id
int64 599M
3.29B
| url
stringlengths 58
61
| html_url
stringlengths 46
51
| number
int64 1
7.72k
| title
stringlengths 1
290
| state
stringclasses 2
values | comments
int64 0
70
| created_at
timestamp[s]date 2020-04-14 10:18:02
2025-08-05 09:28:51
| updated_at
timestamp[s]date 2020-04-27 16:04:17
2025-08-05 11:39:56
| closed_at
timestamp[s]date 2020-04-14 12:01:40
2025-08-01 05:15:45
⌀ | user_login
stringlengths 3
26
| labels
listlengths 0
4
| body
stringlengths 0
228k
⌀ | is_pull_request
bool 2
classes |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
1,214,510,010
|
https://api.github.com/repos/huggingface/datasets/issues/4213
|
https://github.com/huggingface/datasets/pull/4213
| 4,213
|
ETT time series dataset
|
closed
| 2
| 2022-04-25T13:26:18
| 2022-05-05T12:19:21
| 2022-05-05T12:10:35
|
kashif
|
[] |
Ready for review.
| true
|
1,214,498,582
|
https://api.github.com/repos/huggingface/datasets/issues/4212
|
https://github.com/huggingface/datasets/pull/4212
| 4,212
|
[Common Voice] Make sure bytes are correctly deleted if `path` exists
|
closed
| 2
| 2022-04-25T13:18:26
| 2022-04-26T22:54:28
| 2022-04-26T22:48:27
|
patrickvonplaten
|
[] |
`path` should be set to local path inside audio feature if exist so that bytes can correctly be deleted.
| true
|
1,214,361,837
|
https://api.github.com/repos/huggingface/datasets/issues/4211
|
https://github.com/huggingface/datasets/issues/4211
| 4,211
|
DatasetDict containing Datasets with different features when pushed to hub gets remapped features
|
closed
| 10
| 2022-04-25T11:22:54
| 2023-04-06T19:25:50
| 2022-05-20T15:15:30
|
pietrolesci
|
[
"bug"
] |
Hi there,
I am trying to load a dataset to the Hub. This dataset is a `DatasetDict` composed of various splits. Some splits have a different `Feature` mapping. Locally, the DatasetDict preserves the individual features but if I `push_to_hub` and then `load_dataset`, the features are all the same.
Dataset and code to reproduce available [here](https://huggingface.co/datasets/pietrolesci/robust_nli).
In short:
I have 3 feature mapping
```python
Tri_features = Features(
{
"idx": Value(dtype="int64"),
"premise": Value(dtype="string"),
"hypothesis": Value(dtype="string"),
"label": ClassLabel(num_classes=3, names=["entailment", "neutral", "contradiction"]),
}
)
Ent_features = Features(
{
"idx": Value(dtype="int64"),
"premise": Value(dtype="string"),
"hypothesis": Value(dtype="string"),
"label": ClassLabel(num_classes=2, names=["non-entailment", "entailment"]),
}
)
Con_features = Features(
{
"idx": Value(dtype="int64"),
"premise": Value(dtype="string"),
"hypothesis": Value(dtype="string"),
"label": ClassLabel(num_classes=2, names=["non-contradiction", "contradiction"]),
}
)
```
Then I create different datasets
```python
dataset_splits = {}
for split in df["split"].unique():
print(split)
df_split = df.loc[df["split"] == split].copy()
if split in Tri_dataset:
df_split["label"] = df_split["label"].map({"entailment": 0, "neutral": 1, "contradiction": 2})
ds = Dataset.from_pandas(df_split, features=Tri_features)
elif split in Ent_bin_dataset:
df_split["label"] = df_split["label"].map({"non-entailment": 0, "entailment": 1})
ds = Dataset.from_pandas(df_split, features=Ent_features)
elif split in Con_bin_dataset:
df_split["label"] = df_split["label"].map({"non-contradiction": 0, "contradiction": 1})
ds = Dataset.from_pandas(df_split, features=Con_features)
else:
print("ERROR:", split)
dataset_splits[split] = ds
datasets = DatasetDict(dataset_splits)
```
I then push to hub
```python
datasets.push_to_hub("pietrolesci/robust_nli", token="<token>")
```
Finally, I load it from the hub
```python
datasets_loaded_from_hub = load_dataset("pietrolesci/robust_nli")
```
And I get that
```python
datasets["LI_TS"].features != datasets_loaded_from_hub["LI_TS"].features
```
since
```python
"label": ClassLabel(num_classes=2, names=["non-contradiction", "contradiction"])
```
gets remapped to
```python
"label": ClassLabel(num_classes=3, names=["entailment", "neutral", "contradiction"])
```
| false
|
1,214,089,130
|
https://api.github.com/repos/huggingface/datasets/issues/4210
|
https://github.com/huggingface/datasets/issues/4210
| 4,210
|
TypeError: Cannot cast array data from dtype('O') to dtype('int64') according to the rule 'safe'
|
closed
| 5
| 2022-04-25T07:28:42
| 2022-05-31T12:16:31
| 2022-05-31T12:16:31
|
loretoparisi
|
[
"bug"
] |
### System Info
```shell
- `transformers` version: 4.18.0
- Platform: Linux-5.4.144+-x86_64-with-Ubuntu-18.04-bionic
- Python version: 3.7.13
- Huggingface_hub version: 0.5.1
- PyTorch version (GPU?): 1.10.0+cu111 (True)
- Tensorflow version (GPU?): 2.8.0 (True)
- Flax version (CPU?/GPU?/TPU?): not installed (NA)
- Jax version: not installed
- JaxLib version: not installed
- Using GPU in script?: <fill in>
- Using distributed or parallel set-up in script?: <fill in>
```
### Who can help?
@LysandreJik
### Information
- [ ] The official example scripts
- [X] My own modified scripts
### Tasks
- [ ] An officially supported task in the `examples` folder (such as GLUE/SQuAD, ...)
- [X] My own task or dataset (give details below)
### Reproduction
```python
from datasets import load_dataset,Features,Value,ClassLabel
class_names = ["cmn","deu","rus","fra","eng","jpn","spa","ita","kor","vie","nld","epo","por","tur","heb","hun","ell","ind","ara","arz","fin","bul","yue","swe","ukr","bel","que","ces","swh","nno","wuu","nob","zsm","est","kat","pol","lat","urd","sqi","isl","fry","afr","ron","fao","san","bre","tat","yid","uig","uzb","srp","qya","dan","pes","slk","eus","cycl","acm","tgl","lvs","kaz","hye","hin","lit","ben","cat","bos","hrv","tha","orv","cha","mon","lzh","scn","gle","mkd","slv","frm","glg","vol","ain","jbo","tok","ina","nds","mal","tlh","roh","ltz","oss","ido","gla","mlt","sco","ast","jav","oci","ile","ota","xal","tel","sjn","nov","khm","tpi","ang","aze","tgk","tuk","chv","hsb","dsb","bod","sme","cym","mri","ksh","kmr","ewe","kab","ber","tpw","udm","lld","pms","lad","grn","mlg","xho","pnb","grc","hat","lao","npi","cor","nah","avk","mar","guj","pan","kir","myv","prg","sux","crs","ckt","bak","zlm","hil","cbk","chr","nav","lkt","enm","arq","lin","abk","pcd","rom","gsw","tam","zul","awa","wln","amh","bar","hbo","mhr","bho","mrj","ckb","osx","pfl","mgm","sna","mah","hau","kan","nog","sin","glv","dng","kal","liv","vro","apc","jdt","fur","che","haw","yor","crh","pdc","ppl","kin","shs","mnw","tet","sah","kum","ngt","nya","pus","hif","mya","moh","wol","tir","ton","lzz","oar","lug","brx","non","mww","hak","nlv","ngu","bua","aym","vec","ibo","tkl","bam","kha","ceb","lou","fuc","smo","gag","lfn","arg","umb","tyv","kjh","oji","cyo","urh","kzj","pam","srd","lmo","swg","mdf","gil","snd","tso","sot","zza","tsn","pau","som","egl","ady","asm","ori","dtp","cho","max","kam","niu","sag","ilo","kaa","fuv","nch","hoc","iba","gbm","sun","war","mvv","pap","ary","kxi","csb","pag","cos","rif","kek","krc","aii","ban","ssw","tvl","mfe","tah","bvy","bcl","hnj","nau","nst","afb","quc","min","tmw","mad","bjn","mai","cjy","got","hsn","gan","tzl","dws","ldn","afh","sgs","krl","vep","rue","tly","mic","ext","izh","sma","jam","cmo","mwl","kpv","koi","bis","ike","run","evn","ryu","mnc","aoz","otk","kas","aln","akl","yua","shy","fkv","gos","fij","thv","zgh","gcf","cay","xmf","tig","div","lij","rap","hrx","cpi","tts","gaa","tmr","iii","ltg","bzt","syc","emx","gom","chg","osp","stq","frr","fro","nys","toi","new","phn","jpa","rel","drt","chn","pli","laa","bal","hdn","hax","mik","ajp","xqa","pal","crk","mni","lut","ayl","ood","sdh","ofs","nus","kiu","diq","qxq","alt","bfz","klj","mus","srn","guc","lim","zea","shi","mnr","bom","sat","szl"]
features = Features({ 'label': ClassLabel(names=class_names), 'text': Value('string')})
num_labels = features['label'].num_classes
data_files = { "train": "train.csv", "test": "test.csv" }
sentences = load_dataset("loretoparisi/tatoeba-sentences",
data_files=data_files,
delimiter='\t',
column_names=['label', 'text'],
features = features
```
ERROR:
```
ClassLabel(num_classes=403, names=['cmn', 'deu', 'rus', 'fra', 'eng', 'jpn', 'spa', 'ita', 'kor', 'vie', 'nld', 'epo', 'por', 'tur', 'heb', 'hun', 'ell', 'ind', 'ara', 'arz', 'fin', 'bul', 'yue', 'swe', 'ukr', 'bel', 'que', 'ces', 'swh', 'nno', 'wuu', 'nob', 'zsm', 'est', 'kat', 'pol', 'lat', 'urd', 'sqi', 'isl', 'fry', 'afr', 'ron', 'fao', 'san', 'bre', 'tat', 'yid', 'uig', 'uzb', 'srp', 'qya', 'dan', 'pes', 'slk', 'eus', 'cycl', 'acm', 'tgl', 'lvs', 'kaz', 'hye', 'hin', 'lit', 'ben', 'cat', 'bos', 'hrv', 'tha', 'orv', 'cha', 'mon', 'lzh', 'scn', 'gle', 'mkd', 'slv', 'frm', 'glg', 'vol', 'ain', 'jbo', 'tok', 'ina', 'nds', 'mal', 'tlh', 'roh', 'ltz', 'oss', 'ido', 'gla', 'mlt', 'sco', 'ast', 'jav', 'oci', 'ile', 'ota', 'xal', 'tel', 'sjn', 'nov', 'khm', 'tpi', 'ang', 'aze', 'tgk', 'tuk', 'chv', 'hsb', 'dsb', 'bod', 'sme', 'cym', 'mri', 'ksh', 'kmr', 'ewe', 'kab', 'ber', 'tpw', 'udm', 'lld', 'pms', 'lad', 'grn', 'mlg', 'xho', 'pnb', 'grc', 'hat', 'lao', 'npi', 'cor', 'nah', 'avk', 'mar', 'guj', 'pan', 'kir', 'myv', 'prg', 'sux', 'crs', 'ckt', 'bak', 'zlm', 'hil', 'cbk', 'chr', 'nav', 'lkt', 'enm', 'arq', 'lin', 'abk', 'pcd', 'rom', 'gsw', 'tam', 'zul', 'awa', 'wln', 'amh', 'bar', 'hbo', 'mhr', 'bho', 'mrj', 'ckb', 'osx', 'pfl', 'mgm', 'sna', 'mah', 'hau', 'kan', 'nog', 'sin', 'glv', 'dng', 'kal', 'liv', 'vro', 'apc', 'jdt', 'fur', 'che', 'haw', 'yor', 'crh', 'pdc', 'ppl', 'kin', 'shs', 'mnw', 'tet', 'sah', 'kum', 'ngt', 'nya', 'pus', 'hif', 'mya', 'moh', 'wol', 'tir', 'ton', 'lzz', 'oar', 'lug', 'brx', 'non', 'mww', 'hak', 'nlv', 'ngu', 'bua', 'aym', 'vec', 'ibo', 'tkl', 'bam', 'kha', 'ceb', 'lou', 'fuc', 'smo', 'gag', 'lfn', 'arg', 'umb', 'tyv', 'kjh', 'oji', 'cyo', 'urh', 'kzj', 'pam', 'srd', 'lmo', 'swg', 'mdf', 'gil', 'snd', 'tso', 'sot', 'zza', 'tsn', 'pau', 'som', 'egl', 'ady', 'asm', 'ori', 'dtp', 'cho', 'max', 'kam', 'niu', 'sag', 'ilo', 'kaa', 'fuv', 'nch', 'hoc', 'iba', 'gbm', 'sun', 'war', 'mvv', 'pap', 'ary', 'kxi', 'csb', 'pag', 'cos', 'rif', 'kek', 'krc', 'aii', 'ban', 'ssw', 'tvl', 'mfe', 'tah', 'bvy', 'bcl', 'hnj', 'nau', 'nst', 'afb', 'quc', 'min', 'tmw', 'mad', 'bjn', 'mai', 'cjy', 'got', 'hsn', 'gan', 'tzl', 'dws', 'ldn', 'afh', 'sgs', 'krl', 'vep', 'rue', 'tly', 'mic', 'ext', 'izh', 'sma', 'jam', 'cmo', 'mwl', 'kpv', 'koi', 'bis', 'ike', 'run', 'evn', 'ryu', 'mnc', 'aoz', 'otk', 'kas', 'aln', 'akl', 'yua', 'shy', 'fkv', 'gos', 'fij', 'thv', 'zgh', 'gcf', 'cay', 'xmf', 'tig', 'div', 'lij', 'rap', 'hrx', 'cpi', 'tts', 'gaa', 'tmr', 'iii', 'ltg', 'bzt', 'syc', 'emx', 'gom', 'chg', 'osp', 'stq', 'frr', 'fro', 'nys', 'toi', 'new', 'phn', 'jpa', 'rel', 'drt', 'chn', 'pli', 'laa', 'bal', 'hdn', 'hax', 'mik', 'ajp', 'xqa', 'pal', 'crk', 'mni', 'lut', 'ayl', 'ood', 'sdh', 'ofs', 'nus', 'kiu', 'diq', 'qxq', 'alt', 'bfz', 'klj', 'mus', 'srn', 'guc', 'lim', 'zea', 'shi', 'mnr', 'bom', 'sat', 'szl'], id=None)
Value(dtype='string', id=None)
Using custom data configuration loretoparisi--tatoeba-sentences-7b2c5e991f398f39
Downloading and preparing dataset csv/loretoparisi--tatoeba-sentences to /root/.cache/huggingface/datasets/csv/loretoparisi--tatoeba-sentences-7b2c5e991f398f39/0.0.0/433e0ccc46f9880962cc2b12065189766fbb2bee57a221866138fb9203c83519...
Downloading data files: 100%
2/2 [00:18<00:00, 8.06s/it]
Downloading data: 100%
391M/391M [00:13<00:00, 35.3MB/s]
Downloading data: 100%
92.4M/92.4M [00:02<00:00, 36.5MB/s]
Failed to read file '/root/.cache/huggingface/datasets/downloads/933132df9905194ea9faeb30cabca8c49318795612f6495fcb941a290191dd5d' with error <class 'ValueError'>: invalid literal for int() with base 10: 'cmn'
---------------------------------------------------------------------------
TypeError Traceback (most recent call last)
/usr/local/lib/python3.7/dist-packages/pandas/_libs/parsers.pyx in pandas._libs.parsers.TextReader._convert_tokens()
TypeError: Cannot cast array data from dtype('O') to dtype('int64') according to the rule 'safe'
During handling of the above exception, another exception occurred:
ValueError Traceback (most recent call last)
15 frames
/usr/local/lib/python3.7/dist-packages/pandas/_libs/parsers.pyx in pandas._libs.parsers.TextReader._convert_tokens()
ValueError: invalid literal for int() with base 10: 'cmn'
```
while loading without `features` it loads without errors
```
sentences = load_dataset("loretoparisi/tatoeba-sentences",
data_files=data_files,
delimiter='\t',
column_names=['label', 'text']
)
```
but the `label` col seems to be wrong (without the `ClassLabel` object):
```
sentences['train'].features
{'label': Value(dtype='string', id=None),
'text': Value(dtype='string', id=None)}
```
The dataset was https://huggingface.co/datasets/loretoparisi/tatoeba-sentences
Dataset format is:
```
ces Nechci vědět, co je tam uvnitř.
ces Kdo o tom chce slyšet?
deu Tom sagte, er fühle sich nicht wohl.
ber Mel-iyi-d anida-t tura ?
hun Gondom lesz rá rögtön.
ber Mel-iyi-d anida-tt tura ?
deu Ich will dich nicht reden hören.
```
### Expected behavior
```shell
correctly load train and test files.
```
| false
|
1,213,716,426
|
https://api.github.com/repos/huggingface/datasets/issues/4208
|
https://github.com/huggingface/datasets/pull/4208
| 4,208
|
Add CMU MoCap Dataset
|
closed
| 11
| 2022-04-24T17:31:08
| 2022-10-03T09:38:24
| 2022-10-03T09:36:30
|
dnaveenr
|
[
"dataset contribution"
] |
Resolves #3457
Dataset Request : Add CMU Graphics Lab Motion Capture dataset [#3457](https://github.com/huggingface/datasets/issues/3457)
This PR adds the CMU MoCap Dataset.
The authors didn't respond even after multiple follow ups, so I ended up crawling the website to get categories, subcategories and description information. Some of the subjects do not have category/subcategory/description as well. I am using a subject to categories, subcategories and description map (metadata file).
Currently the loading of the dataset works for "asf/amc" and "avi" formats since they have a single download link. But "c3d" and "mpg" have multiple download links (part archives) and dl_manager.download_and_extract() extracts the files to multiple paths, is there a way to extract these multiple archives into one folder ? Any other way to go about this ?
Any suggestions/inputs on this would be helpful. Thank you.
| true
|
1,213,604,615
|
https://api.github.com/repos/huggingface/datasets/issues/4207
|
https://github.com/huggingface/datasets/pull/4207
| 4,207
|
[Minor edit] Fix typo in class name
|
closed
| 0
| 2022-04-24T09:49:37
| 2022-05-05T13:17:47
| 2022-05-05T13:17:47
|
cakiki
|
[] |
Typo: `datasets.DatsetDict` -> `datasets.DatasetDict`
| true
|
1,212,715,581
|
https://api.github.com/repos/huggingface/datasets/issues/4206
|
https://github.com/huggingface/datasets/pull/4206
| 4,206
|
Add Nerval Metric
|
closed
| 1
| 2022-04-22T19:45:00
| 2023-07-11T09:34:56
| 2023-07-11T09:34:55
|
maridda
|
[
"transfer-to-evaluate"
] |
This PR adds readme.md and ner_val.py to metrics.
Nerval is a python package that helps evaluate NER models. It creates classification report and confusion matrix at entity level.
| true
|
1,212,466,138
|
https://api.github.com/repos/huggingface/datasets/issues/4205
|
https://github.com/huggingface/datasets/pull/4205
| 4,205
|
Fix `convert_file_size_to_int` for kilobits and megabits
|
closed
| 1
| 2022-04-22T14:56:21
| 2022-05-03T15:28:42
| 2022-05-03T15:21:48
|
mariosasko
|
[] |
Minor change to fully align this function with the recent change in Transformers (https://github.com/huggingface/transformers/pull/16891)
| true
|
1,212,431,764
|
https://api.github.com/repos/huggingface/datasets/issues/4204
|
https://github.com/huggingface/datasets/pull/4204
| 4,204
|
Add Recall Metric Card
|
closed
| 2
| 2022-04-22T14:24:26
| 2022-05-03T13:23:23
| 2022-05-03T13:16:24
|
emibaylor
|
[] |
What this PR mainly does:
- add metric card for recall metric
- update docs in recall python file
Note: I've also included a .json file with all of the metric card information. I've started compiling the relevant information in this type of .json files, and then using a script I wrote to generate the formatted metric card, as well as the docs to go in the .py file. I figured I'd upload the .json because it could be useful, especially if I also make a PR with the script I'm using (let me know if that's something you think would be beneficial!)
| true
|
1,212,431,067
|
https://api.github.com/repos/huggingface/datasets/issues/4203
|
https://github.com/huggingface/datasets/pull/4203
| 4,203
|
Add Precision Metric Card
|
closed
| 1
| 2022-04-22T14:23:48
| 2022-05-03T14:23:40
| 2022-05-03T14:16:46
|
emibaylor
|
[] |
What this PR mainly does:
- add metric card for precision metric
- update docs in precision python file
Note: I've also included a .json file with all of the metric card information. I've started compiling the relevant information in this type of .json files, and then using a script I wrote to generate the formatted metric card, as well as the docs to go in the .py file. I figured I'd upload the .json because it could be useful, especially if I also make a PR with the script I'm using (let me know if that's something you think would be beneficial!)
| true
|
1,212,326,288
|
https://api.github.com/repos/huggingface/datasets/issues/4202
|
https://github.com/huggingface/datasets/pull/4202
| 4,202
|
Fix some type annotation in doc
|
closed
| 1
| 2022-04-22T12:53:31
| 2022-04-22T15:03:00
| 2022-04-22T14:56:43
|
thomasw21
|
[] | null | true
|
1,212,086,420
|
https://api.github.com/repos/huggingface/datasets/issues/4201
|
https://github.com/huggingface/datasets/pull/4201
| 4,201
|
Update GH template for dataset viewer issues
|
closed
| 2
| 2022-04-22T09:34:44
| 2022-05-06T08:38:43
| 2022-04-26T08:45:55
|
albertvillanova
|
[] |
Update template to use new issue forms instead.
With this PR we can check if this new feature is useful for us.
Once validated, we can update the other templates.
CC: @severo
| true
|
1,211,980,110
|
https://api.github.com/repos/huggingface/datasets/issues/4200
|
https://github.com/huggingface/datasets/pull/4200
| 4,200
|
Add to docs how to load from local script
|
closed
| 1
| 2022-04-22T08:08:25
| 2022-05-06T08:39:25
| 2022-04-23T05:47:25
|
albertvillanova
|
[] |
This option was missing from the docs guide (it was only explained in the docstring of `load_dataset`). Although this is an infrequent use case, there might be some users interested in it.
Related to #4192
CC: @stevhliu
| true
|
1,211,953,308
|
https://api.github.com/repos/huggingface/datasets/issues/4199
|
https://github.com/huggingface/datasets/issues/4199
| 4,199
|
Cache miss during reload for datasets using image fetch utilities through map
|
closed
| 5
| 2022-04-22T07:47:08
| 2022-04-26T17:00:32
| 2022-04-26T13:38:26
|
apsdehal
|
[
"bug"
] |
## Describe the bug
It looks like that result of `.map` operation dataset are missing the cache when you reload the script and always run from scratch. In same interpretor session, they are able to find the cache and reload it. But, when you exit the interpretor and reload it, the downloading starts from scratch.
## Steps to reproduce the bug
Using the example provided in `red_caps` dataset.
```python
from concurrent.futures import ThreadPoolExecutor
from functools import partial
import io
import urllib
import PIL.Image
import datasets
from datasets import load_dataset
from datasets.utils.file_utils import get_datasets_user_agent
def fetch_single_image(image_url, timeout=None, retries=0):
for _ in range(retries + 1):
try:
request = urllib.request.Request(
image_url,
data=None,
headers={"user-agent": get_datasets_user_agent()},
)
with urllib.request.urlopen(request, timeout=timeout) as req:
image = PIL.Image.open(io.BytesIO(req.read()))
break
except Exception:
image = None
return image
def fetch_images(batch, num_threads, timeout=None, retries=0):
fetch_single_image_with_args = partial(fetch_single_image, timeout=timeout, retries=retries)
with ThreadPoolExecutor(max_workers=num_threads) as executor:
batch["image"] = list(executor.map(lambda image_urls: [fetch_single_image_with_args(image_url) for image_url in image_urls], batch["image_url"]))
return batch
def process_image_urls(batch):
processed_batch_image_urls = []
for image_url in batch["image_url"]:
processed_example_image_urls = []
image_url_splits = re.findall(r"http\S+", image_url)
for image_url_split in image_url_splits:
if "imgur" in image_url_split and "," in image_url_split:
for image_url_part in image_url_split.split(","):
if not image_url_part:
continue
image_url_part = image_url_part.strip()
root, ext = os.path.splitext(image_url_part)
if not root.startswith("http"):
root = "http://i.imgur.com/" + root
root = root.split("#")[0]
if not ext:
ext = ".jpg"
ext = re.split(r"[?%]", ext)[0]
image_url_part = root + ext
processed_example_image_urls.append(image_url_part)
else:
processed_example_image_urls.append(image_url_split)
processed_batch_image_urls.append(processed_example_image_urls)
batch["image_url"] = processed_batch_image_urls
return batch
dset = load_dataset("red_caps", "jellyfish")
dset = dset.map(process_image_urls, batched=True, num_proc=4)
features = dset["train"].features.copy()
features["image"] = datasets.Sequence(datasets.Image())
num_threads = 5
dset = dset.map(fetch_images, batched=True, batch_size=50, features=features, fn_kwargs={"num_threads": num_threads})
```
Run this in an interpretor or as a script twice and see that the cache is missed the second time.
## Expected results
At reload there should not be any cache miss
## Actual results
Every time script is run, cache is missed and dataset is built from scratch.
## Environment info
- `datasets` version: 2.1.1.dev0
- Platform: Linux-4.19.0-20-cloud-amd64-x86_64-with-glibc2.10
- Python version: 3.8.13
- PyArrow version: 7.0.0
- Pandas version: 1.4.1
| false
|
1,211,456,559
|
https://api.github.com/repos/huggingface/datasets/issues/4198
|
https://github.com/huggingface/datasets/issues/4198
| 4,198
|
There is no dataset
|
closed
| 0
| 2022-04-21T19:19:26
| 2022-05-03T11:29:05
| 2022-04-22T06:12:25
|
wilfoderek
|
[] |
## Dataset viewer issue for '*name of the dataset*'
**Link:** *link to the dataset viewer page*
*short description of the issue*
Am I the one who added this dataset ? Yes-No
| false
|
1,211,342,558
|
https://api.github.com/repos/huggingface/datasets/issues/4197
|
https://github.com/huggingface/datasets/pull/4197
| 4,197
|
Add remove_columns=True
|
closed
| 4
| 2022-04-21T17:28:13
| 2023-09-24T10:02:32
| 2022-04-22T14:45:30
|
thomasw21
|
[] |
This should fix all the issue we have with in place operations in mapping functions. This is crucial as where we do some weird things like:
```
def apply(batch):
batch_size = len(batch["id"])
batch["text"] = ["potato" for _ range(batch_size)]
return {}
# Columns are: {"id": int}
dset.map(apply, batched=True, remove_columns="text") # crashes because `text` is not in the original columns
dset.map(apply, batched=True) # mapped datasets has `text` column
```
In this PR we suggest to have `remove_columns=True` so that we ignore the input completely, and just use the output to generate mapped dataset. This means that inplace operations won't have any effects anymore.
| true
|
1,211,271,261
|
https://api.github.com/repos/huggingface/datasets/issues/4196
|
https://github.com/huggingface/datasets/issues/4196
| 4,196
|
Embed image and audio files in `save_to_disk`
|
closed
| 0
| 2022-04-21T16:25:18
| 2022-12-14T18:22:59
| 2022-12-14T18:22:59
|
lhoestq
|
[] |
Following https://github.com/huggingface/datasets/pull/4184, currently a dataset saved using `save_to_disk` doesn't actually contain the bytes of the image or audio files. Instead it stores the path to your local files.
Adding `embed_external_files` and set it to True by default to save_to_disk would be kind of a breaking change since some users will get bigger Arrow files when updating the lib, but the advantages are nice:
- the resulting dataset is self contained, in case you want to delete your cache for example or share it with someone else
- users also upload these Arrow files to cloud storage via the fs parameter, and in this case they would expect to upload a self-contained dataset
- consistency with push_to_hub
This can be implemented at the same time as sharding for `save_to_disk` for efficiency, and reuse the helpers from `push_to_hub` to embed the external files.
cc @mariosasko
| false
|
1,210,958,602
|
https://api.github.com/repos/huggingface/datasets/issues/4194
|
https://github.com/huggingface/datasets/pull/4194
| 4,194
|
Support lists of multi-dimensional numpy arrays
|
closed
| 1
| 2022-04-21T12:22:26
| 2022-05-12T15:16:34
| 2022-05-12T15:08:40
|
albertvillanova
|
[] |
Fix #4191.
CC: @SaulLu
| true
|
1,210,734,701
|
https://api.github.com/repos/huggingface/datasets/issues/4193
|
https://github.com/huggingface/datasets/pull/4193
| 4,193
|
Document save_to_disk and push_to_hub on images and audio files
|
closed
| 2
| 2022-04-21T09:04:36
| 2022-04-22T09:55:55
| 2022-04-22T09:49:31
|
lhoestq
|
[] |
Following https://github.com/huggingface/datasets/pull/4187, I explained in the documentation of `save_to_disk` and `push_to_hub` how they handle image and audio data.
| true
|
1,210,692,554
|
https://api.github.com/repos/huggingface/datasets/issues/4192
|
https://github.com/huggingface/datasets/issues/4192
| 4,192
|
load_dataset can't load local dataset,Unable to find ...
|
closed
| 4
| 2022-04-21T08:28:58
| 2022-04-25T16:51:57
| 2022-04-22T07:39:53
|
ahf876828330
|
[
"bug"
] |
Traceback (most recent call last):
File "/home/gs603/ahf/pretrained/model.py", line 48, in <module>
dataset = load_dataset("json",data_files="dataset/dataset_infos.json")
File "/home/gs603/miniconda3/envs/coderepair/lib/python3.7/site-packages/datasets/load.py", line 1675, in load_dataset
**config_kwargs,
File "/home/gs603/miniconda3/envs/coderepair/lib/python3.7/site-packages/datasets/load.py", line 1496, in load_dataset_builder
data_files=data_files,
File "/home/gs603/miniconda3/envs/coderepair/lib/python3.7/site-packages/datasets/load.py", line 1155, in dataset_module_factory
download_mode=download_mode,
File "/home/gs603/miniconda3/envs/coderepair/lib/python3.7/site-packages/datasets/load.py", line 800, in get_module
data_files = DataFilesDict.from_local_or_remote(patterns, use_auth_token=self.downnload_config.use_auth_token)
File "/home/gs603/miniconda3/envs/coderepair/lib/python3.7/site-packages/datasets/data_files.py", line 582, in from_local_or_remote
if not isinstance(patterns_for_key, DataFilesList)
File "/home/gs603/miniconda3/envs/coderepair/lib/python3.7/site-packages/datasets/data_files.py", line 544, in from_local_or_remote
data_files = resolve_patterns_locally_or_by_urls(base_path, patterns, allowed_extensions)
File "/home/gs603/miniconda3/envs/coderepair/lib/python3.7/site-packages/datasets/data_files.py", line 194, in resolve_patterns_locally_or_by_urls
for path in _resolve_single_pattern_locally(base_path, pattern, allowed_extensions):
File "/home/gs603/miniconda3/envs/coderepair/lib/python3.7/site-packages/datasets/data_files.py", line 144, in _resolve_single_pattern_locally
raise FileNotFoundError(error_msg)
FileNotFoundError: Unable to find '/home/gs603/ahf/pretrained/dataset/dataset_infos.json' at /home/gs603/ahf/pretrained
the code is in the model.py,why I can't use the load_dataset function to load my local dataset?
| false
|
1,210,028,090
|
https://api.github.com/repos/huggingface/datasets/issues/4191
|
https://github.com/huggingface/datasets/issues/4191
| 4,191
|
feat: create an `Array3D` column from a list of arrays of dimension 2
|
closed
| 2
| 2022-04-20T18:04:32
| 2022-05-12T15:08:40
| 2022-05-12T15:08:40
|
SaulLu
|
[
"enhancement"
] |
**Is your feature request related to a problem? Please describe.**
It is possible to create an `Array2D` column from a list of arrays of dimension 1. Similarly, I think it might be nice to be able to create a `Array3D` column from a list of lists of arrays of dimension 1.
To illustrate my proposal, let's take the following toy dataset t:
```python
import numpy as np
from datasets import Dataset, features
data_map = {
1: np.array([[0.2, 0,4],[0.19, 0,3]]),
2: np.array([[0.1, 0,4],[0.19, 0,3]]),
}
def create_toy_ds():
my_dict = {"id":[1, 2]}
return Dataset.from_dict(my_dict)
ds = create_toy_ds()
```
The following 2D processing works without any errors raised:
```python
def prepare_dataset_2D(batch):
batch["pixel_values"] = [data_map[index] for index in batch["id"]]
return batch
ds_2D = ds.map(
prepare_dataset_2D,
batched=True,
remove_columns=ds.column_names,
features=features.Features({"pixel_values": features.Array2D(shape=(2, 3), dtype="float32")})
)
```
The following 3D processing doesn't work:
```python
def prepare_dataset_3D(batch):
batch["pixel_values"] = [[data_map[index]] for index in batch["id"]]
return batch
ds_3D = ds.map(
prepare_dataset_3D,
batched=True,
remove_columns=ds.column_names,
features=features.Features({"pixel_values": features.Array3D(shape=(1, 2, 3, dtype="float32")})
)
```
The error raised is:
```
---------------------------------------------------------------------------
ArrowInvalid Traceback (most recent call last)
[<ipython-input-6-676547e4cd41>](https://localhost:8080/#) in <module>()
3 batched=True,
4 remove_columns=ds.column_names,
----> 5 features=features.Features({"pixel_values": features.Array3D(shape=(1, 2, 3), dtype="float32")})
6 )
12 frames
[/usr/local/lib/python3.7/dist-packages/datasets/arrow_dataset.py](https://localhost:8080/#) in map(self, function, with_indices, with_rank, input_columns, batched, batch_size, drop_last_batch, remove_columns, keep_in_memory, load_from_cache_file, cache_file_name, writer_batch_size, features, disable_nullable, fn_kwargs, num_proc, suffix_template, new_fingerprint, desc)
1971 new_fingerprint=new_fingerprint,
1972 disable_tqdm=disable_tqdm,
-> 1973 desc=desc,
1974 )
1975 else:
[/usr/local/lib/python3.7/dist-packages/datasets/arrow_dataset.py](https://localhost:8080/#) in wrapper(*args, **kwargs)
518 self: "Dataset" = kwargs.pop("self")
519 # apply actual function
--> 520 out: Union["Dataset", "DatasetDict"] = func(self, *args, **kwargs)
521 datasets: List["Dataset"] = list(out.values()) if isinstance(out, dict) else [out]
522 for dataset in datasets:
[/usr/local/lib/python3.7/dist-packages/datasets/arrow_dataset.py](https://localhost:8080/#) in wrapper(*args, **kwargs)
485 }
486 # apply actual function
--> 487 out: Union["Dataset", "DatasetDict"] = func(self, *args, **kwargs)
488 datasets: List["Dataset"] = list(out.values()) if isinstance(out, dict) else [out]
489 # re-apply format to the output
[/usr/local/lib/python3.7/dist-packages/datasets/fingerprint.py](https://localhost:8080/#) in wrapper(*args, **kwargs)
456 # Call actual function
457
--> 458 out = func(self, *args, **kwargs)
459
460 # Update fingerprint of in-place transforms + update in-place history of transforms
[/usr/local/lib/python3.7/dist-packages/datasets/arrow_dataset.py](https://localhost:8080/#) in _map_single(self, function, with_indices, with_rank, input_columns, batched, batch_size, drop_last_batch, remove_columns, keep_in_memory, load_from_cache_file, cache_file_name, writer_batch_size, features, disable_nullable, fn_kwargs, new_fingerprint, rank, offset, disable_tqdm, desc, cache_only)
2354 writer.write_table(batch)
2355 else:
-> 2356 writer.write_batch(batch)
2357 if update_data and writer is not None:
2358 writer.finalize() # close_stream=bool(buf_writer is None)) # We only close if we are writing in a file
[/usr/local/lib/python3.7/dist-packages/datasets/arrow_writer.py](https://localhost:8080/#) in write_batch(self, batch_examples, writer_batch_size)
505 col_try_type = try_features[col] if try_features is not None and col in try_features else None
506 typed_sequence = OptimizedTypedSequence(batch_examples[col], type=col_type, try_type=col_try_type, col=col)
--> 507 arrays.append(pa.array(typed_sequence))
508 inferred_features[col] = typed_sequence.get_inferred_type()
509 schema = inferred_features.arrow_schema if self.pa_writer is None else self.schema
/usr/local/lib/python3.7/dist-packages/pyarrow/array.pxi in pyarrow.lib.array()
/usr/local/lib/python3.7/dist-packages/pyarrow/array.pxi in pyarrow.lib._handle_arrow_array_protocol()
[/usr/local/lib/python3.7/dist-packages/datasets/arrow_writer.py](https://localhost:8080/#) in __arrow_array__(self, type)
175 storage = list_of_np_array_to_pyarrow_listarray(data, type=pa_type.value_type)
176 else:
--> 177 storage = pa.array(data, pa_type.storage_dtype)
178 return pa.ExtensionArray.from_storage(pa_type, storage)
179
/usr/local/lib/python3.7/dist-packages/pyarrow/array.pxi in pyarrow.lib.array()
/usr/local/lib/python3.7/dist-packages/pyarrow/array.pxi in pyarrow.lib._sequence_to_array()
/usr/local/lib/python3.7/dist-packages/pyarrow/error.pxi in pyarrow.lib.pyarrow_internal_check_status()
/usr/local/lib/python3.7/dist-packages/pyarrow/error.pxi in pyarrow.lib.check_status()
ArrowInvalid: Can only convert 1-dimensional array values
```
**Describe the solution you'd like**
No error in the second scenario and an identical result to the following snippets.
**Describe alternatives you've considered**
There are other alternatives that work such as:
```python
def prepare_dataset_3D_bis(batch):
batch["pixel_values"] = [[data_map[index].tolist()] for index in batch["id"]]
return batch
ds_3D_bis = ds.map(
prepare_dataset_3D_bis,
batched=True,
remove_columns=ds.column_names,
features=features.Features({"pixel_values": features.Array3D(shape=(1, 2, 3), dtype="float32")})
)
```
or
```python
def prepare_dataset_3D_ter(batch):
batch["pixel_values"] = [data_map[index][np.newaxis, :, :] for index in batch["id"]]
return batch
ds_3D_ter = ds.map(
prepare_dataset_3D_ter,
batched=True,
remove_columns=ds.column_names,
features=features.Features({"pixel_values": features.Array3D(shape=(1, 2, 3), dtype="float32")})
)
```
But both solutions require the user to be aware that `data_map[index]` is an `np.array` type.
cc @lhoestq as we discuss this offline :smile:
| false
|
1,209,901,677
|
https://api.github.com/repos/huggingface/datasets/issues/4190
|
https://github.com/huggingface/datasets/pull/4190
| 4,190
|
Deprecate `shard_size` in `push_to_hub` in favor of `max_shard_size`
|
closed
| 1
| 2022-04-20T16:08:01
| 2022-04-22T13:58:25
| 2022-04-22T13:52:00
|
mariosasko
|
[] |
This PR adds a `max_shard_size` param to `push_to_hub` and deprecates `shard_size` in favor of this new param to have a more descriptive name (a shard has at most the `shard_size` bytes in `push_to_hub`) for the param and to align the API with [Transformers](https://github.com/huggingface/transformers/blob/ff06b177917384137af2d9585697d2d76c40cdfc/src/transformers/modeling_utils.py#L1350).
| true
|
1,209,881,351
|
https://api.github.com/repos/huggingface/datasets/issues/4189
|
https://github.com/huggingface/datasets/pull/4189
| 4,189
|
Document how to use FAISS index for special operations
|
closed
| 1
| 2022-04-20T15:51:56
| 2022-05-06T08:43:10
| 2022-05-06T08:35:52
|
albertvillanova
|
[] |
Document how to use FAISS index for special operations, by accessing the index itself.
Close #4029.
| true
|
1,209,740,957
|
https://api.github.com/repos/huggingface/datasets/issues/4188
|
https://github.com/huggingface/datasets/pull/4188
| 4,188
|
Support streaming cnn_dailymail dataset
|
closed
| 2
| 2022-04-20T14:04:36
| 2022-05-11T13:39:06
| 2022-04-20T15:52:49
|
albertvillanova
|
[] |
Support streaming cnn_dailymail dataset.
Fix #3969.
CC: @severo
| true
|
1,209,721,532
|
https://api.github.com/repos/huggingface/datasets/issues/4187
|
https://github.com/huggingface/datasets/pull/4187
| 4,187
|
Don't duplicate data when encoding audio or image
|
closed
| 5
| 2022-04-20T13:50:37
| 2022-04-21T09:17:00
| 2022-04-21T09:10:47
|
lhoestq
|
[] |
Right now if you pass both the `bytes` and a local `path` for audio or image data, then the `bytes` are unnecessarily written in the Arrow file, while we could just keep the local `path`.
This PR discards the `bytes` when the audio or image file exists locally.
In particular it's common for audio datasets builders to provide both the bytes and the local path in order to work for both streaming (using the bytes) and non-streaming mode (using a local file - which is often required for audio).
cc @patrickvonplaten
| true
|
1,209,463,599
|
https://api.github.com/repos/huggingface/datasets/issues/4186
|
https://github.com/huggingface/datasets/pull/4186
| 4,186
|
Fix outdated docstring about default dataset config
|
closed
| 1
| 2022-04-20T10:04:51
| 2022-04-22T12:54:44
| 2022-04-22T12:48:31
|
lhoestq
|
[] | null | true
|
1,209,429,743
|
https://api.github.com/repos/huggingface/datasets/issues/4185
|
https://github.com/huggingface/datasets/issues/4185
| 4,185
|
Librispeech documentation, clarification on format
|
open
| 8
| 2022-04-20T09:35:55
| 2022-04-21T11:00:53
| null |
albertz
|
[] |
https://github.com/huggingface/datasets/blob/cd3ce34ab1604118351e1978d26402de57188901/datasets/librispeech_asr/librispeech_asr.py#L53
> Note that in order to limit the required storage for preparing this dataset, the audio
> is stored in the .flac format and is not converted to a float32 array. To convert, the audio
> file to a float32 array, please make use of the `.map()` function as follows:
>
> ```python
> import soundfile as sf
> def map_to_array(batch):
> speech_array, _ = sf.read(batch["file"])
> batch["speech"] = speech_array
> return batch
> dataset = dataset.map(map_to_array, remove_columns=["file"])
> ```
Is this still true?
In my case, `ds["train.100"]` returns:
```
Dataset({
features: ['file', 'audio', 'text', 'speaker_id', 'chapter_id', 'id'],
num_rows: 28539
})
```
and taking the first instance yields:
```
{'file': '374-180298-0000.flac',
'audio': {'path': '374-180298-0000.flac',
'array': array([ 7.01904297e-04, 7.32421875e-04, 7.32421875e-04, ...,
-2.74658203e-04, -1.83105469e-04, -3.05175781e-05]),
'sampling_rate': 16000},
'text': 'CHAPTER SIXTEEN I MIGHT HAVE TOLD YOU OF THE BEGINNING OF THIS LIAISON IN A FEW LINES BUT I WANTED YOU TO SEE EVERY STEP BY WHICH WE CAME I TO AGREE TO WHATEVER MARGUERITE WISHED',
'speaker_id': 374,
'chapter_id': 180298,
'id': '374-180298-0000'}
```
The `audio` `array` seems to be already decoded. So such convert/decode code as mentioned in the doc is wrong?
But I wonder, is it actually stored as flac on disk, and the decoding is done on-the-fly? Or was it decoded already during the preparation and is stored as raw samples on disk?
Note that I also used `datasets.load_dataset("librispeech_asr", "clean").save_to_disk(...)` and then `datasets.load_from_disk(...)` in this example. Does this change anything on how it is stored on disk?
A small related question: Actually I would prefer to even store it as mp3 or ogg on disk. Is this easy to convert?
| false
|
1,208,592,669
|
https://api.github.com/repos/huggingface/datasets/issues/4184
|
https://github.com/huggingface/datasets/pull/4184
| 4,184
|
[Librispeech] Add 'all' config
|
closed
| 29
| 2022-04-19T16:27:56
| 2024-08-02T05:03:04
| 2022-04-22T09:45:17
|
patrickvonplaten
|
[] |
Add `"all"` config to Librispeech
Closed #4179
| true
|
1,208,449,335
|
https://api.github.com/repos/huggingface/datasets/issues/4183
|
https://github.com/huggingface/datasets/pull/4183
| 4,183
|
Document librispeech configs
|
closed
| 5
| 2022-04-19T14:26:59
| 2023-09-24T10:02:24
| 2022-04-19T15:15:20
|
lhoestq
|
[] |
Added an example of how to load one config or the other
| true
|
1,208,285,235
|
https://api.github.com/repos/huggingface/datasets/issues/4182
|
https://github.com/huggingface/datasets/issues/4182
| 4,182
|
Zenodo.org download is not responding
|
closed
| 5
| 2022-04-19T12:26:57
| 2022-04-20T07:11:05
| 2022-04-20T07:11:05
|
dkajtoch
|
[
"bug"
] |
## Describe the bug
Source download_url from zenodo.org does not respond.
`_DOWNLOAD_URL = "https://zenodo.org/record/2787612/files/SICK.zip?download=1"`
Other datasets also use zenodo.org to store data and they cannot be downloaded as well.
It would be better to actually use more reliable way to store original data like s3 bucket.
## Steps to reproduce the bug
```python
load_dataset("sick")
```
## Expected results
Dataset should be downloaded.
## Actual results
ConnectionError: Couldn't reach https://zenodo.org/record/2787612/files/SICK.zip?download=1 (ReadTimeout(ReadTimeoutError("HTTPSConnectionPool(host='zenodo.org', port=443): Read timed out. (read timeout=100)")))
## Environment info
- `datasets` version: 2.1.0
- Platform: Darwin-21.4.0-x86_64-i386-64bit
- Python version: 3.7.11
- PyArrow version: 7.0.0
- Pandas version: 1.3.5
| false
|
1,208,194,805
|
https://api.github.com/repos/huggingface/datasets/issues/4181
|
https://github.com/huggingface/datasets/issues/4181
| 4,181
|
Support streaming FLEURS dataset
|
closed
| 9
| 2022-04-19T11:09:56
| 2022-07-25T11:44:02
| 2022-07-25T11:44:02
|
patrickvonplaten
|
[
"dataset bug"
] |
## Dataset viewer issue for '*name of the dataset*'
https://huggingface.co/datasets/google/fleurs
```
Status code: 400
Exception: NotImplementedError
Message: Extraction protocol for TAR archives like 'https://storage.googleapis.com/xtreme_translations/FLEURS/af_za.tar.gz' is not implemented in streaming mode. Please use `dl_manager.iter_archive` instead.
```
Am I the one who added this dataset ? Yes
Can I fix this somehow in the script? @lhoestq @severo
| false
|
1,208,042,320
|
https://api.github.com/repos/huggingface/datasets/issues/4180
|
https://github.com/huggingface/datasets/issues/4180
| 4,180
|
Add some iteration method on a dataset column (specific for inference)
|
closed
| 5
| 2022-04-19T09:15:45
| 2025-06-17T13:08:50
| 2025-06-17T13:08:50
|
Narsil
|
[
"enhancement"
] |
**Is your feature request related to a problem? Please describe.**
A clear and concise description of what the problem is.
Currently, `dataset["audio"]` will load EVERY element in the dataset in RAM, which can be quite big for an audio dataset.
Having an iterator (or sequence) type of object, would make inference with `transformers` 's `pipeline` easier to use and not so memory hungry.
**Describe the solution you'd like**
A clear and concise description of what you want to happen.
For a non breaking change:
```python
for audio in dataset.iterate("audio"):
# {"array": np.array(...), "sampling_rate":...}
```
For a breaking change solution (not necessary), changing the type of `dataset["audio"]` to a sequence type so that
```python
pipe = pipeline(model="...")
for out in pipe(dataset["audio"]):
# {"text":....}
```
could work
**Describe alternatives you've considered**
A clear and concise description of any alternative solutions or features you've considered.
```python
def iterate(dataset, key):
for item in dataset:
yield dataset[key]
for out in pipeline(iterate(dataset, "audio")):
# {"array": ...}
```
This works but requires the helper function which feels slightly clunky.
**Additional context**
Add any other context about the feature request here.
The context is actually to showcase better integration between `pipeline` and `datasets` in the Quicktour demo: https://github.com/huggingface/transformers/pull/16723/files
@lhoestq
| false
|
1,208,001,118
|
https://api.github.com/repos/huggingface/datasets/issues/4179
|
https://github.com/huggingface/datasets/issues/4179
| 4,179
|
Dataset librispeech_asr fails to load
|
closed
| 21
| 2022-04-19T08:45:48
| 2022-07-27T16:10:00
| 2022-07-27T16:10:00
|
albertz
|
[
"bug"
] |
## Describe the bug
The dataset librispeech_asr (standard Librispeech) fails to load.
## Steps to reproduce the bug
```python
datasets.load_dataset("librispeech_asr")
```
## Expected results
It should download and prepare the whole dataset (all subsets).
In [the doc](https://huggingface.co/datasets/librispeech_asr), it says it has two configurations (clean and other).
However, the dataset doc says that not specifying `split` should just load the whole dataset, which is what I want.
Also, in case of this specific dataset, this is also the standard what the community uses. When you look at any publications with results on Librispeech, they always use the whole train dataset for training.
## Actual results
```
...
File "/home/az/.cache/huggingface/modules/datasets_modules/datasets/librispeech_asr/1f4602f6b5fed8d3ab3e3382783173f2e12d9877e98775e34d7780881175096c/librispeech_asr.py", line 119, in LibrispeechASR._split_generators
line: archive_path = dl_manager.download(_DL_URLS[self.config.name])
locals:
archive_path = <not found>
dl_manager = <local> <datasets.utils.download_manager.DownloadManager object at 0x7fc07b426160>
dl_manager.download = <local> <bound method DownloadManager.download of <datasets.utils.download_manager.DownloadManager object at 0x7fc07b426160>>
_DL_URLS = <global> {'clean': {'dev': 'http://www.openslr.org/resources/12/dev-clean.tar.gz', 'test': 'http://www.openslr.org/resources/12/test-clean.tar.gz', 'train.100': 'http://www.openslr.org/resources/12/train-clean-100.tar.gz', 'train.360': 'http://www.openslr.org/resources/12/train-clean-360.tar.gz'}, 'other'...
self = <local> <datasets_modules.datasets.librispeech_asr.1f4602f6b5fed8d3ab3e3382783173f2e12d9877e98775e34d7780881175096c.librispeech_asr.LibrispeechASR object at 0x7fc12a633310>
self.config = <local> BuilderConfig(name='default', version=0.0.0, data_dir='/home/az/i6/setups/2022-03-20--sis/work/i6_core/datasets/huggingface/DownloadAndPrepareHuggingFaceDatasetJob.TV6Nwm6dFReF/output/data_dir', data_files=None, description=None)
self.config.name = <local> 'default', len = 7
KeyError: 'default'
```
## Environment info
<!-- You can run the command `datasets-cli env` and copy-and-paste its output below. -->
- `datasets` version: 2.1.0
- Platform: Linux-5.4.0-107-generic-x86_64-with-glibc2.31
- Python version: 3.9.9
- PyArrow version: 6.0.1
- Pandas version: 1.4.2
| false
|
1,207,787,073
|
https://api.github.com/repos/huggingface/datasets/issues/4178
|
https://github.com/huggingface/datasets/pull/4178
| 4,178
|
[feat] Add ImageNet dataset
|
closed
| 3
| 2022-04-19T06:01:35
| 2022-04-29T21:43:59
| 2022-04-29T21:37:08
|
apsdehal
|
[] |
To use the dataset download the tar file
[imagenet_object_localization_patched2019.tar.gz](https://www.kaggle.com/competitions/imagenet-object-localization-challenge/data?select=imagenet_object_localization_patched2019.tar.gz) from Kaggle and then point the datasets library to it by using:
```py
from datasets import load_dataset
dataset = load_dataset("imagenet",
data_dir="/path/to/imagenet_object_localization_patched2019.tar.gz")
```
Currently train and validation splits are supported.
| true
|
1,207,535,920
|
https://api.github.com/repos/huggingface/datasets/issues/4177
|
https://github.com/huggingface/datasets/pull/4177
| 4,177
|
Adding missing subsets to the `SemEval-2018 Task 1` dataset
|
open
| 1
| 2022-04-18T22:59:30
| 2022-10-05T10:38:16
| null |
micahcarroll
|
[
"dataset contribution"
] |
This dataset for the [1st task of SemEval-2018](https://competitions.codalab.org/competitions/17751) competition was missing all subtasks except for subtask 5. I added another two subtasks (subtask 1 and 2), which are each comprised of 12 additional data subsets: for each language in En, Es, Ar, there are 4 datasets, broken down by emotions (anger, fear, joy, sadness).
## Remaining questions
I wasn't able to find any documentation about how one should make PRs to modify datasets. Because of that, I just did my best to integrate the new data into the code, and tested locally that this worked. I'm sorry if I'm not respecting your contributing guidelines – if they are documented somewhere, I'd appreciate if you could send a pointer!
Not sure how `dataset_infos.json` and `dummy` should be updated. My understanding is that they were automatically generated at the time of the original dataset creation?
| true
|
1,206,515,563
|
https://api.github.com/repos/huggingface/datasets/issues/4176
|
https://github.com/huggingface/datasets/issues/4176
| 4,176
|
Very slow between two operations
|
closed
| 0
| 2022-04-17T23:52:29
| 2022-04-18T00:03:00
| 2022-04-18T00:03:00
|
yanan1116
|
[
"bug"
] |
Hello, in the processing stage, I use two operations. The first one : map + filter, is very fast and it uses the full cores, while the socond step is very slow and did not use full cores.
Also, there is a significant lag between them. Am I missing something ?
```
raw_datasets = raw_datasets.map(split_func,
batched=False,
num_proc=args.preprocessing_num_workers,
load_from_cache_file=not args.overwrite_cache,
desc = "running split para ==>")\
.filter(lambda example: example['text1']!='' and example['text2']!='',
num_proc=args.preprocessing_num_workers, desc="filtering ==>")
processed_datasets = raw_datasets.map(
preprocess_function,
batched=True,
num_proc=args.preprocessing_num_workers,
remove_columns=column_names,
load_from_cache_file=not args.overwrite_cache,
desc="Running tokenizer on dataset===>",
)
```
| false
|
1,205,589,842
|
https://api.github.com/repos/huggingface/datasets/issues/4175
|
https://github.com/huggingface/datasets/pull/4175
| 4,175
|
Add WIT Dataset
|
closed
| 6
| 2022-04-15T13:42:32
| 2023-09-24T10:02:38
| 2022-05-02T14:26:41
|
thomasw21
|
[] |
closes #2981 #2810
@nateraw @hassiahk I've listed you guys as co-author as you've contributed previously to this dataset
| true
|
1,205,575,941
|
https://api.github.com/repos/huggingface/datasets/issues/4174
|
https://github.com/huggingface/datasets/pull/4174
| 4,174
|
Fix when map function modifies input in-place
|
closed
| 1
| 2022-04-15T13:23:15
| 2022-04-15T14:52:07
| 2022-04-15T14:45:58
|
thomasw21
|
[] |
When `function` modifies input in-place, the guarantee that columns in `remove_columns` are contained in `input` doesn't hold true anymore. Therefore we need to relax way we pop elements by checking if that column exists.
| true
|
1,204,657,114
|
https://api.github.com/repos/huggingface/datasets/issues/4173
|
https://github.com/huggingface/datasets/pull/4173
| 4,173
|
Stream private zipped images
|
closed
| 3
| 2022-04-14T15:15:07
| 2022-05-05T14:05:54
| 2022-05-05T13:58:35
|
lhoestq
|
[] |
As mentioned in https://github.com/huggingface/datasets/issues/4139 it's currently not possible to stream private/gated zipped images from the Hub.
This is because `Image.decode_example` does not handle authentication. Indeed decoding requires to access and download the file from the private repository.
In this PR I added authentication to `Image.decode_example` via a `token_per_repo_id` optional argument. I first wanted to just pass `use_auth_token` but a single `Image` instance can be responsible of decoding images from a combination of several datasets together (from `interleave_datasets` for example). Therefore I just used a dictionary `repo_id` -> `token` instead.
I'm getting the `repo_id` from the dataset builder (I replaced the `namepace` attribute with `repo_id`)
I did the same for `Audio.decode_example`
cc @SBrandeis @severo
| true
|
1,204,433,160
|
https://api.github.com/repos/huggingface/datasets/issues/4172
|
https://github.com/huggingface/datasets/pull/4172
| 4,172
|
Update assin2 dataset_infos.json
|
closed
| 1
| 2022-04-14T11:53:06
| 2022-04-15T14:47:42
| 2022-04-15T14:41:22
|
lhoestq
|
[] |
Following comments in https://github.com/huggingface/datasets/issues/4003 we found that it was outdated and casing an error when loading the dataset
| true
|
1,204,413,620
|
https://api.github.com/repos/huggingface/datasets/issues/4170
|
https://github.com/huggingface/datasets/pull/4170
| 4,170
|
to_tf_dataset rewrite
|
closed
| 15
| 2022-04-14T11:30:58
| 2022-06-06T14:31:12
| 2022-06-06T14:22:09
|
Rocketknight1
|
[] |
This PR rewrites almost all of `to_tf_dataset()`, which makes it kind of hard to list all the changes, but the most critical ones are:
- Much better stability and no more dropping unexpected column names (Sorry @NielsRogge)
- Doesn't clobber custom transforms on the data (Sorry @NielsRogge again)
- Much better handling of the situation when the `collate_fn` adds columns that aren't in the dataset.
- Better inference of shapes and data types
- Lots of hacky special-casing code removed
- Can return string columns (as `tf.String`)
- Most arguments have default values, calling the method should be much simpler
- ~~Can accept a `model` argument and only return columns that are valid inputs to that model~~
- Drops the `dummy_labels` argument - this was a workaround for Keras issues that have been resolved by changes in `transformers`. Also remove it from tests and the Overview notebook.
I still have a couple of TODOs remaining and some testing to do, so don't merge yet, but it should be mostly ready for review at this point!
| true
|
1,203,995,869
|
https://api.github.com/repos/huggingface/datasets/issues/4169
|
https://github.com/huggingface/datasets/issues/4169
| 4,169
|
Timit_asr dataset cannot be previewed recently
|
closed
| 5
| 2022-04-14T03:28:31
| 2023-02-03T04:54:57
| 2022-05-06T16:06:51
|
YingLi001
|
[] |
## Dataset viewer issue for '*timit_asr*'
**Link:** *https://huggingface.co/datasets/timit_asr*
Issue: The timit-asr dataset cannot be previewed recently.
Am I the one who added this dataset ? Yes-No
No
| false
|
1,203,867,540
|
https://api.github.com/repos/huggingface/datasets/issues/4168
|
https://github.com/huggingface/datasets/pull/4168
| 4,168
|
Add code examples to API docs
|
closed
| 4
| 2022-04-13T23:03:38
| 2022-04-27T18:53:37
| 2022-04-27T18:48:34
|
stevhliu
|
[
"documentation"
] |
This PR adds code examples for functions related to the base Datasets class to highlight usage. Most of the examples use the `rotten_tomatoes` dataset since it is nice and small. Several things I would appreciate feedback on:
- Do you think it is clearer to make every code example fully reproducible so when users copy the code they can actually run it and get an output? This seems quite repetitive - maybe even unnecessary - but it is definitely clearer. Personally, I think we might be able to get away with not including this since users probably want to try the function on their own dataset. For example:
```py
>>> from datasets import load_dataset
>>> ds = load_dataset("rotten_tomatoes", split="validation")
>>> code example goes here
```
- Should we showcase a function with more than one parameter to highlight different use-cases (it's pretty basic right now, but I'd be happy to add more)?
- For the `class_encode_column` function, let me know if there is a simpler dataset with fewer columns (currently using `winograd_wsc`) so it is easier for users to see what changed.
- Where possible, I try to show the input before and the output after using a function like `flatten` for example. Do you think this is too much and just showing the usage (ie, `>>> ds.flatten()`) will be sufficient?
Thanks :)
| true
|
1,203,761,614
|
https://api.github.com/repos/huggingface/datasets/issues/4167
|
https://github.com/huggingface/datasets/pull/4167
| 4,167
|
Avoid rate limit in update hub repositories
|
closed
| 2
| 2022-04-13T20:32:17
| 2022-04-13T20:56:41
| 2022-04-13T20:50:32
|
lhoestq
|
[] |
use http.extraHeader to avoid rate limit
| true
|
1,203,758,004
|
https://api.github.com/repos/huggingface/datasets/issues/4166
|
https://github.com/huggingface/datasets/pull/4166
| 4,166
|
Fix exact match
|
closed
| 1
| 2022-04-13T20:28:06
| 2022-05-03T12:23:31
| 2022-05-03T12:16:27
|
emibaylor
|
[] |
Clarify docs and add clarifying example to the exact_match metric
| true
|
1,203,730,187
|
https://api.github.com/repos/huggingface/datasets/issues/4165
|
https://github.com/huggingface/datasets/pull/4165
| 4,165
|
Fix google bleu typos, examples
|
closed
| 1
| 2022-04-13T19:59:54
| 2022-05-03T12:23:52
| 2022-05-03T12:16:44
|
emibaylor
|
[] | null | true
|
1,203,661,346
|
https://api.github.com/repos/huggingface/datasets/issues/4164
|
https://github.com/huggingface/datasets/pull/4164
| 4,164
|
Fix duplicate key in multi_news
|
closed
| 1
| 2022-04-13T18:48:24
| 2022-04-13T21:04:16
| 2022-04-13T20:58:02
|
lhoestq
|
[] |
To merge after this job succeeded: https://github.com/huggingface/datasets/runs/6012207928
| true
|
1,203,539,268
|
https://api.github.com/repos/huggingface/datasets/issues/4163
|
https://github.com/huggingface/datasets/issues/4163
| 4,163
|
Optional Content Warning for Datasets
|
open
| 5
| 2022-04-13T16:38:01
| 2022-06-09T20:39:02
| null |
TristanThrush
|
[
"enhancement"
] |
**Is your feature request related to a problem? Please describe.**
A clear and concise description of what the problem is.
We now have hate speech datasets on the hub, like this one: https://huggingface.co/datasets/HannahRoseKirk/HatemojiBuild
I'm wondering if there is an option to select a content warning message that appears before the dataset preview? Otherwise, people immediately see hate speech when clicking on this dataset.
**Describe the solution you'd like**
A clear and concise description of what you want to happen.
Implementation of a content warning message that separates users from the dataset preview until they click out of the warning.
**Describe alternatives you've considered**
A clear and concise description of any alternative solutions or features you've considered.
Possibly just a way to remove the dataset preview completely? I think I like the content warning option better, though.
**Additional context**
Add any other context about the feature request here.
| false
|
1,203,421,909
|
https://api.github.com/repos/huggingface/datasets/issues/4162
|
https://github.com/huggingface/datasets/pull/4162
| 4,162
|
Add Conceptual 12M
|
closed
| 2
| 2022-04-13T14:57:23
| 2022-04-15T08:13:01
| 2022-04-15T08:06:25
|
thomasw21
|
[] | null | true
|
1,203,230,485
|
https://api.github.com/repos/huggingface/datasets/issues/4161
|
https://github.com/huggingface/datasets/pull/4161
| 4,161
|
Add Visual Genome
|
closed
| 4
| 2022-04-13T12:25:24
| 2022-04-21T15:42:49
| 2022-04-21T13:08:52
|
thomasw21
|
[] | null | true
|
1,202,845,874
|
https://api.github.com/repos/huggingface/datasets/issues/4160
|
https://github.com/huggingface/datasets/issues/4160
| 4,160
|
RGBA images not showing
|
closed
| 2
| 2022-04-13T06:59:23
| 2022-06-21T16:43:11
| 2022-06-21T16:43:11
|
cceyda
|
[
"dataset-viewer",
"dataset-viewer-rgba-images"
] |
## Dataset viewer issue for ceyda/smithsonian_butterflies_transparent
[**Link:** *link to the dataset viewer page*](https://huggingface.co/datasets/ceyda/smithsonian_butterflies_transparent)
Am I the one who added this dataset ? Yes
👉 More of a general issue of 'RGBA' png images not being supported
(the dataset itself is just for the huggan sprint and not that important, consider it just an example)
| false
|
1,202,522,153
|
https://api.github.com/repos/huggingface/datasets/issues/4159
|
https://github.com/huggingface/datasets/pull/4159
| 4,159
|
Add `TruthfulQA` dataset
|
closed
| 2
| 2022-04-12T23:19:04
| 2022-06-08T15:51:33
| 2022-06-08T14:43:34
|
jon-tow
|
[] | null | true
|
1,202,376,843
|
https://api.github.com/repos/huggingface/datasets/issues/4158
|
https://github.com/huggingface/datasets/pull/4158
| 4,158
|
Add AUC ROC Metric
|
closed
| 1
| 2022-04-12T20:53:28
| 2022-04-26T19:41:50
| 2022-04-26T19:35:22
|
emibaylor
|
[] | null | true
|
1,202,239,622
|
https://api.github.com/repos/huggingface/datasets/issues/4157
|
https://github.com/huggingface/datasets/pull/4157
| 4,157
|
Fix formatting in BLEU metric card
|
closed
| 1
| 2022-04-12T18:29:51
| 2022-04-13T14:30:25
| 2022-04-13T14:16:34
|
mariosasko
|
[] |
Fix #4148
| true
|
1,202,220,531
|
https://api.github.com/repos/huggingface/datasets/issues/4156
|
https://github.com/huggingface/datasets/pull/4156
| 4,156
|
Adding STSb-TR dataset
|
closed
| 1
| 2022-04-12T18:10:05
| 2022-10-03T09:36:25
| 2022-10-03T09:36:25
|
figenfikri
|
[
"dataset contribution"
] |
Semantic Textual Similarity benchmark Turkish (STSb-TR) dataset introduced in our paper [Semantic Similarity Based Evaluation for Abstractive News Summarization](https://aclanthology.org/2021.gem-1.3.pdf) added.
| true
|
1,202,183,608
|
https://api.github.com/repos/huggingface/datasets/issues/4155
|
https://github.com/huggingface/datasets/pull/4155
| 4,155
|
Make HANS dataset streamable
|
closed
| 1
| 2022-04-12T17:34:13
| 2022-04-13T12:03:46
| 2022-04-13T11:57:35
|
mariosasko
|
[] |
Fix #4133
| true
|
1,202,145,721
|
https://api.github.com/repos/huggingface/datasets/issues/4154
|
https://github.com/huggingface/datasets/pull/4154
| 4,154
|
Generate tasks.json taxonomy from `huggingface_hub`
|
closed
| 7
| 2022-04-12T17:12:46
| 2022-04-14T10:32:32
| 2022-04-14T10:26:13
|
julien-c
|
[] | null | true
|
1,202,040,506
|
https://api.github.com/repos/huggingface/datasets/issues/4153
|
https://github.com/huggingface/datasets/pull/4153
| 4,153
|
Adding Text-based NP Enrichment (TNE) dataset
|
closed
| 3
| 2022-04-12T15:47:03
| 2022-05-03T14:05:48
| 2022-05-03T14:05:48
|
yanaiela
|
[] |
Added the [TNE](https://github.com/yanaiela/TNE) dataset to the library
| true
|
1,202,034,115
|
https://api.github.com/repos/huggingface/datasets/issues/4152
|
https://github.com/huggingface/datasets/issues/4152
| 4,152
|
ArrayND error in pyarrow 5
|
closed
| 2
| 2022-04-12T15:41:40
| 2022-05-04T09:29:46
| 2022-05-04T09:29:46
|
lhoestq
|
[] |
As found in https://github.com/huggingface/datasets/pull/3903, The ArrayND features fail on pyarrow 5:
```python
import pyarrow as pa
from datasets import Array2D
from datasets.table import cast_array_to_feature
arr = pa.array([[[0]]])
feature_type = Array2D(shape=(1, 1), dtype="int64")
cast_array_to_feature(arr, feature_type)
```
raises
```python
---------------------------------------------------------------------------
AttributeError Traceback (most recent call last)
<ipython-input-8-04610f9fa78c> in <module>
----> 1 cast_array_to_feature(pa.array([[[0]]]), Array2D(shape=(1, 1), dtype="int32"))
~/Desktop/hf/datasets/src/datasets/table.py in wrapper(array, *args, **kwargs)
1672 return pa.chunked_array([func(chunk, *args, **kwargs) for chunk in array.chunks])
1673 else:
-> 1674 return func(array, *args, **kwargs)
1675
1676 return wrapper
~/Desktop/hf/datasets/src/datasets/table.py in cast_array_to_feature(array, feature, allow_number_to_str)
1806 return array_cast(array, get_nested_type(feature), allow_number_to_str=allow_number_to_str)
1807 elif not isinstance(feature, (Sequence, dict, list, tuple)):
-> 1808 return array_cast(array, feature(), allow_number_to_str=allow_number_to_str)
1809 raise TypeError(f"Couldn't cast array of type\n{array.type}\nto\n{feature}")
1810
~/Desktop/hf/datasets/src/datasets/table.py in wrapper(array, *args, **kwargs)
1672 return pa.chunked_array([func(chunk, *args, **kwargs) for chunk in array.chunks])
1673 else:
-> 1674 return func(array, *args, **kwargs)
1675
1676 return wrapper
~/Desktop/hf/datasets/src/datasets/table.py in array_cast(array, pa_type, allow_number_to_str)
1705 array = array.storage
1706 if isinstance(pa_type, pa.ExtensionType):
-> 1707 return pa_type.wrap_array(array)
1708 elif pa.types.is_struct(array.type):
1709 if pa.types.is_struct(pa_type) and (
AttributeError: 'Array2DExtensionType' object has no attribute 'wrap_array'
```
The thing is that `cast_array_to_feature` is called when writing an Arrow file, so creating an Arrow dataset using any ArrayND type currently fails.
`wrap_array` has been added in pyarrow 6, so we can either bump the required pyarrow version or fix this for pyarrow 5
| false
|
1,201,837,999
|
https://api.github.com/repos/huggingface/datasets/issues/4151
|
https://github.com/huggingface/datasets/pull/4151
| 4,151
|
Add missing label for emotion description
|
closed
| 0
| 2022-04-12T13:17:37
| 2022-04-12T13:58:50
| 2022-04-12T13:58:50
|
lijiazheng99
|
[] | null | true
|
1,201,689,730
|
https://api.github.com/repos/huggingface/datasets/issues/4150
|
https://github.com/huggingface/datasets/issues/4150
| 4,150
|
Inconsistent splits generation for datasets without loading script (packaged dataset puts everything into a single split)
|
closed
| 0
| 2022-04-12T11:15:55
| 2022-04-28T21:02:44
| 2022-04-28T21:02:44
|
polinaeterna
|
[
"bug"
] |
## Describe the bug
Splits for dataset loaders without scripts are prepared inconsistently. I think it might be confusing for users.
## Steps to reproduce the bug
* If you load a packaged datasets from Hub, it infers splits from directory structure / filenames (check out the data [here](https://huggingface.co/datasets/nateraw/test-imagefolder-dataset)):
```python
ds = load_dataset("nateraw/test-imagefolder-dataset")
print(ds)
### Output:
DatasetDict({
train: Dataset({
features: ['image', 'label'],
num_rows: 6
})
test: Dataset({
features: ['image', 'label'],
num_rows: 4
})
})
```
* If you do the same from locally stored data specifying only directory path you'll get the same:
```python
ds = load_dataset("/path/to/local/data/test-imagefolder-dataset")
print(ds)
### Output:
DatasetDict({
train: Dataset({
features: ['image', 'label'],
num_rows: 6
})
test: Dataset({
features: ['image', 'label'],
num_rows: 4
})
})
```
* However, if you explicitely specify package name (like `imagefolder`, `csv`, `json`), all the data is put into a single split:
```python
ds = load_dataset("imagefolder", data_dir="/path/to/local/data/test-imagefolder-dataset")
print(ds)
### Output:
DatasetDict({
train: Dataset({
features: ['image', 'label'],
num_rows: 10
})
})
```
## Expected results
For `load_dataset("imagefolder", data_dir="/path/to/local/data/test-imagefolder-dataset")` I expect the same output as of the two first options.
| false
|
1,201,389,221
|
https://api.github.com/repos/huggingface/datasets/issues/4149
|
https://github.com/huggingface/datasets/issues/4149
| 4,149
|
load_dataset for winoground returning decoding error
|
closed
| 10
| 2022-04-12T08:16:16
| 2022-05-04T23:40:38
| 2022-05-04T23:40:38
|
odellus
|
[
"bug"
] |
## Describe the bug
I am trying to use datasets to load winoground and I'm getting a JSON decoding error.
## Steps to reproduce the bug
```python
from datasets import load_dataset
token = 'hf_XXXXX' # my HF access token
datasets = load_dataset('facebook/winoground', use_auth_token=token)
```
## Expected results
I downloaded images.zip and examples.jsonl manually. I was expecting to have some trouble decoding json so I didn't use jsonlines but instead was able to get a complete set of 400 examples by doing
```python
import json
with open('examples.jsonl', 'r') as f:
examples = f.read().split('\n')
# Thinking this would error if the JSON is not utf-8 encoded
json_data = [json.loads(x) for x in examples]
print(json_data[-1])
```
and I see
```python
{'caption_0': 'someone is overdoing it',
'caption_1': 'someone is doing it over',
'collapsed_tag': 'Relation',
'id': 399,
'image_0': 'ex_399_img_0',
'image_1': 'ex_399_img_1',
'num_main_preds': 1,
'secondary_tag': 'Morpheme-Level',
'tag': 'Scope, Preposition'}
```
so I'm not sure what's going on here honestly. The file `examples.jsonl` doesn't have non-UTF-8 encoded text.
## Actual results
During the split operation after downloading, datasets encounters an error in the JSON ([trace](https://gist.github.com/odellus/e55d390ca203386bf551f38e0c63a46b) abbreviated for brevity).
```
datasets/packaged_modules/json/json.py:144 in Json._generate_tables(self, files)
...
UnicodeDecodeError: 'utf-8' codec can't decode byte 0xff in position 0: invalid start byte
```
## Environment info
<!-- You can run the command `datasets-cli env` and copy-and-paste its output below. -->
- `datasets` version: 1.18.4
- Platform: Linux-5.13.0-39-generic-x86_64-with-glibc2.29
- Python version: 3.8.10
- PyArrow version: 7.0.0
| false
|
1,201,169,242
|
https://api.github.com/repos/huggingface/datasets/issues/4148
|
https://github.com/huggingface/datasets/issues/4148
| 4,148
|
fix confusing bleu metric example
|
closed
| 0
| 2022-04-12T06:18:26
| 2022-04-13T14:16:34
| 2022-04-13T14:16:34
|
aizawa-naoki
|
[
"enhancement"
] |
**Is your feature request related to a problem? Please describe.**
I would like to see the example in "Metric Card for BLEU" changed.
The 0th element in the predictions list is not closed in square brackets, and the 1st list is missing a comma.
The BLEU score are calculated correctly, but it is difficult to understand, so it would be helpful if you could correct this.
```
>> predictions = [
... ["hello", "there", "general", "kenobi", # <- no closing square bracket.
... ["foo", "bar" "foobar"] # <- no comma between "bar" and "foobar"
... ]
>>> references = [
... [["hello", "there", "general", "kenobi"]],
... [["foo", "bar", "foobar"]]
... ]
>>> bleu = datasets.load_metric("bleu")
>>> results = bleu.compute(predictions=predictions, references=references)
>>> print(results)
{'bleu': 0.6370964381207871, ...
```
**Describe the solution you'd like**
```
>> predictions = [
... ["hello", "there", "general", "kenobi", # <- no closing square bracket.
... ["foo", "bar" "foobar"] # <- no comma between "bar" and "foobar"
... ]
# and
>>> print(results)
{'bleu':1.0, ...
```
| false
|
1,200,756,008
|
https://api.github.com/repos/huggingface/datasets/issues/4147
|
https://github.com/huggingface/datasets/pull/4147
| 4,147
|
Adjust path to datasets tutorial in How-To
|
closed
| 1
| 2022-04-12T01:20:34
| 2022-04-12T08:32:24
| 2022-04-12T08:26:02
|
NimaBoscarino
|
[] |
The link in the How-To overview page to the Datasets tutorials is currently broken. This is just a small adjustment to make it match the format used in https://github.com/huggingface/datasets/blob/master/docs/source/tutorial.md.
(Edit to add: The link in the PR deployment (https://moon-ci-docs.huggingface.co/docs/datasets/pr_4147/en/how_to) is also broken since it's actually hardcoded to `master` and not dynamic to the branch name, but other links seem to behave similarly.)
| true
|
1,200,215,789
|
https://api.github.com/repos/huggingface/datasets/issues/4146
|
https://github.com/huggingface/datasets/issues/4146
| 4,146
|
SAMSum dataset viewer not working
|
closed
| 3
| 2022-04-11T16:22:57
| 2022-04-29T16:26:09
| 2022-04-29T16:26:09
|
aakashnegi10
|
[
"bug"
] |
## Dataset viewer issue for '*name of the dataset*'
**Link:** *link to the dataset viewer page*
*short description of the issue*
Am I the one who added this dataset ? Yes-No
| false
|
1,200,209,781
|
https://api.github.com/repos/huggingface/datasets/issues/4145
|
https://github.com/huggingface/datasets/pull/4145
| 4,145
|
Redirect TIMIT download from LDC
|
closed
| 4
| 2022-04-11T16:17:55
| 2022-04-13T15:39:31
| 2022-04-13T15:33:04
|
lhoestq
|
[] |
LDC data is protected under US copyright laws and under various legal agreements between the Linguistic Data Consortium/the University of Pennsylvania and data providers which prohibit redistribution of that data by anyone other than LDC. Similarly, LDC's membership agreements, non-member user agreement and various corpus-specific license agreements specifically state that users cannot publish, retransmit, disclose, copy, reproduce or redistribute LDC databases to others outside their organizations.
LDC explicitly asked us to remove the download script for the TIMIT dataset. In this PR I remove all means to download the dataset, and redirect users to download the data from https://catalog.ldc.upenn.edu/LDC93S1
| true
|
1,200,016,983
|
https://api.github.com/repos/huggingface/datasets/issues/4144
|
https://github.com/huggingface/datasets/pull/4144
| 4,144
|
Fix splits in local packaged modules, local datasets without script and hub datasets without script
|
closed
| 7
| 2022-04-11T13:57:33
| 2022-04-29T09:12:14
| 2022-04-28T21:02:45
|
polinaeterna
|
[] |
fixes #4150
I suggest to infer splits structure from files when `data_dir` is passed with `get_patterns_locally`, analogous to what's done in `LocalDatasetModuleFactoryWithoutScript` with `self.path`, instead of generating files with `data_dir/**` patterns and putting them all into a single default (train) split.
I would also suggest to align `HubDatasetModuleFactoryWithoutScript` and `LocalDatasetModuleFactoryWithoutScript` with this logic (remove `data_files = os.path.join(data_dir, "**")`). It's not reflected in the current code now as I'd like to discuss it cause I might be unaware of some use cases. @lhoestq @mariosasko @albertvillanova WDYT?
| true
|
1,199,937,961
|
https://api.github.com/repos/huggingface/datasets/issues/4143
|
https://github.com/huggingface/datasets/issues/4143
| 4,143
|
Unable to download `Wikepedia` 20220301.en version
|
closed
| 3
| 2022-04-11T13:00:14
| 2022-08-17T00:37:55
| 2022-04-21T17:04:14
|
beyondguo
|
[
"bug"
] |
## Describe the bug
Unable to download `Wikepedia` dataset, 20220301.en version
## Steps to reproduce the bug
```python
!pip install apache_beam mwparserfromhell
dataset_wikipedia = load_dataset("wikipedia", "20220301.en")
```
## Actual results
```
ValueError: BuilderConfig 20220301.en not found.
Available: ['20200501.aa', '20200501.ab', '20200501.ace', '20200501.ady', '20200501.af', '20200501.ak', '20200501.als', '20200501.am', '20200501.an', '20200501.ang', '20200501.ar', '20200501.arc', '20200501.arz', '20200501.as', '20200501.ast', '20200501.atj', '20200501.av', '20200501.ay', '20200501.az', '20200501.azb', '20200501.ba', '20200501.bar', '20200501.bat-smg', '20200501.bcl', '20200501.be', '20200501.be-x-old', '20200501.bg', '20200501.bh', '20200501.bi', '20200501.bjn', '20200501.bm', '20200501.bn', '20200501.bo', '20200501.bpy', '20200501.br', '20200501.bs', '20200501.bug', '20200501.bxr', '20200501.ca', '20200501.cbk-zam', '20200501.cdo', '20200501.ce', '20200501.ceb', '20200501.ch', '20200501.cho', '20200501.chr', '20200501.chy', '20200501.ckb', '20200501.co', '20200501.cr', '20200501.crh', '20200501.cs', '20200501.csb', '20200501.cu', '20200501.cv', '20200501.cy', '20200501.da', '20200501.de', '20200501.din', '20200501.diq', '20200501.dsb', '20200501.dty', '20200501.dv', '20200501.dz', '20200501.ee', '20200501.el', '20200501.eml', '20200501.en', '20200501.eo', '20200501.es', '20200501.et', '20200501.eu', '20200501.ext', '20200501.fa', '20200501.ff', '20200501.fi', '20200501.fiu-vro', '20200501.fj', '20200501.fo', '20200501.fr', '20200501.frp', '20200501.frr', '20200501.fur', '20200501.fy', '20200501.ga', '20200501.gag', '20200501.gan', '20200501.gd', '20200501.gl', '20200501.glk', '20200501.gn', '20200501.gom', '20200501.gor', '20200501.got', '20200501.gu', '20200501.gv', '20200501.ha', '20200501.hak', '20200501.haw', '20200501.he', '20200501.hi', '20200501.hif', '20200501.ho', '20200501.hr', '20200501.hsb', '20200501.ht', '20200501.hu', '20200501.hy', '20200501.ia', '20200501.id', '20200501.ie', '20200501.ig', '20200501.ii', '20200501.ik', '20200501.ilo', '20200501.inh', '20200501.io', '20200501.is', '20200501.it', '20200501.iu', '20200501.ja', '20200501.jam', '20200501.jbo', '20200501.jv', '20200501.ka', '20200501.kaa', '20200501.kab', '20200501.kbd', '20200501.kbp', '20200501.kg', '20200501.ki', '20200501.kj', '20200501.kk', '20200501.kl', '20200501.km', '20200501.kn', '20200501.ko', '20200501.koi', '20200501.krc', '20200501.ks', '20200501.ksh', '20200501.ku', '20200501.kv', '20200501.kw', '20200501.ky', '20200501.la', '20200501.lad', '20200501.lb', '20200501.lbe', '20200501.lez', '20200501.lfn', '20200501.lg', '20200501.li', '20200501.lij', '20200501.lmo', '20200501.ln', '20200501.lo', '20200501.lrc', '20200501.lt', '20200501.ltg', '20200501.lv', '20200501.mai', '20200501.map-bms', '20200501.mdf', '20200501.mg', '20200501.mh', '20200501.mhr', '20200501.mi', '20200501.min', '20200501.mk', '20200501.ml', '20200501.mn', '20200501.mr', '20200501.mrj', '20200501.ms', '20200501.mt', '20200501.mus', '20200501.mwl', '20200501.my', '20200501.myv', '20200501.mzn', '20200501.na', '20200501.nah', '20200501.nap', '20200501.nds', '20200501.nds-nl', '20200501.ne', '20200501.new', '20200501.ng', '20200501.nl', '20200501.nn', '20200501.no', '20200501.nov', '20200501.nrm', '20200501.nso', '20200501.nv', '20200501.ny', '20200501.oc', '20200501.olo', '20200501.om', '20200501.or', '20200501.os', '20200501.pa', '20200501.pag', '20200501.pam', '20200501.pap', '20200501.pcd', '20200501.pdc', '20200501.pfl', '20200501.pi', '20200501.pih', '20200501.pl', '20200501.pms', '20200501.pnb', '20200501.pnt', '20200501.ps', '20200501.pt', '20200501.qu', '20200501.rm', '20200501.rmy', '20200501.rn', '20200501.ro', '20200501.roa-rup', '20200501.roa-tara', '20200501.ru', '20200501.rue', '20200501.rw', '20200501.sa', '20200501.sah', '20200501.sat', '20200501.sc', '20200501.scn', '20200501.sco', '20200501.sd', '20200501.se', '20200501.sg', '20200501.sh', '20200501.si', '20200501.simple', '20200501.sk', '20200501.sl', '20200501.sm', '20200501.sn', '20200501.so', '20200501.sq', '20200501.sr', '20200501.srn', '20200501.ss', '20200501.st', '20200501.stq', '20200501.su', '20200501.sv', '20200501.sw', '20200501.szl', '20200501.ta', '20200501.tcy', '20200501.te', '20200501.tet', '20200501.tg', '20200501.th', '20200501.ti', '20200501.tk', '20200501.tl', '20200501.tn', '20200501.to', '20200501.tpi', '20200501.tr', '20200501.ts', '20200501.tt', '20200501.tum', '20200501.tw', '20200501.ty', '20200501.tyv', '20200501.udm', '20200501.ug', '20200501.uk', '20200501.ur', '20200501.uz', '20200501.ve', '20200501.vec', '20200501.vep', '20200501.vi', '20200501.vls', '20200501.vo', '20200501.wa', '20200501.war', '20200501.wo', '20200501.wuu', '20200501.xal', '20200501.xh', '20200501.xmf', '20200501.yi', '20200501.yo', '20200501.za', '20200501.zea', '20200501.zh', '20200501.zh-classical', '20200501.zh-min-nan', '20200501.zh-yue', '20200501.zu']
```
## Environment info
<!-- You can run the command `datasets-cli env` and copy-and-paste its output below. -->
- `datasets` version: 2.0.0
- Platform: Ubuntu
- Python version: 3.6
- PyArrow version: 6.0.1
| false
|
1,199,794,750
|
https://api.github.com/repos/huggingface/datasets/issues/4142
|
https://github.com/huggingface/datasets/issues/4142
| 4,142
|
Add ObjectFolder 2.0 dataset
|
open
| 1
| 2022-04-11T10:57:51
| 2022-10-05T10:30:49
| null |
osanseviero
|
[
"dataset request"
] |
## Adding a Dataset
- **Name:** ObjectFolder 2.0
- **Description:** ObjectFolder 2.0 is a dataset of 1,000 objects in the form of implicit representations. It contains 1,000 Object Files each containing the complete multisensory profile for an object instance.
- **Paper:** [*link to the dataset paper if available*](https://arxiv.org/abs/2204.02389)
- **Data:** https://github.com/rhgao/ObjectFolder
Instructions to add a new dataset can be found [here](https://github.com/huggingface/datasets/blob/master/ADD_NEW_DATASET.md).
| false
|
1,199,610,885
|
https://api.github.com/repos/huggingface/datasets/issues/4141
|
https://github.com/huggingface/datasets/issues/4141
| 4,141
|
Why is the dataset not visible under the dataset preview section?
|
closed
| 0
| 2022-04-11T08:36:42
| 2022-04-11T18:55:32
| 2022-04-11T17:09:49
|
Nid989
|
[
"dataset-viewer"
] |
## Dataset viewer issue for '*name of the dataset*'
**Link:** *link to the dataset viewer page*
*short description of the issue*
Am I the one who added this dataset ? Yes-No
| false
|
1,199,492,356
|
https://api.github.com/repos/huggingface/datasets/issues/4140
|
https://github.com/huggingface/datasets/issues/4140
| 4,140
|
Error loading arxiv data set
|
closed
| 3
| 2022-04-11T07:06:34
| 2022-04-12T16:24:08
| 2022-04-12T16:24:08
|
yjqiu
|
[
"bug"
] |
## Describe the bug
A clear and concise description of what the bug is.
I met the error below when loading arxiv dataset via `nlp.load_dataset('scientific_papers', 'arxiv',)`.
```
Traceback (most recent call last):
File "scripts/summarization.py", line 354, in <module>
main(args)
File "scripts/summarization.py", line 306, in main
model.hf_datasets = nlp.load_dataset('scientific_papers', 'arxiv')
File "/opt/conda/envs/longformer/lib/python3.7/site-packages/nlp/load.py", line 549, in load_dataset
download_config=download_config, download_mode=download_mode, ignore_verifications=ignore_verifications,
File "/opt/conda/envs/longformer/lib/python3.7/site-packages/nlp/builder.py", line 463, in download_and_prepare
dl_manager=dl_manager, verify_infos=verify_infos, **download_and_prepare_kwargs
File "/opt/conda/envs/longformer/lib/python3.7/site-packages/nlp/builder.py", line 522, in _download_and_prepare
self.info.download_checksums, dl_manager.get_recorded_sizes_checksums(), "dataset source files"
File "/opt/conda/envs/longformer/lib/python3.7/site-packages/nlp/utils/info_utils.py", line 38, in verify_checksums
raise NonMatchingChecksumError(error_msg + str(bad_urls))
nlp.utils.info_utils.NonMatchingChecksumError: Checksums didn't match for dataset source files:
['https://drive.google.com/uc?id=1b3rmCSIoh6VhD4HKWjI4HOW-cSwcwbeC&export=download', 'https://drive.google.com/uc?id=1lvsqvsFi3W-pE1SqNZI0s8NR9rC1tsja&export=download']
```
I then tried to ignore verification steps by `ignore_verifications=True` and there is another error.
```
Traceback (most recent call last):
File "/opt/conda/envs/longformer/lib/python3.7/site-packages/nlp/builder.py", line 537, in _download_and_prepare
self._prepare_split(split_generator, **prepare_split_kwargs)
File "/opt/conda/envs/longformer/lib/python3.7/site-packages/nlp/builder.py", line 810, in _prepare_split
for key, record in utils.tqdm(generator, unit=" examples", total=split_info.num_examples, leave=False):
File "/opt/conda/envs/longformer/lib/python3.7/site-packages/tqdm/std.py", line 1195, in __iter__
for obj in iterable:
File "/opt/conda/envs/longformer/lib/python3.7/site-packages/nlp/datasets/scientific_papers/9e4f2cfe3d8494e9f34a84ce49c3214605b4b52a3d8eb199104430d04c52cc12/scientific_papers.py", line 108, in _generate_examples
with open(path, encoding="utf-8") as f:
NotADirectoryError: [Errno 20] Not a directory: '/home/username/.cache/huggingface/datasets/downloads/c0deae7af7d9c87f25dfadf621f7126f708d7dcac6d353c7564883084a000076/arxiv-dataset/train.txt'
During handling of the above exception, another exception occurred:
Traceback (most recent call last):
File "scripts/summarization.py", line 354, in <module>
main(args)
File "scripts/summarization.py", line 306, in main
model.hf_datasets = nlp.load_dataset('scientific_papers', 'arxiv', ignore_verifications=True)
File "/opt/conda/envs/longformer/lib/python3.7/site-packages/nlp/load.py", line 549, in load_dataset
download_config=download_config, download_mode=download_mode, ignore_verifications=ignore_verifications,
File "/opt/conda/envs/longformer/lib/python3.7/site-packages/nlp/builder.py", line 463, in download_and_prepare
dl_manager=dl_manager, verify_infos=verify_infos, **download_and_prepare_kwargs
File "/opt/conda/envs/longformer/lib/python3.7/site-packages/nlp/builder.py", line 539, in _download_and_prepare
raise OSError("Cannot find data file. " + (self.manual_download_instructions or ""))
OSError: Cannot find data file.
```
## Steps to reproduce the bug
```python
# Sample code to reproduce the bug
```
## Expected results
A clear and concise description of the expected results.
## Actual results
Specify the actual results or traceback.
## Environment info
<!-- You can run the command `datasets-cli env` and copy-and-paste its output below. -->
- `datasets` version:
- Platform:
- Python version:
- PyArrow version:
| false
|
1,199,443,822
|
https://api.github.com/repos/huggingface/datasets/issues/4139
|
https://github.com/huggingface/datasets/issues/4139
| 4,139
|
Dataset viewer issue for Winoground
|
closed
| 11
| 2022-04-11T06:11:41
| 2022-06-21T16:43:58
| 2022-06-21T16:43:58
|
alcinos
|
[
"dataset-viewer",
"dataset-viewer-gated"
] |
## Dataset viewer issue for 'Winoground'
**Link:** [*link to the dataset viewer page*](https://huggingface.co/datasets/facebook/winoground/viewer/facebook--winoground/train)
*short description of the issue*
Getting 401, message='Unauthorized'
The dataset is subject to authorization, but I can access the files from the interface, so I assume I'm granted to access it. I'd assume the permission somehow doesn't propagate to the dataset viewer tool.
Am I the one who added this dataset ? No
| false
|
1,199,291,730
|
https://api.github.com/repos/huggingface/datasets/issues/4138
|
https://github.com/huggingface/datasets/issues/4138
| 4,138
|
Incorrect Russian filenames encoding after extraction by datasets.DownloadManager.download_and_extract()
|
closed
| 5
| 2022-04-11T02:07:13
| 2022-04-19T03:15:46
| 2022-04-16T15:46:29
|
iluvvatar
|
[] |
## Dataset viewer issue for 'MalakhovIlya/RuREBus'
**Link:** https://huggingface.co/datasets/MalakhovIlya/RuREBus
**Description**
Using os.walk(topdown=False) in DatasetBuilder causes following error:
Status code: 400
Exception: TypeError
Message: xwalk() got an unexpected keyword argument 'topdown'
Couldn't find where "xwalk" come from. How can I fix this?
Am I the one who added this dataset ? Yes
| false
|
1,199,000,453
|
https://api.github.com/repos/huggingface/datasets/issues/4137
|
https://github.com/huggingface/datasets/pull/4137
| 4,137
|
Add single dataset citations for TweetEval
|
closed
| 2
| 2022-04-10T11:51:54
| 2022-04-12T07:57:22
| 2022-04-12T07:51:15
|
gchhablani
|
[] |
This PR adds single data citations as per request of the original creators of the TweetEval dataset.
This is a recent email from the creator:
> Could I ask you a favor? Would you be able to add at the end of the README the citations of the single datasets as well? You can just copy our readme maybe? https://github.com/cardiffnlp/tweeteval#citing-tweeteval
(just to be sure that the creator of the single datasets also get credits when tweeteval is used)
Please let me know if this looks okay or if any changes are needed.
Thanks,
Gunjan
| true
|
1,198,307,610
|
https://api.github.com/repos/huggingface/datasets/issues/4135
|
https://github.com/huggingface/datasets/pull/4135
| 4,135
|
Support streaming xtreme dataset for PAN-X config
|
closed
| 1
| 2022-04-09T06:19:48
| 2022-05-06T08:39:40
| 2022-04-11T06:54:14
|
albertvillanova
|
[] |
Support streaming xtreme dataset for PAN-X config.
| true
|
1,197,937,146
|
https://api.github.com/repos/huggingface/datasets/issues/4134
|
https://github.com/huggingface/datasets/issues/4134
| 4,134
|
ELI5 supporting documents
|
open
| 1
| 2022-04-08T23:36:27
| 2022-04-13T13:52:46
| null |
saurabh-0077
|
[
"question"
] |
if i am using dense search to create supporting documents for eli5 how much time it will take bcz i read somewhere that it takes about 18 hrs??
| false
|
1,197,830,623
|
https://api.github.com/repos/huggingface/datasets/issues/4133
|
https://github.com/huggingface/datasets/issues/4133
| 4,133
|
HANS dataset preview broken
|
closed
| 3
| 2022-04-08T21:06:15
| 2022-04-13T11:57:34
| 2022-04-13T11:57:34
|
pietrolesci
|
[
"streaming"
] |
## Dataset viewer issue for '*hans*'
**Link:** [https://huggingface.co/datasets/hans](https://huggingface.co/datasets/hans)
HANS dataset preview is broken with error 400
Am I the one who added this dataset ? No
| false
|
1,197,661,720
|
https://api.github.com/repos/huggingface/datasets/issues/4132
|
https://github.com/huggingface/datasets/pull/4132
| 4,132
|
Support streaming xtreme dataset for PAWS-X config
|
closed
| 1
| 2022-04-08T18:25:32
| 2022-05-06T08:39:42
| 2022-04-08T21:02:44
|
albertvillanova
|
[] |
Support streaming xtreme dataset for PAWS-X config.
| true
|
1,197,472,249
|
https://api.github.com/repos/huggingface/datasets/issues/4131
|
https://github.com/huggingface/datasets/pull/4131
| 4,131
|
Support streaming xtreme dataset for udpos config
|
closed
| 1
| 2022-04-08T15:30:49
| 2022-05-06T08:39:46
| 2022-04-08T16:28:07
|
albertvillanova
|
[] |
Support streaming xtreme dataset for udpos config.
| true
|
1,197,456,857
|
https://api.github.com/repos/huggingface/datasets/issues/4130
|
https://github.com/huggingface/datasets/pull/4130
| 4,130
|
Add SBU Captions Photo Dataset
|
closed
| 1
| 2022-04-08T15:17:39
| 2022-04-12T10:47:31
| 2022-04-12T10:41:29
|
thomasw21
|
[] | null | true
|
1,197,376,796
|
https://api.github.com/repos/huggingface/datasets/issues/4129
|
https://github.com/huggingface/datasets/issues/4129
| 4,129
|
dataset metadata for reproducibility
|
open
| 1
| 2022-04-08T14:17:28
| 2023-09-29T09:23:56
| null |
nbroad1881
|
[
"enhancement"
] |
When pulling a dataset from the hub, it would be useful to have some metadata about the specific dataset and version that is used. The metadata could then be passed to the `Trainer` which could then be saved to a model card. This is useful for people who run many experiments on different versions (commits/branches) of the same dataset.
The dataset could have a list of “source datasets” metadata and ignore what happens to them before arriving in the Trainer (i.e. ignore mapping, filtering, etc.).
Here is a basic representation (made by @lhoestq )
```python
>>> from datasets import load_dataset
>>>
>>> my_dataset = load_dataset(...)["train"]
>>> my_dataset = my_dataset.map(...)
>>>
>>> my_dataset.sources
[HFHubDataset(repo_id=..., revision=..., arguments={...})]
```
| false
|
1,197,326,311
|
https://api.github.com/repos/huggingface/datasets/issues/4128
|
https://github.com/huggingface/datasets/pull/4128
| 4,128
|
More robust `cast_to_python_objects` in `TypedSequence`
|
closed
| 1
| 2022-04-08T13:33:35
| 2022-04-13T14:07:41
| 2022-04-13T14:01:16
|
mariosasko
|
[] |
Adds a fallback to run an expensive version of `cast_to_python_objects` which exhaustively checks entire lists to avoid the `ArrowInvalid: Could not convert` error in `TypedSequence`. Currently, this error can happen in situations where only some images are decoded in `map`, in which case `cast_to_python_objects` fails to recognize that it needs to cast `PIL.Image` objects if they are not at the beginning of the sequence and stops after the first image dictionary (e.g., if `data` is `[{"bytes": None, "path": "some path"}, PIL.Image(), ...]`)
Fix #4124
| true
|
1,197,297,756
|
https://api.github.com/repos/huggingface/datasets/issues/4127
|
https://github.com/huggingface/datasets/pull/4127
| 4,127
|
Add configs with processed data in medical_dialog dataset
|
closed
| 1
| 2022-04-08T13:08:16
| 2022-05-06T08:39:50
| 2022-04-08T16:20:51
|
albertvillanova
|
[] |
There exist processed data files that do not require parsing the raw data files (which can take long time).
Fix #4122.
| true
|
1,196,665,194
|
https://api.github.com/repos/huggingface/datasets/issues/4126
|
https://github.com/huggingface/datasets/issues/4126
| 4,126
|
dataset viewer issue for common_voice
|
closed
| 2
| 2022-04-07T23:34:28
| 2022-04-25T13:42:17
| 2022-04-25T13:42:16
|
laphang
|
[
"dataset-viewer",
"audio_column"
] |
## Dataset viewer issue for 'common_voice'
**Link:** https://huggingface.co/datasets/common_voice
Server Error
Status code: 400
Exception: TypeError
Message: __init__() got an unexpected keyword argument 'audio_column'
Am I the one who added this dataset ? No
| false
|
1,196,633,936
|
https://api.github.com/repos/huggingface/datasets/issues/4125
|
https://github.com/huggingface/datasets/pull/4125
| 4,125
|
BIG-bench
|
closed
| 21
| 2022-04-07T22:33:30
| 2022-06-08T17:57:48
| 2022-06-08T17:32:32
|
andersjohanandreassen
|
[] |
This PR adds all BIG-bench json tasks to huggingface/datasets.
| true
|
1,196,469,842
|
https://api.github.com/repos/huggingface/datasets/issues/4124
|
https://github.com/huggingface/datasets/issues/4124
| 4,124
|
Image decoding often fails when transforming Image datasets
|
closed
| 7
| 2022-04-07T19:17:25
| 2022-04-13T14:01:16
| 2022-04-13T14:01:16
|
RafayAK
|
[
"bug"
] |
## Describe the bug
When transforming/modifying images in an image dataset using the `map` function the PIL images often fail to decode in time for the image transforms, causing errors.
Using a debugger it is easy to see what the problem is, the Image decode invocation does not take place and the resulting image passed around is still raw bytes:
```
[{'bytes': b'\x89PNG\r\n\x1a\n\x00\x00\x00\rIHDR\x00\x00\x00 \x00\x00\x00 \x08\x02\x00\x00\x00\xfc\x18\xed\xa3\x00\x00\x08\x02IDATx\x9cEVIs[\xc7\x11\xeemf\xde\x82\x8d\x80\x08\x89"\xb5V\\\xb6\x94(\xe5\x9f\x90\xca5\x7f$\xa7T\xe5\x9f&9\xd9\x8a\\.\xdb\xa4$J\xa4\x00\x02x\xc0{\xb3t\xe7\x00\xca\x99\xd3\\f\xba\xba\xbf\xa5?|\xfa\xf4\xa2\xeb\xba\xedv\xa3f^\xf8\xd5\x0bY\xb6\x10\xb3\xaaDq\xcd\x83\x87\xdf5\xf3gZ\x1a\x04\x0f\xa0fp\xfa\xe0\xd4\x07?\x9dN\xc4\xb1\x99\xfd\xf2\xcb/\x97\x97\x97H\xa2\xaaf\x16\x82\xaf\xeb\xca{\xbf\xd9l.\xdf\x7f\xfa\xcb_\xff&\x88\x08\x00\x80H\xc0\x80@.;\x0f\x8c@#v\xe3\xe5\xfc\xd1\x9f\xee6q\xbf\xdf\xa6\x14\'\x93\xf1\xc3\xe5\xe3\xd1x\x14c\x8c1\xa5\x1c\x9dsM\xd3\xb4\xed\x08\x89SJ)\xa5\xedv\xbb^\xafNO\x97D\x84Hf ....
```
## Steps to reproduce the bug
```python
from datasets import load_dataset, Dataset
import numpy as np
# seeded NumPy random number generator for reprodducinble results.
rng = np.random.default_rng(seed=0)
test_dataset = load_dataset('cifar100', split="test")
def preprocess_data(dataset):
"""
Helper function to pre-process HuggingFace Cifar-100 Dataset to remove fine_label and coarse_label columns and
add is_flipped column
Args:
dataset: HuggingFace CIFAR-100 Dataset Object
Returns:
new_dataset: A Dataset object with "img" and "is_flipped" columns only
"""
# remove fine_label and coarse_label columns
new_dataset = dataset.remove_columns(['fine_label', 'coarse_label'])
# add the column for is_flipped
new_dataset = new_dataset.add_column(name="is_flipped", column=np.zeros((len(new_dataset)), dtype=np.uint8))
return new_dataset
def generate_flipped_data(example, p=0.5):
"""
A Dataset mapping function that transforms some of the images up-side-down.
If the probability value (p) is 0.5 approximately half the images will be flipped upside-down
Args:
example: An example from the dataset containing a Python dictionary with "img" and "is_flipped" key-value pair
p: the probability of flipping the image up-side-down, Default 0.5
Returns:
example: A Dataset object
"""
# example['img'] = example['img']
if rng.random() > p: # the flip the image and set is_flipped column to 1
example['img'] = example['img'].transpose(
1) # ImageOps.flip(example['img']) #example['img'].transpose(Image.FLIP_TOP_BOTTOM)
example['is_flipped'] = 1
return example
my_test = preprocess_data(test_dataset)
my_test = my_test.map(generate_flipped_data)
```
## Expected results
The dataset should be transformed without problems.
## Actual results
```
/home/rafay/anaconda3/envs/pytorch_new/bin/python /home/rafay/Documents/you_only_live_once/upside_down_detector/create_dataset.py
Reusing dataset cifar100 (/home/rafay/.cache/huggingface/datasets/cifar100/cifar100/1.0.0/f365c8b725c23e8f0f8d725c3641234d9331cd2f62919d1381d1baa5b3ba3142)
Reusing dataset cifar100 (/home/rafay/.cache/huggingface/datasets/cifar100/cifar100/1.0.0/f365c8b725c23e8f0f8d725c3641234d9331cd2f62919d1381d1baa5b3ba3142)
20%|█▉ | 1999/10000 [00:00<00:01, 5560.44ex/s]
Traceback (most recent call last):
File "/home/rafay/anaconda3/envs/pytorch_new/lib/python3.10/site-packages/datasets/arrow_dataset.py", line 2326, in _map_single
writer.write(example)
File "/home/rafay/anaconda3/envs/pytorch_new/lib/python3.10/site-packages/datasets/arrow_writer.py", line 441, in write
self.write_examples_on_file()
File "/home/rafay/anaconda3/envs/pytorch_new/lib/python3.10/site-packages/datasets/arrow_writer.py", line 399, in write_examples_on_file
self.write_batch(batch_examples=batch_examples)
File "/home/rafay/anaconda3/envs/pytorch_new/lib/python3.10/site-packages/datasets/arrow_writer.py", line 492, in write_batch
arrays.append(pa.array(typed_sequence))
File "pyarrow/array.pxi", line 230, in pyarrow.lib.array
File "pyarrow/array.pxi", line 110, in pyarrow.lib._handle_arrow_array_protocol
File "/home/rafay/anaconda3/envs/pytorch_new/lib/python3.10/site-packages/datasets/arrow_writer.py", line 185, in __arrow_array__
out = pa.array(cast_to_python_objects(data, only_1d_for_numpy=True))
File "pyarrow/array.pxi", line 316, in pyarrow.lib.array
File "pyarrow/array.pxi", line 39, in pyarrow.lib._sequence_to_array
File "pyarrow/error.pxi", line 143, in pyarrow.lib.pyarrow_internal_check_status
File "pyarrow/error.pxi", line 99, in pyarrow.lib.check_status
pyarrow.lib.ArrowInvalid: Could not convert <PIL.Image.Image image mode=RGB size=32x32 at 0x7F56AEE61DE0> with type Image: did not recognize Python value type when inferring an Arrow data type
During handling of the above exception, another exception occurred:
Traceback (most recent call last):
File "/home/rafay/Documents/you_only_live_once/upside_down_detector/create_dataset.py", line 55, in <module>
my_test = my_test.map(generate_flipped_data)
File "/home/rafay/anaconda3/envs/pytorch_new/lib/python3.10/site-packages/datasets/arrow_dataset.py", line 1953, in map
return self._map_single(
File "/home/rafay/anaconda3/envs/pytorch_new/lib/python3.10/site-packages/datasets/arrow_dataset.py", line 519, in wrapper
out: Union["Dataset", "DatasetDict"] = func(self, *args, **kwargs)
File "/home/rafay/anaconda3/envs/pytorch_new/lib/python3.10/site-packages/datasets/arrow_dataset.py", line 486, in wrapper
out: Union["Dataset", "DatasetDict"] = func(self, *args, **kwargs)
File "/home/rafay/anaconda3/envs/pytorch_new/lib/python3.10/site-packages/datasets/fingerprint.py", line 458, in wrapper
out = func(self, *args, **kwargs)
File "/home/rafay/anaconda3/envs/pytorch_new/lib/python3.10/site-packages/datasets/arrow_dataset.py", line 2360, in _map_single
writer.finalize()
File "/home/rafay/anaconda3/envs/pytorch_new/lib/python3.10/site-packages/datasets/arrow_writer.py", line 522, in finalize
self.write_examples_on_file()
File "/home/rafay/anaconda3/envs/pytorch_new/lib/python3.10/site-packages/datasets/arrow_writer.py", line 399, in write_examples_on_file
self.write_batch(batch_examples=batch_examples)
File "/home/rafay/anaconda3/envs/pytorch_new/lib/python3.10/site-packages/datasets/arrow_writer.py", line 492, in write_batch
arrays.append(pa.array(typed_sequence))
File "pyarrow/array.pxi", line 230, in pyarrow.lib.array
File "pyarrow/array.pxi", line 110, in pyarrow.lib._handle_arrow_array_protocol
File "/home/rafay/anaconda3/envs/pytorch_new/lib/python3.10/site-packages/datasets/arrow_writer.py", line 185, in __arrow_array__
out = pa.array(cast_to_python_objects(data, only_1d_for_numpy=True))
File "pyarrow/array.pxi", line 316, in pyarrow.lib.array
File "pyarrow/array.pxi", line 39, in pyarrow.lib._sequence_to_array
File "pyarrow/error.pxi", line 143, in pyarrow.lib.pyarrow_internal_check_status
File "pyarrow/error.pxi", line 99, in pyarrow.lib.check_status
pyarrow.lib.ArrowInvalid: Could not convert <PIL.Image.Image image mode=RGB size=32x32 at 0x7F56AEE61DE0> with type Image: did not recognize Python value type when inferring an Arrow data type
Process finished with exit code 1
```
## Environment info
<!-- You can run the command `datasets-cli env` and copy-and-paste its output below. -->
- `datasets` version: 2.0.0
- Platform: Linux(Fedora 35)
- Python version: 3.10
- PyArrow version: 7.0.0
| false
|
1,196,367,512
|
https://api.github.com/repos/huggingface/datasets/issues/4123
|
https://github.com/huggingface/datasets/issues/4123
| 4,123
|
Building C4 takes forever
|
closed
| 1
| 2022-04-07T17:41:30
| 2023-06-26T22:01:29
| 2023-06-26T22:01:29
|
StellaAthena
|
[
"bug"
] |
## Describe the bug
C4-en is a 300 GB dataset. However, when I try to download it through the hub it takes over _six hours_ to generate the train/test split from the downloaded files. This is an absurd amount of time and an unnecessary waste of resources.
## Steps to reproduce the bug
```python
c4 = datasets.load("c4", "en")
```
## Expected results
I would like to be able to download pre-split data.
## Environment info
- `datasets` version: 2.0.0
- Platform: Linux-5.13.0-35-generic-x86_64-with-glibc2.29
- Python version: 3.8.10
- PyArrow version: 7.0.0
- Pandas version: 1.4.1
| false
|
1,196,095,072
|
https://api.github.com/repos/huggingface/datasets/issues/4122
|
https://github.com/huggingface/datasets/issues/4122
| 4,122
|
medical_dialog zh has very slow _generate_examples
|
closed
| 3
| 2022-04-07T14:00:51
| 2022-04-08T16:20:51
| 2022-04-08T16:20:51
|
nbroad1881
|
[
"bug"
] |
## Describe the bug
After downloading the files from Google Drive, `load_dataset("medical_dialog", "zh", data_dir="./")` takes an unreasonable amount of time. Generating the train/test split for 33% of the dataset takes over 4.5 hours.
## Steps to reproduce the bug
The easiest way I've found to download files from Google Drive is to use `gdown` and use Google Colab because the download speeds will be very high due to the fact that they are both in Google Cloud.
```python
file_ids = [
"1AnKxGEuzjeQsDHHqL3NqI_aplq2hVL_E",
"1tt7weAT1SZknzRFyLXOT2fizceUUVRXX",
"1A64VBbsQ_z8wZ2LDox586JIyyO6mIwWc",
"1AKntx-ECnrxjB07B6BlVZcFRS4YPTB-J",
"1xUk8AAua_x27bHUr-vNoAuhEAjTxOvsu",
"1ezKTfe7BgqVN5o-8Vdtr9iAF0IueCSjP",
"1tA7bSOxR1RRNqZst8cShzhuNHnayUf7c",
"1pA3bCFA5nZDhsQutqsJcH3d712giFb0S",
"1pTLFMdN1A3ro-KYghk4w4sMz6aGaMOdU",
"1dUSnG0nUPq9TEQyHd6ZWvaxO0OpxVjXD",
"1UfCH05nuWiIPbDZxQzHHGAHyMh8dmPQH",
]
for i in file_ids:
url = f"https://drive.google.com/uc?id={i}"
!gdown $url
from datasets import load_dataset
ds = load_dataset("medical_dialog", "zh", data_dir="./")
```
## Expected results
Faster load time
## Actual results
`Generating train split: 33%: 625519/1921127 [4:31:03<31:39:20, 11.37 examples/s]`
## Environment info
- `datasets` version: 2.0.0
- Platform: Linux-5.4.144+-x86_64-with-Ubuntu-18.04-bionic
- Python version: 3.7.13
- PyArrow version: 6.0.1
- Pandas version: 1.3.5
@vrindaprabhu , could you take a look at this since you implemented it? I think the `_generate_examples` function might need to be rewritten
| false
|
1,196,000,018
|
https://api.github.com/repos/huggingface/datasets/issues/4121
|
https://github.com/huggingface/datasets/issues/4121
| 4,121
|
datasets.load_metric can not load a local metirc
|
closed
| 1
| 2022-04-07T12:48:56
| 2023-01-18T14:30:46
| 2022-04-07T13:53:27
|
SadGare
|
[
"bug"
] |
## Describe the bug
No matter how I hard try to tell load_metric that I want to load a local metric file, it still continues to fetch things on the Internet. And unfortunately it says 'ConnectionError: Couldn't reach'. However I can download this file without connectionerror and tell load_metric its local directory. And it comes back where it begins...
## Steps to reproduce the bug
```python
metric = load_metric(path=r'C:\Users\Gare\PycharmProjects\Gare\blue\bleu.py')
ConnectionError: Couldn't reach https://github.com/tensorflow/nmt/raw/master/nmt/scripts/bleu.py
metric = load_metric(path='bleu')
ConnectionError: Couldn't reach https://raw.githubusercontent.com/huggingface/datasets/1.12.1/metrics/bleu/bleu.py
metric = load_metric(path='./blue/bleu.py')
ConnectionError: Couldn't reach https://github.com/tensorflow/nmt/raw/master/nmt/scripts/bleu.py
```
## Expected results
I do read the docs [here](https://huggingface.co/docs/datasets/package_reference/loading_methods#datasets.load_metric). There are no other parameters that help function to distinguish from local and online file but path. As what I code above, it should load from local.
## Actual results
> metric = load_metric(path=r'C:\Users\Gare\PycharmProjects\Gare\blue\bleu.py')
> ~\AppData\Local\Temp\ipykernel_19636\1855752034.py in <module>
----> 1 metric = load_metric(path=r'C:\Users\Gare\PycharmProjects\Gare\blue\bleu.py')
D:\Program Files\Anaconda\envs\Gare\lib\site-packages\datasets\load.py in load_metric(path, config_name, process_id, num_process, cache_dir, experiment_id, keep_in_memory, download_config, download_mode, script_version, **metric_init_kwargs)
817 if data_files is None and data_dir is not None:
818 data_files = os.path.join(data_dir, "**")
--> 819
820 self.name = name
821 self.revision = revision
D:\Program Files\Anaconda\envs\Gare\lib\site-packages\datasets\load.py in prepare_module(path, script_version, download_config, download_mode, dataset, force_local_path, dynamic_modules_path, return_resolved_file_path, return_associated_base_path, data_files, **download_kwargs)
639 self,
640 path: str,
--> 641 download_config: Optional[DownloadConfig] = None,
642 download_mode: Optional[DownloadMode] = None,
643 dynamic_modules_path: Optional[str] = None,
D:\Program Files\Anaconda\envs\Gare\lib\site-packages\datasets\utils\file_utils.py in cached_path(url_or_filename, download_config, **download_kwargs)
297 token = hf_api.HfFolder.get_token()
298 if token:
--> 299 headers["authorization"] = f"Bearer {token}"
300 return headers
301
D:\Program Files\Anaconda\envs\Gare\lib\site-packages\datasets\utils\file_utils.py in get_from_cache(url, cache_dir, force_download, proxies, etag_timeout, resume_download, user_agent, local_files_only, use_etag, max_retries, use_auth_token)
604 def _resumable_file_manager():
605 with open(incomplete_path, "a+b") as f:
--> 606 yield f
607
608 temp_file_manager = _resumable_file_manager
ConnectionError: Couldn't reach https://github.com/tensorflow/nmt/raw/master/nmt/scripts/bleu.py
## Environment info
<!-- You can run the command `datasets-cli env` and copy-and-paste its output below. -->
- `datasets` version: 2.0.0
- Platform: Windows-10-10.0.22000-SP0
- Python version: 3.7.13
- PyArrow version: 7.0.0
- Pandas version: 1.3.4
Any advice would be appreciated.
| false
|
1,195,887,430
|
https://api.github.com/repos/huggingface/datasets/issues/4120
|
https://github.com/huggingface/datasets/issues/4120
| 4,120
|
Representing dictionaries (json) objects as features
|
open
| 0
| 2022-04-07T11:07:41
| 2022-04-07T11:07:41
| null |
yanaiela
|
[
"enhancement"
] |
In the process of adding a new dataset to the hub, I stumbled upon the inability to represent dictionaries that contain different key names, unknown in advance (and may differ between samples), original asked in the [forum](https://discuss.huggingface.co/t/representing-nested-dictionary-with-different-keys/16442).
For instance:
```
sample1 = {"nps": {
"a": {"id": 0, "text": "text1"},
"b": {"id": 1, "text": "text2"},
}}
sample2 = {"nps": {
"a": {"id": 0, "text": "text1"},
"b": {"id": 1, "text": "text2"},
"c": {"id": 2, "text": "text3"},
}}
sample3 = {"nps": {
"a": {"id": 0, "text": "text1"},
"b": {"id": 1, "text": "text2"},
"c": {"id": 2, "text": "text3"},
"d": {"id": 3, "text": "text4"},
}}
```
the `nps` field cannot be represented as a Feature while maintaining its original structure.
@lhoestq suggested to add JSON as a new feature type, which will solve this problem.
It seems like an alternative solution would be to change the original data format, which isn't an optimal solution in my case. Moreover, JSON is a common structure, that will likely to be useful in future datasets as well.
| false
|
1,195,641,298
|
https://api.github.com/repos/huggingface/datasets/issues/4119
|
https://github.com/huggingface/datasets/pull/4119
| 4,119
|
Hotfix failing CI tests on Windows
|
closed
| 1
| 2022-04-07T07:38:46
| 2022-04-07T09:47:24
| 2022-04-07T07:57:13
|
albertvillanova
|
[] |
This PR makes a hotfix for our CI Windows tests: https://app.circleci.com/pipelines/github/huggingface/datasets/11092/workflows/9cfdb1dd-0fec-4fe0-8122-5f533192ebdc/jobs/67414
Fix #4118
I guess this issue is related to this PR:
- huggingface/huggingface_hub#815
| true
|
1,195,638,944
|
https://api.github.com/repos/huggingface/datasets/issues/4118
|
https://github.com/huggingface/datasets/issues/4118
| 4,118
|
Failing CI tests on Windows
|
closed
| 0
| 2022-04-07T07:36:25
| 2022-04-07T07:57:13
| 2022-04-07T07:57:13
|
albertvillanova
|
[
"bug"
] |
## Describe the bug
Our CI Windows tests are failing from yesterday: https://app.circleci.com/pipelines/github/huggingface/datasets/11092/workflows/9cfdb1dd-0fec-4fe0-8122-5f533192ebdc/jobs/67414
| false
|
1,195,552,406
|
https://api.github.com/repos/huggingface/datasets/issues/4117
|
https://github.com/huggingface/datasets/issues/4117
| 4,117
|
AttributeError: module 'huggingface_hub' has no attribute 'hf_api'
|
closed
| 13
| 2022-04-07T05:52:36
| 2024-05-07T09:24:35
| 2022-04-19T15:36:35
|
arymbe
|
[
"bug"
] |
## Describe the bug
Could you help me please. I got this following error.
AttributeError: module 'huggingface_hub' has no attribute 'hf_api'
## Steps to reproduce the bug
when I imported the datasets
# Sample code to reproduce the bug
from datasets import list_datasets, load_dataset, list_metrics, load_metric
## Environment info
<!-- You can run the command `datasets-cli env` and copy-and-paste its output below. -->
- `datasets` version: 2.0.0
- Platform: macOS-12.3-x86_64-i386-64bit
- Python version: 3.8.9
- PyArrow version: 7.0.0
- Pandas version: 1.3.5
- Huggingface-hub: 0.5.0
- Transformers: 4.18.0
Thank you in advance.
| false
|
1,194,926,459
|
https://api.github.com/repos/huggingface/datasets/issues/4116
|
https://github.com/huggingface/datasets/pull/4116
| 4,116
|
Pretty print dataset info files
|
closed
| 5
| 2022-04-06T17:40:48
| 2022-04-08T11:28:01
| 2022-04-08T11:21:53
|
mariosasko
|
[] |
Adds indentation to the `dataset_infos.json` file when saving for nicer diffs.
(suggested by @julien-c)
This PR also updates the info files of the GH datasets. Note that this change adds more than **10 MB** to the repo size (the total file size before the change: 29.672298 MB, after: 41.666475 MB), so I'm not sure this change is a good idea.
`src/datasets/info.py` is the only relevant file for reviewers.
| true
|
1,194,907,555
|
https://api.github.com/repos/huggingface/datasets/issues/4115
|
https://github.com/huggingface/datasets/issues/4115
| 4,115
|
ImageFolder add option to ignore some folders like '.ipynb_checkpoints'
|
closed
| 5
| 2022-04-06T17:29:43
| 2022-06-01T13:04:16
| 2022-06-01T13:04:16
|
cceyda
|
[
"enhancement"
] |
**Is your feature request related to a problem? Please describe.**
I sometimes like to peek at the dataset images from jupyterlab. thus '.ipynb_checkpoints' folder appears where my dataset is and (just realized) leads to accidental duplicate image additions. I think this is an easy enough thing to miss especially if the dataset is very large.
**Describe the solution you'd like**
maybe have an option `ignore` or something .gitignore style
`dataset = load_dataset("imagefolder", data_dir="./data/original", ignore="regex?")`
**Describe alternatives you've considered**
Could filter out manually
| false
|
1,194,855,345
|
https://api.github.com/repos/huggingface/datasets/issues/4114
|
https://github.com/huggingface/datasets/issues/4114
| 4,114
|
Allow downloading just some columns of a dataset
|
open
| 14
| 2022-04-06T16:38:46
| 2025-02-17T15:10:56
| null |
osanseviero
|
[
"enhancement"
] |
**Is your feature request related to a problem? Please describe.**
Some people are interested in doing label analysis of a CV dataset without downloading all the images. Downloading the whole dataset does not always makes sense for this kind of use case
**Describe the solution you'd like**
Be able to just download some columns of a dataset, such as doing
```python
load_dataset("huggan/wikiart",columns=["artist", "genre"])
```
Although this might make things a bit complicated in terms of local caching of datasets.
| false
|
1,194,843,532
|
https://api.github.com/repos/huggingface/datasets/issues/4113
|
https://github.com/huggingface/datasets/issues/4113
| 4,113
|
Multiprocessing with FileLock fails in python 3.9
|
closed
| 1
| 2022-04-06T16:27:09
| 2022-11-28T11:49:14
| 2022-11-28T11:49:14
|
lhoestq
|
[
"bug"
] |
On python 3.9, this code hangs:
```python
from multiprocessing import Pool
from filelock import FileLock
def run(i):
print(f"got the lock in multi process [{i}]")
with FileLock("tmp.lock"):
with Pool(2) as pool:
pool.map(run, range(2))
```
This is because the subprocesses try to acquire the lock from the main process for some reason. This is not the case in older versions of python.
This can cause many issues in python 3.9. In particular, we use multiprocessing to fetch data files when you load a dataset (as long as there are >16 data files). Therefore `imagefolder` hangs, and I expect any dataset that needs to download >16 files to hang as well.
Let's see if we can fix this and have a CI that runs on 3.9.
cc @mariosasko @julien-c
| false
|
1,194,752,765
|
https://api.github.com/repos/huggingface/datasets/issues/4112
|
https://github.com/huggingface/datasets/issues/4112
| 4,112
|
ImageFolder with Grayscale images dataset
|
closed
| 3
| 2022-04-06T15:10:00
| 2022-04-22T10:21:53
| 2022-04-22T10:21:52
|
chainyo
|
[] |
Hi, I'm facing a problem with a grayscale images dataset I have uploaded [here](https://huggingface.co/datasets/ChainYo/rvl-cdip) (RVL-CDIP)
I'm getting an error while I want to use images for training a model with PyTorch DataLoader. Here is the full traceback:
```bash
AttributeError: Caught AttributeError in DataLoader worker process 0.
Original Traceback (most recent call last):
File "/home/chainyo/miniconda3/envs/gan-bird/lib/python3.8/site-packages/torch/utils/data/_utils/worker.py", line 287, in _worker_loop
data = fetcher.fetch(index)
File "/home/chainyo/miniconda3/envs/gan-bird/lib/python3.8/site-packages/torch/utils/data/_utils/fetch.py", line 49, in fetch
data = [self.dataset[idx] for idx in possibly_batched_index]
File "/home/chainyo/miniconda3/envs/gan-bird/lib/python3.8/site-packages/torch/utils/data/_utils/fetch.py", line 49, in <listcomp>
data = [self.dataset[idx] for idx in possibly_batched_index]
File "/home/chainyo/miniconda3/envs/gan-bird/lib/python3.8/site-packages/datasets/arrow_dataset.py", line 1765, in __getitem__
return self._getitem(
File "/home/chainyo/miniconda3/envs/gan-bird/lib/python3.8/site-packages/datasets/arrow_dataset.py", line 1750, in _getitem
formatted_output = format_table(
File "/home/chainyo/miniconda3/envs/gan-bird/lib/python3.8/site-packages/datasets/formatting/formatting.py", line 532, in format_table
return formatter(pa_table, query_type=query_type)
File "/home/chainyo/miniconda3/envs/gan-bird/lib/python3.8/site-packages/datasets/formatting/formatting.py", line 281, in __call__
return self.format_row(pa_table)
File "/home/chainyo/miniconda3/envs/gan-bird/lib/python3.8/site-packages/datasets/formatting/torch_formatter.py", line 58, in format_row
return self.recursive_tensorize(row)
File "/home/chainyo/miniconda3/envs/gan-bird/lib/python3.8/site-packages/datasets/formatting/torch_formatter.py", line 54, in recursive_tensorize
return map_nested(self._recursive_tensorize, data_struct, map_list=False)
File "/home/chainyo/miniconda3/envs/gan-bird/lib/python3.8/site-packages/datasets/utils/py_utils.py", line 314, in map_nested
mapped = [
File "/home/chainyo/miniconda3/envs/gan-bird/lib/python3.8/site-packages/datasets/utils/py_utils.py", line 315, in <listcomp>
_single_map_nested((function, obj, types, None, True, None))
File "/home/chainyo/miniconda3/envs/gan-bird/lib/python3.8/site-packages/datasets/utils/py_utils.py", line 267, in _single_map_nested
return {k: _single_map_nested((function, v, types, None, True, None)) for k, v in pbar}
File "/home/chainyo/miniconda3/envs/gan-bird/lib/python3.8/site-packages/datasets/utils/py_utils.py", line 267, in <dictcomp>
return {k: _single_map_nested((function, v, types, None, True, None)) for k, v in pbar}
File "/home/chainyo/miniconda3/envs/gan-bird/lib/python3.8/site-packages/datasets/utils/py_utils.py", line 251, in _single_map_nested
return function(data_struct)
File "/home/chainyo/miniconda3/envs/gan-bird/lib/python3.8/site-packages/datasets/formatting/torch_formatter.py", line 51, in _recursive_tensorize
return self._tensorize(data_struct)
File "/home/chainyo/miniconda3/envs/gan-bird/lib/python3.8/site-packages/datasets/formatting/torch_formatter.py", line 38, in _tensorize
if np.issubdtype(value.dtype, np.integer):
AttributeError: 'bytes' object has no attribute 'dtype'
```
I don't really understand why the image is still a bytes object while I used transformations on it. Here the code I used to upload the dataset (and it worked well):
```python
train_dataset = load_dataset("imagefolder", data_dir="data/train")
train_dataset = train_dataset["train"]
test_dataset = load_dataset("imagefolder", data_dir="data/test")
test_dataset = test_dataset["train"]
val_dataset = load_dataset("imagefolder", data_dir="data/val")
val_dataset = val_dataset["train"]
dataset = DatasetDict({
"train": train_dataset,
"val": val_dataset,
"test": test_dataset
})
dataset.push_to_hub("ChainYo/rvl-cdip")
```
Now here is the code I am using to get the dataset and prepare it for training:
```python
img_size = 512
batch_size = 128
normalize = [(0.5), (0.5)]
data_dir = "ChainYo/rvl-cdip"
dataset = load_dataset(data_dir, split="train")
transforms = transforms.Compose([
transforms.Resize(img_size),
transforms.CenterCrop(img_size),
transforms.ToTensor(),
transforms.Normalize(*normalize)
])
transformed_dataset = dataset.with_transform(transforms)
transformed_dataset.set_format(type="torch", device="cuda")
train_dataloader = torch.utils.data.DataLoader(
transformed_dataset, batch_size=batch_size, shuffle=True, num_workers=4, pin_memory=True
)
```
But this get me the error above. I don't understand why it's doing this kind of weird thing?
Do I need to map something on the dataset? Something like this:
```python
labels = dataset.features["label"].names
num_labels = dataset.features["label"].num_classes
def preprocess_data(examples):
images = [ex.convert("RGB") for ex in examples["image"]]
labels = [ex for ex in examples["label"]]
return {"images": images, "labels": labels}
features = Features({
"images": Image(decode=True, id=None),
"labels": ClassLabel(num_classes=num_labels, names=labels)
})
decoded_dataset = dataset.map(preprocess_data, remove_columns=dataset.column_names, features=features, batched=True, batch_size=100)
```
| false
|
1,194,660,699
|
https://api.github.com/repos/huggingface/datasets/issues/4111
|
https://github.com/huggingface/datasets/pull/4111
| 4,111
|
Update security policy
|
closed
| 1
| 2022-04-06T13:59:51
| 2022-04-07T09:46:30
| 2022-04-07T09:40:27
|
albertvillanova
|
[] | null | true
|
1,194,581,375
|
https://api.github.com/repos/huggingface/datasets/issues/4110
|
https://github.com/huggingface/datasets/pull/4110
| 4,110
|
Matthews Correlation Metric Card
|
closed
| 1
| 2022-04-06T12:59:35
| 2022-05-03T13:43:17
| 2022-05-03T13:36:13
|
emibaylor
|
[] | null | true
|
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.