142812594771385,
0.30432945489883423,
-0.017380662262439728,
-0.3451733887195587,
-0.001589260995388031,
-0.2504502534866333,
0.1799796223640442,
-0.2980843186378479,
-0.3518190383911133,
-0.014309242367744446,
0.22457793354988098,
-0.051153820008039474,
-0.4006478190422058,
-0.060093075037002563,
0.1267896294593811,
0.09874902665615082,
0.1411825716495514,
-0.23552784323692322,
-0.2179218977689743,
0.28058183193206787,
-0.06922587007284164,
-0.07415220141410828,
0.032699450850486755,
0.1559431254863739,
-0.4639070928096771,
-0.21319936215877533,
-0.03263254463672638,
0.19737188518047333,
0.26600706577301025,
0.28038570284843445,
0.13937042653560638,
-0.07972268015146255,
-0.2499895989894867,
-0.001124952919781208,
-0.0405939482152462,
-0.15223562717437744,
0.531467080116272,
-0.15184012055397034,
0.0904516875743866,
-0.20528092980384827,
0.3922710418701172,
0.1825878918170929,
0.1486111581325531,
-0.02756732702255249,
0.11244922876358032,
-0.1182924211025238,
-0.11030668020248413,
-0.09317553043365479,
-0.06052488088607788,
-0.11856216937303543,
0.008454520255327225,
-0.29406440258026123,
0.19962066411972046,
0.27630722522735596,
-0.016298582777380943,
0.07249584794044495,
0.3035910129547119,
-0.25033047795295715,
-0.457568883895874,
0.10085532814264297,
0.08082135021686554,
-0.07574636489152908,
-0.10017263889312744,
0.08798982203006744,
-0.11817143857479095,
0.15497618913650513,
0.6689578890800476,
-0.2593654990196228,
-0.22264665365219116,
0.009598851203918457,
0.2004629671573639,
0.08540656417608261,
0.1396985650062561,
-0.08338335156440735,
-0.07921189069747925,
-0.12510380148887634,
0.07765056192874908,
0.024542182683944702,
-0.1204695850610733,
-0.0428435355424881,
0.19635778665542603,
0.21678051352500916,
-0.09578903019428253,
-0.23582208156585693,
-0.324596643447876,
-0.3343942165374756,
-0.10725420713424683,
0.20262300968170166,
0.012907840311527252,
-0.0461440235376358,
-0.08299568295478821,
0.6750946640968323,
0.08769296854734421,
-0.3021237552165985,
0.24845552444458008,
-0.0822640061378479,
-0.050996020436286926,
-0.17750069499015808,
0.19915303587913513,
0.3842018246650696,
-0.12150600552558899,
-0.17338357865810394,
-0.23282815515995026,
0.37631648778915405,
-0.08606995642185211,
0.4022093415260315,
-0.08369158208370209,
0.15544922649860382,
-0.16203370690345764,
0.5856527090072632,
0.23155951499938965,
-0.04679866507649422,
-0.20633524656295776,
0.28424015641212463,
0.10000159591436386,
-0.20904818177223206,
-0.20226672291755676,
0.051618073135614395,
0.33272844552993774,
0.1875542551279068,
0.4447084665298462,
0.39823317527770996,
-0.12155617773532867,
-0.2642560303211212,
0.09424786269664764,
0.5052286386489868,
-0.19930879771709442,
0.20498985052108765,
0.0005274559371173382,
0.13664285838603973,
0.19577041268348694,
-0.1093682050704956,
0.22564071416854858,
0.03534727543592453,
0.304769366979599,
-0.3092609941959381,
0.7399809956550598,
0.1800457090139389,
-0.04088064283132553,
0.26147162914276123,
-0.4083422124385834,
-0.027955103665590286,
0.0511932373046875,
-0.27629613876342773,
0.21890032291412354,
-0.11413562297821045,
0.05269940197467804,
-0.2647417187690735,
-0.22853687405586243,
-0.014623750001192093,
0.2962864339351654,
-0.002693724352866411,
0.031127575784921646,
-0.5655739307403564,
-0.008411988615989685,
-0.20485199987888336,
0.0021113306283950806,
0.028320545330643654,
-0.03827006742358208,
0.5511711835861206,
0.07702211290597916,
-0.041578423231840134,
-0.21449188888072968,
-0.06528092920780182,
0.05460263416171074,
0.14385336637496948,
-0.28721684217453003,
0.2275797724723816,
0.3027360439300537,
-0.13196931779384613,
0.07006557285785675,
0.1666833758354187,
0.07416398823261261,
0.26832813024520874,
-0.053863245993852615,
0.2339264303445816,
0.3050244450569153,
-0.06516563147306442,
-0.09334377199411392,
0.40177035331726074,
0.03326905518770218,
-0.15578347444534302,
0.2595016062259674,
-0.00023682788014411926,
0.017757251858711243,
0.2346586138010025,
0.2817247807979584,
-0.06593585014343262,
0.025959346443414688,
0.459255188703537,
-0.06756645441055298,
0.14546972513198853,
0.13626807928085327,
0.005246706306934357,
-0.37799328565597534,
-0.01038924977183342,
0.350979745388031,
-0.45905059576034546,
0.05125473812222481,
-0.022730836644768715,
0.004870064556598663,
-0.1468215137720108,
0.5160329937934875,
0.16500578820705414,
0.03988877683877945,
-0.2612239420413971,
-0.29604411125183105,
-0.5132619738578796,
0.08484984189271927,
-0.4117299020290375,
-0.16564467549324036,
-0.02849666215479374,
0.43281829357147217,
0.07081688940525055,
0.19639864563941956,
-0.3687134385108948,
0.08428403735160828,
-0.1745309680700302,
0.5773721933364868,
-0.2791088819503784,
-0.08937016129493713,
-0.5378432273864746,
0.15635201334953308,
0.09824883192777634,
-0.3539040684700012,
0.45212867856025696,
0.26110634207725525,
-0.1391412615776062,
0.07427337765693665,
0.005071297287940979,
0.26286131143569946,
0.005235630087554455,
-0.3676958382129669,
0.09250601381063461,
0.33780398964881897,
0.0506775826215744,
0.09396832436323166,
-0.0835827887058258,
0.12262168526649475,
-0.4526345729827881,
0.19286896288394928,
0.09724687039852142,
0.3620455265045166,
-0.19569164514541626,
-0.15890656411647797,
-0.29978421330451965,
0.35243454575538635,
0.03415100276470184,
-0.20737066864967346,
-0.27364835143089294,
0.08183158934116364,
-0.1954088807106018,
0.2924795150756836,
0.3380967974662781,
0.14995163679122925,
-0.23145905137062073,
-0.12317950278520584,
-0.31143924593925476,
-0.1804192215204239,
0.3792820870876312,
-0.47693243622779846,
0.045383088290691376,
-0.4637560248374939,
0.23137111961841583,
0.2531643509864807,
0.1970571130514145,
-0.25030016899108887,
-0.3180619478225708,
0.22695736587047577,
-0.32072004675865173,
0.057510972023010254,
0.07607784122228622,
0.12044914066791534,
-0.2699720561504364,
-0.1574166715145111,
-0.40562427043914795,
0.032632339745759964,
-0.5015792846679688,
0.22198708355426788,
-0.30460089445114136
] |
https://github.com/huggingface/datasets/issues/6623 | In the case one at least one of the noes will get an empty/incomplete batch. The data is not repeated in that case. If the training loop doesn't take this into account it can lead to unexpected behaviors indeed.
In the future we'd like to add a feature that would allow the nodes to ignore the last batch, this way all the nodes would only have full batches. | streaming datasets doesn't work properly with multi-node | ### Feature request
Let’s say I have a dataset with 5 samples with values [1, 2, 3, 4, 5], with 2 GPUs (for DDP) and batch size of 2. This dataset is an `IterableDataset` since I am streaming it.
Now I split the dataset using `split_dataset_by_node` to ensure it doesn’t get repeated. And since it’s already splitted, I don’t have to use `DistributedSampler` (also they don't work with iterable datasets anyway)?
But in this case I noticed that the:
First iteraton:
first GPU will get → [1, 2]
first GPU will get → [3, 4]
Second iteraton:
first GPU will get → [5]
first GPU will get → Nothing
which actually creates an issue since in case of `DistributedSampler`, the samples are repeated internally to ensure non of the GPUs at any iteration is missing any data for gradient sync.
So my questions are:
1. Here since splitting is happening before hand, how to make sure each GPU get’s a batch at each iteration to avoid gradient sync issues?
2. Do we need to use `DistributedSampler`? If yes, how?
3. in the docstrings of `split_dataset_by_node`, this is mentioned: *"If the dataset has a number of shards that is a factor of `world_size` (i.e. if `dataset.n_shards % world_size == 0`), then the shards are evenly assigned across the nodes, which is the most optimized. Otherwise, each node keeps 1 example out of `world_size`, skipping the other examples."* Can you explain the last part here?
4. If `dataset.n_shards % world_size != 0`, is it possible to shard the streaming dataset on the fly to avoid the case where data is missing?
### Motivation
Somehow streaming datasets should work with DDP since for big LLMs a lot of data is required and DDP/multi-node is mostly used to train such models and streaming can actually help solve the data part of it.
### Your contribution
Yes, I can help in submitting the PR once we get mutual understanding on how it should behave. | 68 | streaming datasets doesn't work properly with multi-node
### Feature request
Let’s say I have a dataset with 5 samples with values [1, 2, 3, 4, 5], with 2 GPUs (for DDP) and batch size of 2. This dataset is an `IterableDataset` since I am streaming it.
Now I split the dataset using `split_dataset_by_node` to ensure it doesn’t get repeated. And since it’s already splitted, I don’t have to use `DistributedSampler` (also they don't work with iterable datasets anyway)?
But in this case I noticed that the:
First iteraton:
first GPU will get → [1, 2]
first GPU will get → [3, 4]
Second iteraton:
first GPU will get → [5]
first GPU will get → Nothing
which actually creates an issue since in case of `DistributedSampler`, the samples are repeated internally to ensure non of the GPUs at any iteration is missing any data for gradient sync.
So my questions are:
1. Here since splitting is happening before hand, how to make sure each GPU get’s a batch at each iteration to avoid gradient sync issues?
2. Do we need to use `DistributedSampler`? If yes, how?
3. in the docstrings of `split_dataset_by_node`, this is mentioned: *"If the dataset has a number of shards that is a factor of `world_size` (i.e. if `dataset.n_shards % world_size == 0`), then the shards are evenly assigned across the nodes, which is the most optimized. Otherwise, each node keeps 1 example out of `world_size`, skipping the other examples."* Can you explain the last part here?
4. If `dataset.n_shards % world_size != 0`, is it possible to shard the streaming dataset on the fly to avoid the case where data is missing?
### Motivation
Somehow streaming datasets should work with DDP since for big LLMs a lot of data is required and DDP/multi-node is mostly used to train such models and streaming can actually help solve the data part of it.
### Your contribution
Yes, I can help in submitting the PR once we get mutual understanding on how it should behave.
In the case one at least one of the noes will get an empty/incomplete batch. The data is not repeated in that case. If the training loop doesn't take this into account it can lead to unexpected behaviors indeed.
In the future we'd like to add a feature that would allow the nodes to ignore the last batch, this way all the nodes would only have full batches. | [
-0.5885023474693298,
-0.2717367708683014,
-0.022129017859697342,
0.05329866707324982,
-0.05288459733128548,
-0.24468931555747986,
0.28833165764808655,
-0.16213496029376984,
-0.05456581711769104,
0.17409411072731018,
0.13106313347816467,
0.20888292789459229,
-0.21015878021717072,
0.1710551381111145,
-0.05287248641252518,
-0.2316754013299942,
0.07644330710172653,
0.01057659462094307,
0.04034364968538284,
-0.2694319486618042,
-0.07909917831420898,
-0.16820268332958221,
-0.22020316123962402,
-0.4776022434234619,
-0.23010557889938354,
-0.34483352303504944,
-0.1880214959383011,
0.17915897071361542,
0.04292941093444824,
-0.0007918775081634521,
-0.024187937378883362,
0.21599668264389038,
0.22667305171489716,
0.26554247736930847,
-0.00011949403415201232,
-0.17549662292003632,
0.14307358860969543,
-0.033987656235694885,
-0.2797178626060486,
-0.10833810269832611,
-0.4525478482246399,
0.12000128626823425,
-0.0856480672955513,
-0.15769720077514648,
-0.005377188324928284,
0.0469653457403183,
0.16009725630283356,
-0.5938501954078674,
0.4552592635154724,
-0.03413572907447815,
0.08512648195028305,
0.01408693939447403,
-0.2336328625679016,
0.11152130365371704,
-0.18026015162467957,
0.3543376922607422,
0.005941025912761688,
-0.023008089512586594,
0.3836512565612793,
-0.10668213665485382,
-0.2726716697216034,
0.3208426833152771,
-0.17977496981620789,
0.3480873107910156,
-0.009544376283884048,
-0.1916971206665039,
-0.13937494158744812,
-0.3502672016620636,
-0.16309091448783875,
0.19770018756389618,
0.24191229045391083,
-0.12461245805025101,
-0.39788925647735596,
-0.521531343460083,
-0.14842446148395538,
-0.3500770926475525,
-0.23801052570343018,
0.18176981806755066,
-0.2481730580329895,
-0.0703870952129364,
-0.4510071277618408,
-0.23377658426761627,
-0.039797209203243256,
0.11789821088314056,
0.15419495105743408,
0.7694438695907593,
0.16675050556659698,
0.23403680324554443,
-0.1394328773021698,
0.29044508934020996,
0.3079175055027008,
-0.2309572994709015,
0.09334338456392288,
-0.13752999901771545,
-0.4325198531150818,
0.018086589872837067,
0.05734461918473244,
0.04305759817361832,
0.16307535767555237,
0.47958433628082275,
0.30966776609420776,
0.039084240794181824,
0.2876565754413605,
0.16717436909675598,
0.39613404870033264,
-0.18216004967689514,
-0.07782210409641266,
0.13634347915649414,
0.2166709154844284,
-0.017559371888637543,
0.08888927102088928,
0.020971247926354408,
0.21538354456424713,
0.23815034329891205,
-0.24705252051353455,
0.18153789639472961,
0.21558597683906555,
0.06319627165794373,
-0.057585135102272034,
-0.33602985739707947,
-0.03557462990283966,
-0.1887940913438797,
0.13759249448776245,
0.04293625056743622,
0.03659946843981743,
0.24672368168830872,
-0.4028873145580292,
0.15768982470035553,
-0.16016718745231628,
-0.5503530502319336,
-0.21842163801193237,
-0.3020685017108917,
-0.15051349997520447,
0.36816567182540894,
0.14705823361873627,
-0.0330776646733284,
0.2690260112285614,
-0.02857164293527603,
0.10693713277578354,
0.1553390473127365,
-0.07142554968595505,
-0.27262142300605774,
0.024798860773444176,
0.23033371567726135,
0.18815961480140686,
0.09010680019855499,
0.036062948405742645,
0.08091837167739868,
-0.07104326784610748,
0.17591574788093567,
-0.01980457827448845,
-0.33053672313690186,
0.19016502797603607,
0.08777797967195511,
-0.019701950252056122,
-0.013415093533694744,
-0.5710844993591309,
0.4722989499568939,
0.330960214138031,
-0.22632598876953125,
0.00015896186232566833,
-0.056308429688215256,
-0.37576544284820557,
-0.22712896764278412,
0.32718145847320557,
0.4236840605735779,
-0.08281472325325012,
-0.11036442220211029,
-0.07954616844654083,
0.10516591370105743,
0.559253990650177,
0.4052501320838928,
-0.17483340203762054,
-0.07906554639339447,
0.020856671035289764,
0.15972900390625,
0.16132764518260956,
-0.2096678912639618,
-0.25260594487190247,
0.4820108711719513,
-0.24641740322113037,
0.2487398236989975,
0.35342103242874146,
0.1166631281375885,
0.617304801940918,
-0.17541232705116272,
-0.07058148086071014,
0.6404229402542114,
-0.4585624933242798,
0.08418096601963043,
-0.38688284158706665,
-0.005343355238437653,
-0.10611443221569061,
0.25615841150283813,
0.16894984245300293,
0.12541131675243378,
0.24003133177757263,
0.2963806092739105,
0.3235984146595001,
0.08058828860521317,
0.35313987731933594,
-0.06989148259162903,
0.2761307954788208,
-0.2106955647468567,
0.049497976899147034,
-0.11211924254894257,
-0.6091260313987732,
0.28869563341140747,
0.07801954448223114,
-0.27191469073295593,
0.0680549144744873,
-0.23662494122982025,
0.44753289222717285,
-0.15206006169319153,
-0.0822150856256485,
-0.23689866065979004,
-0.03141500800848007,
0.26023662090301514,
0.2944309711456299,
-0.10916776955127716,
-0.3123038113117218,
0.2191813886165619,
-0.42277610301971436,
-0.02126145362854004,
-0.26497286558151245,
0.5215786099433899,
0.21814259886741638,
-0.21846994757652283,
-0.0010435879230499268,
0.045710932463407516,
0.14224129915237427,
-0.15118201076984406,
0.015415862202644348,
0.33918988704681396,
0.2921460270881653,
0.18274448812007904,
-0.08908434212207794,
0.1525992751121521,
0.40113458037376404,
-0.08527597039937973,
0.06607906520366669,
0.24203136563301086,
0.003439906984567642,
-0.3257346451282501,
-0.27416977286338806,
0.18110360205173492,
-0.15961357951164246,
0.3226219117641449,
0.12815937399864197,
0.05069836974143982,
0.2503218650817871,
-0.04333620145916939,
-0.1160278394818306,
0.07843576371669769,
0.10563630610704422,
-0.3318949341773987,
0.23888325691223145,
-0.10024002194404602,
-0.5517885088920593,
0.38689136505126953,
-0.007618499919772148,
0.06681092083454132,
-0.025056421756744385,
-0.183919757604599,
-0.0696144700050354,
-0.13988539576530457,
0.47877833247184753,
0.5028660297393799,
0.47347626090049744,
0.17877982556819916,
0.3006388545036316,
-0.029582621529698372,
0.09126587957143784,
-0.3178114593029022,
-0.0831855833530426,
0.04881472885608673,
0.23611204326152802,
0.2955628037452698,
-0.11971695721149445,
-0.04904571920633316,
-0.1970173865556717,
-0.23455947637557983,
0.3719569742679596,
0.05490919202566147,
-0.1710328459739685,
0.06568951159715652,
-0.31892722845077515,
-0.4927380681037903,
-0.22923643887043,
-0.045837175101041794,
-0.05014248937368393,
-0.17293314635753632,
0.06627776473760605,
0.7270770072937012,
-0.06778944283723831,
0.28438833355903625,
0.47979599237442017,
0.4485873579978943,
-0.21059271693229675,
0.1654486060142517,
-0.14575327932834625,
-0.18223978579044342,
0.0032022353261709213,
0.0369364470243454,
0.22898724675178528,
0.09430357813835144,
0.33727672696113586,
0.28389203548431396,
-0.16278596222400665,
-0.2981051802635193,
0.10762080550193787,
-0.09899180382490158,
0.00272422656416893,
-0.0963461697101593,
-0.17546817660331726,
-0.25050976872444153,
-0.03215603902935982,
0.003959033638238907,
-0.07086239755153656,
-0.24277536571025848,
-0.02813585475087166,
-0.013979624956846237,
-0.05706073343753815,
-0.2292422503232956,
-0.10658164322376251,
-0.008825637400150299,
-0.4428950250148773,
-0.33722180128097534,
0.03422681242227554,
-0.023857880383729935,
0.21627111732959747,
-0.106937475502491,
0.09920388460159302,
-0.317745178937912,
0.13228967785835266,
-0.2115560919046402,
0.004646945744752884,
-0.2248537242412567,
0.10413463413715363,
-0.20951318740844727,
-0.3363834023475647,
-0.18087321519851685,
0.016225799918174744,
-0.032758019864559174,
0.7001858353614807,
-0.3680460453033447,
-0.022672176361083984,
0.22942234575748444,
-0.28300631046295166,
0.13002845644950867,
-0.01745235174894333,
0.1841202676296234,
-0.21101859211921692,
0.0608733668923378,
-0.09832748770713806,
0.30754420161247253,
0.09404551982879639,
0.11913974583148956,
0.3478055000305176,
0.24098807573318481,
0.2969389855861664,
0.18660897016525269,
0.9571035504341125,
0.4824317395687103,
-0.04224644973874092,
0.14983172714710236,
0.19642145931720734,
-0.056599751114845276,
0.012742290273308754,
-0.0643312931060791,
0.3703346848487854,
-0.0830574780702591,
-0.045186202973127365,
0.12593674659729004,
-0.21250998973846436,
-0.34287744760513306,
-0.2323751598596573,
0.0019959434866905212,
0.11636404693126678,
-0.08302852511405945,
0.09244918823242188,
-0.2529964745044708,
0.17085915803909302,
0.15478980541229248,
0.08479215204715729,
0.20761753618717194,
-0.03211395442485809,
-0.4283945858478546,
0.06317974627017975,
0.27598223090171814,
0.016005147248506546,
-0.16827118396759033,
0.08855253458023071,
-0.33251258730888367,
0.20825934410095215,
0.11089914292097092,
0.41010433435440063,
-0.09692040830850601,
0.0012383460998535156,
0.20487654209136963,
0.16055065393447876,
0.2482411116361618,
-0.4635753929615021,
-0.15831811726093292,
-0.19165050983428955,
-0.3073832094669342,
0.14557787775993347,
-0.06610328704118729,
-0.13106635212898254,
0.39359205961227417,
-0.03613927215337753,
0.06572027504444122,
-0.19327493011951447,
0.00100698322057724,
-0.2515191435813904,
-0.15094858407974243,
-0.19483473896980286,
0.015966925770044327,
-0.22142896056175232,
-0.05043540522456169,
-0.26836690306663513,
0.20394307374954224,
0.11528261750936508,
-0.03895575553178787,
-0.2019885629415512,
-0.12639975547790527,
-0.17505629360675812,
0.09587056189775467,
-0.061167456209659576,
0.06695056706666946,
0.04935375973582268,
0.040436554700136185,
0.21243011951446533,
0.20874544978141785,
0.3953591585159302,
0.5721637606620789,
0.3576987087726593,
-0.10423833131790161,
-0.20544858276844025,
0.21369221806526184,
-0.012208551168441772,
0.35178130865097046,
0.4592912197113037,
0.059458423405885696,
0.12396785616874695,
0.199274480342865,
0.22486010193824768,
-0.3814557194709778,
-0.3206067383289337,
-0.04009976610541344,
-0.01633504405617714,
-0.3836514949798584,
-0.6445526480674744,
0.19172415137290955,
-0.08655966073274612,
-0.09299278259277344,
0.46199873089790344,
-0.20214125514030457,
-0.18130603432655334,
0.3067103624343872,
0.11880967020988464,
0.7116556763648987,
0.2004757821559906,
-0.1280091106891632,
-0.08082973957061768,
0.14721162617206573,
0.6211409568786621,
-0.12598812580108643,
0.30969175696372986,
-0.09895922243595123,
-0.34365129470825195,
-0.09388183802366257,
0.02876719832420349,
0.218648761510849,
0.3569571375846863,
-0.09297424554824829,
0.3380039632320404,
0.2797406017780304,
0.05031014606356621,
0.024454642087221146,
0.38943755626678467,
0.2260688841342926,
-0.29189732670783997,
0.05191393569111824,
0.040939681231975555,
-0.26360800862312317,
0.1095682680606842,
-0.029113762080669403,
0.06142812594771385,
0.30432945489883423,
-0.017380662262439728,
-0.3451733887195587,
-0.001589260995388031,
-0.2504502534866333,
0.1799796223640442,
-0.2980843186378479,
-0.3518190383911133,
-0.014309242367744446,
0.22457793354988098,
-0.051153820008039474,
-0.4006478190422058,
-0.060093075037002563,
0.1267896294593811,
0.09874902665615082,
0.1411825716495514,
-0.23552784323692322,
-0.2179218977689743,
0.28058183193206787,
-0.06922587007284164,
-0.07415220141410828,
0.032699450850486755,
0.1559431254863739,
-0.4639070928096771,
-0.21319936215877533,
-0.03263254463672638,
0.19737188518047333,
0.26600706577301025,
0.28038570284843445,
0.13937042653560638,
-0.07972268015146255,
-0.2499895989894867,
-0.001124952919781208,
-0.0405939482152462,
-0.15223562717437744,
0.531467080116272,
-0.15184012055397034,
0.0904516875743866,
-0.20528092980384827,
0.3922710418701172,
0.1825878918170929,
0.1486111581325531,
-0.02756732702255249,
0.11244922876358032,
-0.1182924211025238,
-0.11030668020248413,
-0.09317553043365479,
-0.06052488088607788,
-0.11856216937303543,
0.008454520255327225,
-0.29406440258026123,
0.19962066411972046,
0.27630722522735596,
-0.016298582777380943,
0.07249584794044495,
0.3035910129547119,
-0.25033047795295715,
-0.457568883895874,
0.10085532814264297,
0.08082135021686554,
-0.07574636489152908,
-0.10017263889312744,
0.08798982203006744,
-0.11817143857479095,
0.15497618913650513,
0.6689578890800476,
-0.2593654990196228,
-0.22264665365219116,
0.009598851203918457,
0.2004629671573639,
0.08540656417608261,
0.1396985650062561,
-0.08338335156440735,
-0.07921189069747925,
-0.12510380148887634,
0.07765056192874908,
0.024542182683944702,
-0.1204695850610733,
-0.0428435355424881,
0.19635778665542603,
0.21678051352500916,
-0.09578903019428253,
-0.23582208156585693,
-0.324596643447876,
-0.3343942165374756,
-0.10725420713424683,
0.20262300968170166,
0.012907840311527252,
-0.0461440235376358,
-0.08299568295478821,
0.6750946640968323,
0.08769296854734421,
-0.3021237552165985,
0.24845552444458008,
-0.0822640061378479,
-0.050996020436286926,
-0.17750069499015808,
0.19915303587913513,
0.3842018246650696,
-0.12150600552558899,
-0.17338357865810394,
-0.23282815515995026,
0.37631648778915405,
-0.08606995642185211,
0.4022093415260315,
-0.08369158208370209,
0.15544922649860382,
-0.16203370690345764,
0.5856527090072632,
0.23155951499938965,
-0.04679866507649422,
-0.20633524656295776,
0.28424015641212463,
0.10000159591436386,
-0.20904818177223206,
-0.20226672291755676,
0.051618073135614395,
0.33272844552993774,
0.1875542551279068,
0.4447084665298462,
0.39823317527770996,
-0.12155617773532867,
-0.2642560303211212,
0.09424786269664764,
0.5052286386489868,
-0.19930879771709442,
0.20498985052108765,
0.0005274559371173382,
0.13664285838603973,
0.19577041268348694,
-0.1093682050704956,
0.22564071416854858,
0.03534727543592453,
0.304769366979599,
-0.3092609941959381,
0.7399809956550598,
0.1800457090139389,
-0.04088064283132553,
0.26147162914276123,
-0.4083422124385834,
-0.027955103665590286,
0.0511932373046875,
-0.27629613876342773,
0.21890032291412354,
-0.11413562297821045,
0.05269940197467804,
-0.2647417187690735,
-0.22853687405586243,
-0.014623750001192093,
0.2962864339351654,
-0.002693724352866411,
0.031127575784921646,
-0.5655739307403564,
-0.008411988615989685,
-0.20485199987888336,
0.0021113306283950806,
0.028320545330643654,
-0.03827006742358208,
0.5511711835861206,
0.07702211290597916,
-0.041578423231840134,
-0.21449188888072968,
-0.06528092920780182,
0.05460263416171074,
0.14385336637496948,
-0.28721684217453003,
0.2275797724723816,
0.3027360439300537,
-0.13196931779384613,
0.07006557285785675,
0.1666833758354187,
0.07416398823261261,
0.26832813024520874,
-0.053863245993852615,
0.2339264303445816,
0.3050244450569153,
-0.06516563147306442,
-0.09334377199411392,
0.40177035331726074,
0.03326905518770218,
-0.15578347444534302,
0.2595016062259674,
-0.00023682788014411926,
0.017757251858711243,
0.2346586138010025,
0.2817247807979584,
-0.06593585014343262,
0.025959346443414688,
0.459255188703537,
-0.06756645441055298,
0.14546972513198853,
0.13626807928085327,
0.005246706306934357,
-0.37799328565597534,
-0.01038924977183342,
0.350979745388031,
-0.45905059576034546,
0.05125473812222481,
-0.022730836644768715,
0.004870064556598663,
-0.1468215137720108,
0.5160329937934875,
0.16500578820705414,
0.03988877683877945,
-0.2612239420413971,
-0.29604411125183105,
-0.5132619738578796,
0.08484984189271927,
-0.4117299020290375,
-0.16564467549324036,
-0.02849666215479374,
0.43281829357147217,
0.07081688940525055,
0.19639864563941956,
-0.3687134385108948,
0.08428403735160828,
-0.1745309680700302,
0.5773721933364868,
-0.2791088819503784,
-0.08937016129493713,
-0.5378432273864746,
0.15635201334953308,
0.09824883192777634,
-0.3539040684700012,
0.45212867856025696,
0.26110634207725525,
-0.1391412615776062,
0.07427337765693665,
0.005071297287940979,
0.26286131143569946,
0.005235630087554455,
-0.3676958382129669,
0.09250601381063461,
0.33780398964881897,
0.0506775826215744,
0.09396832436323166,
-0.0835827887058258,
0.12262168526649475,
-0.4526345729827881,
0.19286896288394928,
0.09724687039852142,
0.3620455265045166,
-0.19569164514541626,
-0.15890656411647797,
-0.29978421330451965,
0.35243454575538635,
0.03415100276470184,
-0.20737066864967346,
-0.27364835143089294,
0.08183158934116364,
-0.1954088807106018,
0.2924795150756836,
0.3380967974662781,
0.14995163679122925,
-0.23145905137062073,
-0.12317950278520584,
-0.31143924593925476,
-0.1804192215204239,
0.3792820870876312,
-0.47693243622779846,
0.045383088290691376,
-0.4637560248374939,
0.23137111961841583,
0.2531643509864807,
0.1970571130514145,
-0.25030016899108887,
-0.3180619478225708,
0.22695736587047577,
-0.32072004675865173,
0.057510972023010254,
0.07607784122228622,
0.12044914066791534,
-0.2699720561504364,
-0.1574166715145111,
-0.40562427043914795,
0.032632339745759964,
-0.5015792846679688,
0.22198708355426788,
-0.30460089445114136
] |
https://github.com/huggingface/datasets/issues/6618 | Hi! Can you please share the error's stack trace so we can see where it comes from? | While importing load_dataset from datasets | ### Describe the bug
cannot import name 'DEFAULT_CIPHERS' from 'urllib3.util.ssl_' this is the error i received
### Steps to reproduce the bug
from datasets import load_dataset
### Expected behavior
No errors
### Environment info
python 3.11.5 | 17 | While importing load_dataset from datasets
### Describe the bug
cannot import name 'DEFAULT_CIPHERS' from 'urllib3.util.ssl_' this is the error i received
### Steps to reproduce the bug
from datasets import load_dataset
### Expected behavior
No errors
### Environment info
python 3.11.5
Hi! Can you please share the error's stack trace so we can see where it comes from? | [
0.02553679049015045,
-0.0960899293422699,
-0.01202813908457756,
0.30808398127555847,
0.20100738108158112,
0.12118995189666748,
0.3076993227005005,
-0.02027871645987034,
0.16383129358291626,
0.16387607157230377,
0.07356870919466019,
0.20870698988437653,
-0.0453374907374382,
0.22486478090286255,
-0.010323697701096535,
-0.08854030817747116,
0.03795933723449707,
-0.11755792796611786,
-0.1343417465686798,
-0.08287185430526733,
-0.26581519842147827,
0.28992894291877747,
-0.18643690645694733,
-0.03193062171339989,
0.16879205405712128,
0.014141449704766273,
-0.1925136148929596,
0.29297828674316406,
-0.30503565073013306,
-0.3261713683605194,
0.38788196444511414,
-0.1012846827507019,
0.3977859318256378,
0.08643285930156708,
-0.00011234589328523725,
-0.07913771271705627,
0.36974188685417175,
0.015161074697971344,
-0.03355153650045395,
-0.470453143119812,
-0.5806116461753845,
-0.43142029643058777,
0.1406441628932953,
-0.32072538137435913,
0.36564499139785767,
0.1480671912431717,
-0.14596295356750488,
-0.14254990220069885,
0.2450360655784607,
0.36443665623664856,
0.24022173881530762,
0.19046014547348022,
0.15952931344509125,
-0.10644888132810593,
0.33428871631622314,
-0.44106757640838623,
-0.1060420572757721,
0.3757067620754242,
-0.06672178208827972,
-0.05535488203167915,
0.3541308343410492,
0.24926592409610748,
-0.21048688888549805,
0.030058875679969788,
0.27987414598464966,
0.021131590008735657,
-0.014930907636880875,
-0.4703759253025055,
-0.1284697949886322,
0.2075527012348175,
0.38516291975975037,
-0.3053736686706543,
-0.032272983342409134,
0.06701798737049103,
0.1988714635372162,
-0.18061639368534088,
0.48838120698928833,
0.05424904078245163,
0.010762669146060944,
0.13253472745418549,
0.24273887276649475,
-0.1690889149904251,
-0.0841400995850563,
0.19292102754116058,
-0.20189189910888672,
0.4724831283092499,
-0.16227538883686066,
0.061334606260061264,
-0.065354123711586,
-0.24784313142299652,
0.4389994144439697,
-0.3556544780731201,
0.12223808467388153,
0.08670903742313385,
-0.23202630877494812,
0.14321982860565186,
0.14614872634410858,
0.11581660807132721,
-0.02202237769961357,
0.1160096526145935,
-0.07869714498519897,
0.021549362689256668,
-0.09361046552658081,
0.29374366998672485,
0.1633535474538803,
0.09156365692615509,
0.028090087696909904,
0.2239925116300583,
0.4188316762447357,
0.20877200365066528,
0.03118882328271866,
0.2086254209280014,
-0.01875334605574608,
-0.18117916584014893,
0.15382318198680878,
0.06474143266677856,
0.46892186999320984,
-0.09849834442138672,
-0.3385747969150543,
0.23375433683395386,
-0.04177612438797951,
-0.13675415515899658,
0.05483110249042511,
0.4141411781311035,
-0.1588600128889084,
0.18822985887527466,
0.08482615649700165,
0.04564640298485756,
-0.05759667232632637,
0.17443877458572388,
-0.16757142543792725,
-0.1858064830303192,
-0.1376136988401413,
-0.10949859768152237,
-0.0001993328332901001,
-0.18338394165039062,
0.145832359790802,
0.20561841130256653,
0.07790271937847137,
0.07166261970996857,
0.15232157707214355,
-0.0703311562538147,
-0.043130021542310715,
0.16476751863956451,
0.15672409534454346,
0.3111204504966736,
0.1860833466053009,
0.0002232864499092102,
-0.11120244115591049,
0.11449362337589264,
-0.7039685249328613,
-0.26508817076683044,
-0.19842146337032318,
0.21204417943954468,
-0.2765454649925232,
-0.19229993224143982,
-0.03251150995492935,
-0.1258443295955658,
-0.09724493324756622,
0.03746570646762848,
-0.32956838607788086,
-0.37241846323013306,
0.03843879699707031,
-0.2060108333826065,
0.24037885665893555,
0.16477572917938232,
-0.06738559156656265,
-0.16966889798641205,
-0.01306772232055664,
-0.2086106836795807,
0.2378397285938263,
0.03765609860420227,
-0.12443415820598602,
-0.22122246026992798,
-0.08149541914463043,
0.08899974822998047,
0.3034837245941162,
-0.2857157588005066,
-0.5515396595001221,
0.170106440782547,
0.09735492616891861,
-0.10354150086641312,
0.005366731435060501,
0.008233118802309036,
-0.0037278831005096436,
0.1105399951338768,
0.21718564629554749,
0.37016603350639343,
0.035913653671741486,
0.04334833472967148,
-0.041801389306783676,
-0.28849679231643677,
0.1526760309934616,
0.1679544448852539,
-0.04115274176001549,
-0.000554896891117096,
0.0013352669775485992,
-0.21094930171966553,
0.062196582555770874,
-0.1401839554309845,
-0.24210090935230255,
0.4701537489891052,
0.41093066334724426,
0.17404705286026,
0.2530912756919861,
0.10452811419963837,
-0.46245601773262024,
0.22233419120311737,
-0.01731663942337036,
0.20014724135398865,
-0.4171617031097412,
0.09273664653301239,
-0.3285067677497864,
0.08436769992113113,
-0.2936386168003082,
0.30734941363334656,
0.060402996838092804,
0.21983295679092407,
-0.31776756048202515,
0.0037487968802452087,
-0.23493114113807678,
0.03868138790130615,
-0.10754875093698502,
0.1017371267080307,
-0.3172740042209625,
0.44401589035987854,
-0.22986288368701935,
0.06304416060447693,
0.09637399017810822,
0.06909254938364029,
0.2709537744522095,
0.01788158342242241,
-0.292052298784256,
0.036968279629945755,
0.09752985835075378,
0.31266719102859497,
-0.15353235602378845,
-0.024366751313209534,
0.05274584889411926,
-0.5685095191001892,
0.15895748138427734,
0.2342654913663864,
0.1007973700761795,
0.061194755136966705,
0.35015758872032166,
0.28063470125198364,
-0.00516026746481657,
0.10783815383911133,
-0.12380313873291016,
0.08735813200473785,
0.2918430268764496,
-0.21040640771389008,
-0.10326173901557922,
-0.32125478982925415,
0.3325873613357544,
-0.04028424993157387,
0.09697955846786499,
-0.05868356674909592,
-0.13566268980503082,
-0.24745750427246094,
0.3271335959434509,
0.032567523419857025,
0.20735184848308563,
0.08520438522100449,
-0.011331640183925629,
-0.08867636322975159,
0.021434729918837547,
0.1905156522989273,
0.3946812152862549,
0.02285800129175186,
-0.09310115873813629,
0.3434097170829773,
0.03356029465794563,
-0.17750315368175507,
0.1573403775691986,
0.08903671056032181,
0.09929852187633514,
0.2100106030702591,
0.06358657777309418,
0.08367235213518143,
-0.22510580718517303,
-0.12658239901065826,
0.09842506051063538,
0.38376811146736145,
-0.41497671604156494,
0.21336553990840912,
-0.24509447813034058,
-0.26065146923065186,
0.03101949393749237,
-0.10516061633825302,
-0.12542366981506348,
-0.4427874684333801,
-0.2907395362854004,
0.018458833917975426,
-0.1781488060951233,
0.16627927124500275,
-0.058390527963638306,
0.037290215492248535,
0.22221453487873077,
-0.19413454830646515,
-0.2670815587043762,
-0.20541200041770935,
0.02568824589252472,
0.010299254208803177,
0.18086323142051697,
-0.1541556566953659,
0.38244643807411194,
-0.07234328985214233,
-0.1035650223493576,
-0.2071174532175064,
-0.36752068996429443,
0.07965318858623505,
0.012024980038404465,
0.2917082905769348,
0.447550505399704,
0.1187218427658081,
-0.05477599799633026,
-0.23145447671413422,
0.23283550143241882,
-0.03668242320418358,
-0.06612265855073929,
0.09331726282835007,
-0.16670581698417664,
-0.09558870643377304,
0.08349746465682983,
-0.3906209170818329,
-0.38127920031547546,
-0.27383506298065186,
-0.37455353140830994,
0.07367315888404846,
0.09302444756031036,
-0.10719664394855499,
-0.1559578776359558,
0.168332040309906,
0.14945662021636963,
0.10890140384435654,
-0.12337008863687515,
-0.2913774847984314,
0.39757710695266724,
-0.12663279473781586,
-0.410009503364563,
0.12785768508911133,
-0.12480725347995758,
0.19086246192455292,
-0.14185720682144165,
-0.5603805780410767,
-0.2046869397163391,
-0.2922917604446411,
0.23654061555862427,
-0.18254493176937103,
0.27309492230415344,
-0.12263210117816925,
-0.022450430318713188,
0.05519453436136246,
-0.20833942294120789,
-0.24277645349502563,
-0.28831276297569275,
0.023180227726697922,
0.23047509789466858,
0.2023102045059204,
0.6140865683555603,
-0.16907694935798645,
0.4722696542739868,
0.18029338121414185,
0.04849185049533844,
0.5941764712333679,
-0.03194278106093407,
0.1750902682542801,
0.0852806493639946,
-0.3505101501941681,
-0.17463424801826477,
0.028778210282325745,
0.10837754607200623,
0.208194300532341,
-0.1578506976366043,
0.04659174382686615,
-0.2692102789878845,
-0.14825987815856934,
0.0319950208067894,
-0.16385304927825928,
-0.08868938684463501,
-0.3130051791667938,
0.2694419324398041,
0.18875616788864136,
-0.03472227603197098,
-0.21305407583713531,
-0.1690765917301178,
0.044219087809324265,
0.20786404609680176,
-0.298356831073761,
0.01270950585603714,
-0.1603713035583496,
0.10816431790590286,
-0.42568233609199524,
0.3534225821495056,
0.09792344272136688,
0.5259202122688293,
-0.05136891454458237,
0.114077128469944,
0.07762209326028824,
-0.030013060197234154,
0.2806844413280487,
-0.2411973923444748,
0.023505859076976776,
0.19895149767398834,
-0.023517228662967682,
-0.46887660026550293,
-0.1088968813419342,
-0.22055523097515106,
-0.09264123439788818,
0.21160855889320374,
0.36250659823417664,
-0.27955129742622375,
0.11899307370185852,
0.1485779583454132,
0.2514075040817261,
-0.14247334003448486,
-0.2576497495174408,
-0.3717425763607025,
-0.5525254011154175,
-0.43312180042266846,
-0.09729410707950592,
-0.025235481560230255,
0.4001227021217346,
-0.03392941877245903,
0.09381294995546341,
0.004072990268468857,
-0.09178128838539124,
0.15717074275016785,
0.15633222460746765,
0.11591438949108124,
0.1413794755935669,
0.21654166281223297,
-0.04448956623673439,
-0.007740922272205353,
0.0193906482309103,
0.7926288843154907,
-0.22684241831302643,
-0.752051055431366,
0.03541440889239311,
0.17691637575626373,
0.06764912605285645,
0.056745368987321854,
-0.1391167789697647,
0.1583767682313919,
0.10856713354587555,
0.08371792733669281,
0.06939806044101715,
-0.022349802777171135,
0.27946561574935913,
0.08454032242298126,
-0.3639776408672333,
-0.3415968716144562,
0.3725995123386383,
-0.01877298206090927,
-0.10212233662605286,
0.6557703614234924,
0.007818667218089104,
-0.2670885920524597,
0.32141023874282837,
-0.2184034287929535,
0.7601138353347778,
0.13949176669120789,
0.1308879852294922,
0.32902154326438904,
-0.01617056131362915,
0.29306668043136597,
0.21962884068489075,
-0.21902452409267426,
-0.5138139128684998,
-0.1626586765050888,
0.04293408244848251,
-0.3877057731151581,
0.20296277105808258,
-0.2660449147224426,
-0.17098791897296906,
0.050633348524570465,
-0.12980441749095917,
0.3608745038509369,
0.2298247367143631,
0.21563223004341125,
-0.5483413934707642,
-0.015150494873523712,
-0.24506810307502747,
0.12755589187145233,
0.008852840401232243,
0.2002710998058319,
-0.23067934811115265,
0.049620337784290314,
0.14581741392612457,
-0.38610416650772095,
-0.04333392530679703,
0.3424680233001709,
-0.39683935046195984,
0.40160906314849854,
-0.055586449801921844,
-0.1538185179233551,
0.25431373715400696,
0.49903926253318787,
0.45388659834861755,
0.06058214232325554,
-0.27650073170661926,
0.2977127134799957,
-0.02539921924471855,
-0.004333019256591797,
0.0551142580807209,
0.08744567632675171,
0.1115647628903389,
-0.23576563596725464,
-0.16960029304027557,
-0.0009936895221471786,
0.0420696847140789,
-0.20326289534568787,
-0.033931538462638855,
0.18590210378170013,
0.3568348288536072,
-0.2725362181663513,
-0.3188658654689789,
-0.222861647605896,
0.29758763313293457,
-0.27748727798461914,
0.14368972182273865,
0.35668107867240906,
-0.13710632920265198,
0.3518802523612976,
-0.14317303895950317,
-0.2584018111228943,
0.014135591685771942,
0.47884976863861084,
-0.022925566881895065,
0.08739778399467468,
0.5913832187652588,
0.08617697656154633,
-0.062259286642074585,
-0.08214464038610458,
0.5878227949142456,
-0.11495646834373474,
-0.002090558409690857,
0.1880767047405243,
0.06586512178182602,
0.013117387890815735,
-0.061525601893663406,
0.3155566453933716,
0.06943336874246597,
0.0395694226026535,
0.14521344006061554,
-0.5375162363052368,
0.0061392709612846375,
0.18694359064102173,
0.08143475651741028,
0.07237565517425537,
0.1346597522497177,
-0.06980019807815552,
-0.228815957903862,
-0.41632699966430664,
-0.26560235023498535,
0.3345631957054138,
-0.1805942952632904,
0.08545296639204025,
0.3096538186073303,
0.11549876630306244,
0.06723625957965851,
-0.146728515625,
0.12154195457696915,
0.055145613849163055,
-0.041824426501989365,
-0.24200597405433655,
-0.40284061431884766,
0.09718911349773407,
0.18265514075756073,
-0.18150216341018677,
-0.01669454202055931,
-0.3710613250732422,
-0.02149149961769581,
0.03705817833542824,
0.1491749882698059,
0.25388485193252563,
0.044770631939172745,
0.3062055706977844,
-0.17457987368106842,
0.031749751418828964,
-0.23335519433021545,
0.28131985664367676,
-0.30885088443756104,
0.24844227731227875,
0.017061255872249603,
0.20253270864486694,
-0.0539546012878418,
0.0664324015378952,
-0.15804767608642578,
-0.09617718309164047,
-0.10917764157056808,
0.11650724709033966,
0.127533420920372,
-0.25002992153167725,
0.23031191527843475,
0.018334709107875824,
0.024464361369609833,
0.38603782653808594,
-0.6166849732398987,
0.08431268483400345,
0.2909267246723175,
0.2108510434627533,
-0.2309926301240921,
-0.14894576370716095,
-0.2835090756416321,
-0.04929504916071892,
-0.12954723834991455,
-0.1665658801794052,
0.09825330972671509,
0.017937511205673218,
-0.20924536883831024,
-0.11590873450040817,
0.3093053996562958,
0.20610260963439941,
0.5103486180305481,
0.4604107439517975,
-0.1999596655368805,
0.22079309821128845,
-0.006039422005414963,
-0.16184519231319427,
-0.12051212787628174,
0.46578022837638855,
-0.07719740271568298,
0.31568798422813416,
-0.00045314058661460876,
0.1795179396867752,
0.19262570142745972,
-0.5440988540649414,
0.24890415370464325,
-0.2068140208721161,
-0.14889080822467804,
-0.07643263041973114,
0.06448167562484741,
0.23883236944675446,
-0.24922755360603333,
0.09092073142528534,
-0.2246820032596588,
0.22443801164627075,
0.011106785386800766,
-0.1982215791940689,
-0.19329330325126648,
0.020183712244033813,
0.10265935957431793,
0.24963678419589996,
-0.13173076510429382,
-0.06405779719352722,
-0.21984347701072693,
0.2845182418823242,
-0.2986399233341217,
-0.4421229064464569,
0.23060156404972076,
0.1698843240737915,
-0.10468582808971405,
-0.21554255485534668,
0.2352343052625656,
0.11307738721370697,
-0.018512576818466187,
-0.07876655459403992,
0.0574258491396904,
0.4801429212093353,
0.23772750794887543,
0.12982773780822754,
0.009986285120248795,
0.14351913332939148,
-0.1271299421787262,
-0.08871768414974213,
0.11760953068733215,
0.24251362681388855,
0.5193853378295898,
0.3090222477912903,
0.15749908983707428,
-0.25264444947242737,
-0.3226427733898163,
0.30113211274147034,
0.022636067122220993,
-0.04213828593492508,
0.437865287065506,
-0.2764911651611328,
0.19283407926559448,
-0.1713656485080719,
0.19691793620586395,
-0.3575226664543152,
-0.08649910241365433,
0.554878830909729,
0.1537276804447174,
0.03971704840660095,
-0.1745000183582306,
0.06168470159173012,
-0.10925501585006714,
0.5514641404151917,
0.32812950015068054,
0.16909654438495636,
-0.2846253514289856,
-0.2328631728887558,
-0.8396716117858887,
0.3168759346008301,
0.14502321183681488,
-0.2916586697101593,
0.10789478570222855,
-0.2316846251487732,
-0.027788467705249786,
0.09198953211307526,
-0.06913568079471588,
0.17910915613174438,
-0.008718784898519516,
-0.12986159324645996,
-0.2974386215209961,
-0.3717603385448456,
0.06685040891170502,
0.1673821657896042,
-0.028477828949689865,
-0.35745424032211304,
0.12037373334169388,
-0.12687498331069946,
-0.0017765238881111145,
-0.388700008392334,
-0.36220186948776245,
-0.09259793907403946,
0.03938313201069832,
0.6595345735549927,
0.20601888000965118,
0.7294602990150452,
-0.027380533516407013,
0.07398427277803421,
0.06865066289901733,
-0.6393170952796936,
-0.0544687956571579,
0.21481209993362427,
0.007433081045746803,
0.09006096422672272,
-0.009896393865346909,
-0.20865051448345184,
-0.282707154750824,
0.24239249527454376,
-0.1323133409023285,
0.10101615637540817,
-0.31497225165367126,
0.19020810723304749,
0.161448135972023,
0.249250590801239,
-0.10973535478115082,
0.0990716814994812,
-0.012795727699995041,
0.23401901125907898,
-0.14983145892620087,
-0.33034569025039673,
0.521776556968689,
-0.20002710819244385,
-0.2910715341567993,
-0.15048156678676605,
0.20611584186553955,
-0.05463743954896927,
-0.14991602301597595,
-0.40658867359161377,
0.11120045185089111,
0.3926159143447876,
-0.14109447598457336,
-0.37175965309143066,
0.4308154881000519,
0.158900648355484,
0.29581135511398315,
-0.07537175714969635,
0.12054537236690521,
0.25460895895957947,
-0.37321382761001587,
0.054065585136413574,
-0.1200055480003357
] |
https://github.com/huggingface/datasets/issues/6612 | "Hi ! We recently updated `cnn_dailymail` and now `datasets>=2.14` is needed to load it.\r\n\r\nYou (...TRUNCATED) | cnn_dailymail repeats itself | "### Describe the bug\r\n\r\nWhen I try to load `cnn_dailymail` dataset, it takes longer than usual (...TRUNCATED) | 25 | "cnn_dailymail repeats itself \n ### Describe the bug\r\n\r\nWhen I try to load `cnn_dailymail` data(...TRUNCATED) | [0.07091876864433289,-0.30874955654144287,0.029564738273620605,0.6589643955230713,-0.008779346942901(...TRUNCATED) |
End of preview. Expand
in Dataset Viewer.
README.md exists but content is empty.
Use the Edit dataset card button to edit it.
- Downloads last month
- 0