title stringlengths 1 290 | body stringlengths 0 228k ⌀ | html_url stringlengths 46 51 | comments list | pull_request dict | number int64 1 5.59k | is_pull_request bool 2 classes |
|---|---|---|---|---|---|---|
check if trasnformers has PreTrainedTokenizerBase | Fix #598 | https://github.com/huggingface/datasets/pull/601 | [] | {
"url": "https://api.github.com/repos/huggingface/datasets/pulls/601",
"html_url": "https://github.com/huggingface/datasets/pull/601",
"diff_url": "https://github.com/huggingface/datasets/pull/601.diff",
"patch_url": "https://github.com/huggingface/datasets/pull/601.patch",
"merged_at": "2020-09-10T11:01:36"
} | 601 | true |
Pickling error when loading dataset | Hi,
I modified line 136 in the original [run_language_modeling.py](https://github.com/huggingface/transformers/blob/master/examples/language-modeling/run_language_modeling.py) as:
```
# line 136: return LineByLineTextDataset(tokenizer=tokenizer, file_path=file_path, block_size=args.block_size)
dataset = load_dataset("text", data_files=file_path, split="train")
dataset = dataset.map(lambda ex: tokenizer(ex["text"], add_special_tokens=True,
truncation=True, max_length=args.block_size), batched=True)
dataset.set_format(type='torch', columns=['input_ids'])
return dataset
```
When I run this with transformers (3.1.0) and nlp (0.4.0), I get the following error:
```
Traceback (most recent call last):
File "src/run_language_modeling.py", line 319, in <module>
main()
File "src/run_language_modeling.py", line 248, in main
get_dataset(data_args, tokenizer=tokenizer, cache_dir=model_args.cache_dir) if training_args.do_train else None
File "src/run_language_modeling.py", line 139, in get_dataset
dataset = dataset.map(lambda ex: tokenizer(ex["text"], add_special_tokens=True, truncation=True, max_length=args.block_size), batched=True)
File "/data/nlp/src/nlp/arrow_dataset.py", line 1136, in map
new_fingerprint=new_fingerprint,
File "/data/nlp/src/nlp/fingerprint.py", line 158, in wrapper
self._fingerprint, transform, kwargs_for_fingerprint
File "/data/nlp/src/nlp/fingerprint.py", line 105, in update_fingerprint
hasher.update(transform_args[key])
File "/data/nlp/src/nlp/fingerprint.py", line 57, in update
self.m.update(self.hash(value).encode("utf-8"))
File "/data/nlp/src/nlp/fingerprint.py", line 53, in hash
return cls.hash_default(value)
File "/data/nlp/src/nlp/fingerprint.py", line 46, in hash_default
return cls.hash_bytes(dumps(value))
File "/data/nlp/src/nlp/utils/py_utils.py", line 362, in dumps
dump(obj, file)
File "/data/nlp/src/nlp/utils/py_utils.py", line 339, in dump
Pickler(file, recurse=True).dump(obj)
File "/root/miniconda3/envs/py3.6/lib/python3.6/site-packages/dill/_dill.py", line 446, in dump
StockPickler.dump(self, obj)
File "/root/miniconda3/envs/py3.6/lib/python3.6/pickle.py", line 409, in dump
self.save(obj)
File "/root/miniconda3/envs/py3.6/lib/python3.6/pickle.py", line 476, in save
f(self, obj) # Call unbound method with explicit self
File "/root/miniconda3/envs/py3.6/lib/python3.6/site-packages/dill/_dill.py", line 1438, in save_function
obj.__dict__, fkwdefaults), obj=obj)
File "/root/miniconda3/envs/py3.6/lib/python3.6/pickle.py", line 610, in save_reduce
save(args)
File "/root/miniconda3/envs/py3.6/lib/python3.6/pickle.py", line 476, in save
f(self, obj) # Call unbound method with explicit self
File "/root/miniconda3/envs/py3.6/lib/python3.6/pickle.py", line 751, in save_tuple
save(element)
File "/root/miniconda3/envs/py3.6/lib/python3.6/pickle.py", line 476, in save
f(self, obj) # Call unbound method with explicit self
File "/root/miniconda3/envs/py3.6/lib/python3.6/pickle.py", line 736, in save_tuple
save(element)
File "/root/miniconda3/envs/py3.6/lib/python3.6/pickle.py", line 476, in save
f(self, obj) # Call unbound method with explicit self
File "/root/miniconda3/envs/py3.6/lib/python3.6/site-packages/dill/_dill.py", line 1170, in save_cell
pickler.save_reduce(_create_cell, (f,), obj=obj)
File "/root/miniconda3/envs/py3.6/lib/python3.6/pickle.py", line 610, in save_reduce
save(args)
File "/root/miniconda3/envs/py3.6/lib/python3.6/pickle.py", line 476, in save
f(self, obj) # Call unbound method with explicit self
File "/root/miniconda3/envs/py3.6/lib/python3.6/pickle.py", line 736, in save_tuple
save(element)
File "/root/miniconda3/envs/py3.6/lib/python3.6/pickle.py", line 521, in save
self.save_reduce(obj=obj, *rv)
File "/root/miniconda3/envs/py3.6/lib/python3.6/pickle.py", line 605, in save_reduce
save(cls)
File "/root/miniconda3/envs/py3.6/lib/python3.6/pickle.py", line 476, in save
f(self, obj) # Call unbound method with explicit self
File "/root/miniconda3/envs/py3.6/lib/python3.6/site-packages/dill/_dill.py", line 1365, in save_type
obj.__bases__, _dict), obj=obj)
File "/root/miniconda3/envs/py3.6/lib/python3.6/pickle.py", line 610, in save_reduce
save(args)
File "/root/miniconda3/envs/py3.6/lib/python3.6/pickle.py", line 476, in save
f(self, obj) # Call unbound method with explicit self
File "/root/miniconda3/envs/py3.6/lib/python3.6/pickle.py", line 751, in save_tuple
save(element)
File "/root/miniconda3/envs/py3.6/lib/python3.6/pickle.py", line 476, in save
f(self, obj) # Call unbound method with explicit self
File "/root/miniconda3/envs/py3.6/lib/python3.6/site-packages/dill/_dill.py", line 933, in save_module_dict
StockPickler.save_dict(pickler, obj)
File "/root/miniconda3/envs/py3.6/lib/python3.6/pickle.py", line 821, in save_dict
self._batch_setitems(obj.items())
File "/root/miniconda3/envs/py3.6/lib/python3.6/pickle.py", line 847, in _batch_setitems
save(v)
File "/root/miniconda3/envs/py3.6/lib/python3.6/pickle.py", line 476, in save
f(self, obj) # Call unbound method with explicit self
File "/root/miniconda3/envs/py3.6/lib/python3.6/site-packages/dill/_dill.py", line 933, in save_module_dict
StockPickler.save_dict(pickler, obj)
File "/root/miniconda3/envs/py3.6/lib/python3.6/pickle.py", line 821, in save_dict
self._batch_setitems(obj.items())
File "/root/miniconda3/envs/py3.6/lib/python3.6/pickle.py", line 847, in _batch_setitems
save(v)
File "/root/miniconda3/envs/py3.6/lib/python3.6/pickle.py", line 507, in save
self.save_global(obj, rv)
File "/root/miniconda3/envs/py3.6/lib/python3.6/pickle.py", line 927, in save_global
(obj, module_name, name))
_pickle.PicklingError: Can't pickle typing.Union[str, NoneType]: it's not the same object as typing.Union
``` | https://github.com/huggingface/datasets/issues/600 | [
"When I change from python3.6 to python3.8, it works! ",
"Does it work when you install `nlp` from source on python 3.6?",
"No, still the pickling error.",
"I wasn't able to reproduce on google colab (python 3.6.9 as well) with \r\n\r\npickle==4.0\r\ndill=0.3.2\r\ntransformers==3.1.0\r\ndatasets=1.0.1 (also t... | null | 600 | false |
Add MATINF dataset | @lhoestq The command to create metadata failed. I guess it's because the zip is not downloaded from a remote address? How to solve that? Also the CI fails and I don't know how to fix that :( | https://github.com/huggingface/datasets/pull/599 | [
"Hi ! sorry for the late response\r\n\r\nCould you try to rebase from master ? We changed the named of the library last week so you have to include this change in your code.\r\n\r\nCan you give me more details about the error you get when running the cli command ?\r\n\r\nNote that in case of a manual download you h... | {
"url": "https://api.github.com/repos/huggingface/datasets/pulls/599",
"html_url": "https://github.com/huggingface/datasets/pull/599",
"diff_url": "https://github.com/huggingface/datasets/pull/599.diff",
"patch_url": "https://github.com/huggingface/datasets/pull/599.patch",
"merged_at": null
} | 599 | true |
The current version of the package on github has an error when loading dataset | Instead of downloading the package from pip, downloading the version from source will result in an error when loading dataset (the pip version is completely fine):
To recreate the error:
First, installing nlp directly from source:
```
git clone https://github.com/huggingface/nlp.git
cd nlp
pip install -e .
```
Then run:
```
from nlp import load_dataset
dataset = load_dataset('wikitext', 'wikitext-2-v1',split = 'train')
```
will give error:
```
>>> dataset = load_dataset('wikitext', 'wikitext-2-v1',split = 'train')
Checking /home/zeyuy/.cache/huggingface/datasets/84a754b488511b109e2904672d809c041008416ae74e38f9ee0c80a8dffa1383.2e21f48d63b5572d19c97e441fbb802257cf6a4c03fbc5ed8fae3d2c2273f59e.py for additional imports.
Found main folder for dataset https://raw.githubusercontent.com/huggingface/nlp/0.4.0/datasets/wikitext/wikitext.py at /home/zeyuy/.cache/huggingface/modules/nlp_modules/datasets/wikitext
Found specific version folder for dataset https://raw.githubusercontent.com/huggingface/nlp/0.4.0/datasets/wikitext/wikitext.py at /home/zeyuy/.cache/huggingface/modules/nlp_modules/datasets/wikitext/5de6e79516446f747fcccc09aa2614fa159053b75909594d28d262395f72d89d
Found script file from https://raw.githubusercontent.com/huggingface/nlp/0.4.0/datasets/wikitext/wikitext.py to /home/zeyuy/.cache/huggingface/modules/nlp_modules/datasets/wikitext/5de6e79516446f747fcccc09aa2614fa159053b75909594d28d262395f72d89d/wikitext.py
Found dataset infos file from https://raw.githubusercontent.com/huggingface/nlp/0.4.0/datasets/wikitext/dataset_infos.json to /home/zeyuy/.cache/huggingface/modules/nlp_modules/datasets/wikitext/5de6e79516446f747fcccc09aa2614fa159053b75909594d28d262395f72d89d/dataset_infos.json
Found metadata file for dataset https://raw.githubusercontent.com/huggingface/nlp/0.4.0/datasets/wikitext/wikitext.py at /home/zeyuy/.cache/huggingface/modules/nlp_modules/datasets/wikitext/5de6e79516446f747fcccc09aa2614fa159053b75909594d28d262395f72d89d/wikitext.json
Loading Dataset Infos from /home/zeyuy/.cache/huggingface/modules/nlp_modules/datasets/wikitext/5de6e79516446f747fcccc09aa2614fa159053b75909594d28d262395f72d89d
Overwrite dataset info from restored data version.
Loading Dataset info from /home/zeyuy/.cache/huggingface/datasets/wikitext/wikitext-2-v1/1.0.0/5de6e79516446f747fcccc09aa2614fa159053b75909594d28d262395f72d89d
Reusing dataset wikitext (/home/zeyuy/.cache/huggingface/datasets/wikitext/wikitext-2-v1/1.0.0/5de6e79516446f747fcccc09aa2614fa159053b75909594d28d262395f72d89d)
Constructing Dataset for split train, from /home/zeyuy/.cache/huggingface/datasets/wikitext/wikitext-2-v1/1.0.0/5de6e79516446f747fcccc09aa2614fa159053b75909594d28d262395f72d89d
Traceback (most recent call last):
File "<stdin>", line 1, in <module>
File "/home/zeyuy/transformers/examples/language-modeling/nlp/src/nlp/load.py", line 600, in load_dataset
ds = builder_instance.as_dataset(split=split, ignore_verifications=ignore_verifications)
File "/home/zeyuy/transformers/examples/language-modeling/nlp/src/nlp/builder.py", line 611, in as_dataset
datasets = utils.map_nested(
File "/home/zeyuy/transformers/examples/language-modeling/nlp/src/nlp/utils/py_utils.py", line 216, in map_nested
return function(data_struct)
File "/home/zeyuy/transformers/examples/language-modeling/nlp/src/nlp/builder.py", line 631, in _build_single_dataset
ds = self._as_dataset(
File "/home/zeyuy/transformers/examples/language-modeling/nlp/src/nlp/builder.py", line 704, in _as_dataset
return Dataset(**dataset_kwargs)
File "/home/zeyuy/transformers/examples/language-modeling/nlp/src/nlp/arrow_dataset.py", line 188, in __init__
self._fingerprint = generate_fingerprint(self)
File "/home/zeyuy/transformers/examples/language-modeling/nlp/src/nlp/fingerprint.py", line 91, in generate_fingerprint
hasher.update(key)
File "/home/zeyuy/transformers/examples/language-modeling/nlp/src/nlp/fingerprint.py", line 57, in update
self.m.update(self.hash(value).encode("utf-8"))
File "/home/zeyuy/transformers/examples/language-modeling/nlp/src/nlp/fingerprint.py", line 53, in hash
return cls.hash_default(value)
File "/home/zeyuy/transformers/examples/language-modeling/nlp/src/nlp/fingerprint.py", line 46, in hash_default
return cls.hash_bytes(dumps(value))
File "/home/zeyuy/transformers/examples/language-modeling/nlp/src/nlp/utils/py_utils.py", line 361, in dumps
with _no_cache_fields(obj):
File "/home/zeyuy/miniconda3/lib/python3.8/contextlib.py", line 113, in __enter__
return next(self.gen)
File "/home/zeyuy/transformers/examples/language-modeling/nlp/src/nlp/utils/py_utils.py", line 348, in _no_cache_fields
if isinstance(obj, tr.PreTrainedTokenizerBase) and hasattr(obj, "cache") and isinstance(obj.cache, dict):
AttributeError: module 'transformers' has no attribute 'PreTrainedTokenizerBase'
```
| https://github.com/huggingface/datasets/issues/598 | [
"Thanks for reporting !\r\nWhich version of transformers are you using ?\r\nIt looks like it doesn't have the PreTrainedTokenizerBase class",
"I was using transformer 2.9. And I switch to the latest transformer package. Everything works just fine!!\r\n\r\nThanks for helping! I should look more carefully next time... | null | 598 | false |
Indices incorrect with multiprocessing | When `num_proc` > 1, the indices argument passed to the map function is incorrect:
```python
d = load_dataset('imdb', split='test[:1%]')
def fn(x, inds):
print(inds)
return x
d.select(range(10)).map(fn, with_indices=True, batched=True)
# [0, 1]
# [0, 1, 2, 3, 4, 5, 6, 7, 8, 9]
d.select(range(10)).map(fn, with_indices=True, batched=True, num_proc=2)
# [0, 1]
# [0, 1]
# [0, 1, 2, 3, 4]
# [0, 1, 2, 3, 4]
```
As you can see, the subset passed to each thread is indexed from 0 to N which doesn't reflect their positions in `d`. | https://github.com/huggingface/datasets/issues/597 | [
"I fixed a bug that could cause this issue earlier today. Could you pull the latest version and try again ?",
"Still the case on master.\r\nI guess we should have an offset in the multi-procs indeed (hopefully it's enough).\r\n\r\nAlso, side note is that we should add some logging before the \"test\" to say we ar... | null | 597 | false |
[style/quality] Moving to isort 5.0.0 + style/quality on datasets and metrics | Move the repo to isort 5.0.0.
Also start testing style/quality on datasets and metrics.
Specific rule: we allow F401 (unused imports) in metrics to be able to add imports to detect early on missing dependencies.
Maybe we could add this in datasets but while cleaning this I've seen many example of really unused imports in dataset so maybe it's better to have it as a line-by-line nova instead of a general rule like in metrics. | https://github.com/huggingface/datasets/pull/596 | [
"Ready for review @lhoestq, just updated a few 156 files here"
] | {
"url": "https://api.github.com/repos/huggingface/datasets/pulls/596",
"html_url": "https://github.com/huggingface/datasets/pull/596",
"diff_url": "https://github.com/huggingface/datasets/pull/596.diff",
"patch_url": "https://github.com/huggingface/datasets/pull/596.patch",
"merged_at": "2020-09-10T10:05:03"
} | 596 | true |
`Dataset`/`DatasetDict` has no attribute 'save_to_disk' | Hi,
As the title indicates, both `Dataset` and `DatasetDict` classes don't seem to have the `save_to_disk` method. While the file [`arrow_dataset.py`](https://github.com/huggingface/nlp/blob/34bf0b03bfe03e7f77b8fec1cd48f5452c4fc7c1/src/nlp/arrow_dataset.py) in the repo here has the method, the file `arrow_dataset.py` which is saved after `pip install nlp -U` in my `conda` environment DOES NOT contain the `save_to_disk` method. I even tried `pip install git+https://github.com/huggingface/nlp.git ` and still no luck. Do I need to install the library in another way? | https://github.com/huggingface/datasets/issues/595 | [
"`pip install git+https://github.com/huggingface/nlp.git` should have done the job.\r\n\r\nDid you uninstall `nlp` before installing from github ?",
"> Did you uninstall `nlp` before installing from github ?\r\n\r\nI did not. I created a new environment and installed `nlp` directly from `github` and it worked!\r\... | null | 595 | false |
Fix germeval url | Continuation of #593 but without the dummy data hack | https://github.com/huggingface/datasets/pull/594 | [] | {
"url": "https://api.github.com/repos/huggingface/datasets/pulls/594",
"html_url": "https://github.com/huggingface/datasets/pull/594",
"diff_url": "https://github.com/huggingface/datasets/pull/594.diff",
"patch_url": "https://github.com/huggingface/datasets/pull/594.patch",
"merged_at": "2020-09-09T13:34:34"
} | 594 | true |
GermEval 2014: new download urls | Hi,
unfortunately, the download links for the GermEval 2014 dataset have changed: they're now located on a Google Drive.
I changed the URLs and bump version from 1.0.0 to 2.0.0. | https://github.com/huggingface/datasets/pull/593 | [
"/cc: @vblagoje",
"Closing this one as #594 is merged (same changes except the dummy data hack)",
"Awesome @stefan-it ! @lhoestq how soon can I use the fixed GermEval dataset in HF token classification examples?",
"I've manually updated the script on S3, so you can actually use it right now with\r\n```python\... | {
"url": "https://api.github.com/repos/huggingface/datasets/pulls/593",
"html_url": "https://github.com/huggingface/datasets/pull/593",
"diff_url": "https://github.com/huggingface/datasets/pull/593.diff",
"patch_url": "https://github.com/huggingface/datasets/pull/593.patch",
"merged_at": null
} | 593 | true |
Test in memory and on disk | I added test parameters to do every test both in memory and on disk.
I also found a bug in concatenate_dataset thanks to the new tests and fixed it. | https://github.com/huggingface/datasets/pull/592 | [] | {
"url": "https://api.github.com/repos/huggingface/datasets/pulls/592",
"html_url": "https://github.com/huggingface/datasets/pull/592",
"diff_url": "https://github.com/huggingface/datasets/pull/592.diff",
"patch_url": "https://github.com/huggingface/datasets/pull/592.patch",
"merged_at": "2020-09-09T13:50:03"
} | 592 | true |
fix #589 (backward compat) | Fix #589 | https://github.com/huggingface/datasets/pull/591 | [] | {
"url": "https://api.github.com/repos/huggingface/datasets/pulls/591",
"html_url": "https://github.com/huggingface/datasets/pull/591",
"diff_url": "https://github.com/huggingface/datasets/pull/591.diff",
"patch_url": "https://github.com/huggingface/datasets/pull/591.patch",
"merged_at": "2020-09-09T08:57:54"
} | 591 | true |
The process cannot access the file because it is being used by another process (windows) | Hi, I consistently get the following error when developing in my PC (windows 10):
```
train_dataset = train_dataset.map(convert_to_features, batched=True)
File "C:\Users\saareliad\AppData\Local\Continuum\miniconda3\envs\py38\lib\site-packages\nlp\arrow_dataset.py", line 970, in map
shutil.move(tmp_file.name, cache_file_name)
File "C:\Users\saareliad\AppData\Local\Continuum\miniconda3\envs\py38\lib\shutil.py", line 803, in move
os.unlink(src)
PermissionError: [WinError 32] The process cannot access the file because it is being used by another process: 'C:\\Users\\saareliad\\.cache\\huggingface\\datasets\\squad\\plain_text\\1.0.0\\408a8fa46a1e2805445b793f1022e743428ca739a34809fce872f0c7f17b44ab\\tmpsau1bep1'
``` | https://github.com/huggingface/datasets/issues/590 | [
"Hi, which version of `nlp` are you using?\r\n\r\nBy the way we'll be releasing today a significant update fixing many issues (but also comprising a few breaking changes).\r\nYou can see more informations here #545 and try it by installing from source from the master branch.",
"I'm using version 0.4.0.\r\n\r\n",
... | null | 590 | false |
Cannot use nlp.load_dataset text, AttributeError: module 'nlp.utils' has no attribute 'logging' |
```
Traceback (most recent call last):
File "<stdin>", line 1, in <module>
File "/root/anaconda3/envs/pytorch/lib/python3.7/site-packages/nlp/load.py", line 533, in load_dataset
builder_cls = import_main_class(module_path, dataset=True)
File "/root/anaconda3/envs/pytorch/lib/python3.7/site-packages/nlp/load.py", line 61, in import_main_class
module = importlib.import_module(module_path)
File "/root/anaconda3/envs/pytorch/lib/python3.7/importlib/__init__.py", line 127, in import_module
return _bootstrap._gcd_import(name[level:], package, level)
File "<frozen importlib._bootstrap>", line 1006, in _gcd_import
File "<frozen importlib._bootstrap>", line 983, in _find_and_load
File "<frozen importlib._bootstrap>", line 967, in _find_and_load_unlocked
File "<frozen importlib._bootstrap>", line 677, in _load_unlocked
File "<frozen importlib._bootstrap_external>", line 728, in exec_module
File "<frozen importlib._bootstrap>", line 219, in _call_with_frames_removed
File "/root/anaconda3/envs/pytorch/lib/python3.7/site-packages/nlp/datasets/text/5dc629379536c4037d9c2063e1caa829a1676cf795f8e030cd90a537eba20c08/text.py", line 9, in <module>
logger = nlp.utils.logging.get_logger(__name__)
AttributeError: module 'nlp.utils' has no attribute 'logging'
```
Occurs on the following code, or any code including the load_dataset('text'):
```
dataset = load_dataset("text", data_files=file_path, split="train")
dataset = dataset.map(lambda ex: tokenizer(ex["text"], add_special_tokens=True,
truncation=True, max_length=args.block_size), batched=True)
dataset.set_format(type='torch', columns=['input_ids'])
return dataset
``` | https://github.com/huggingface/datasets/issues/589 | [] | null | 589 | false |
Support pathlike obj in load dataset | Fix #582
(I recreated the PR, I got an issue with git) | https://github.com/huggingface/datasets/pull/588 | [] | {
"url": "https://api.github.com/repos/huggingface/datasets/pulls/588",
"html_url": "https://github.com/huggingface/datasets/pull/588",
"diff_url": "https://github.com/huggingface/datasets/pull/588.diff",
"patch_url": "https://github.com/huggingface/datasets/pull/588.patch",
"merged_at": "2020-09-08T07:45:17"
} | 588 | true |
Support pathlike obj in load dataset | Fix #582 | https://github.com/huggingface/datasets/pull/587 | [] | {
"url": "https://api.github.com/repos/huggingface/datasets/pulls/587",
"html_url": "https://github.com/huggingface/datasets/pull/587",
"diff_url": "https://github.com/huggingface/datasets/pull/587.diff",
"patch_url": "https://github.com/huggingface/datasets/pull/587.patch",
"merged_at": null
} | 587 | true |
Better message when data files is empty | Fix #581 | https://github.com/huggingface/datasets/pull/586 | [] | {
"url": "https://api.github.com/repos/huggingface/datasets/pulls/586",
"html_url": "https://github.com/huggingface/datasets/pull/586",
"diff_url": "https://github.com/huggingface/datasets/pull/586.diff",
"patch_url": "https://github.com/huggingface/datasets/pull/586.patch",
"merged_at": "2020-09-09T09:00:07"
} | 586 | true |
Fix select for pyarrow < 1.0.0 | Fix #583 | https://github.com/huggingface/datasets/pull/585 | [] | {
"url": "https://api.github.com/repos/huggingface/datasets/pulls/585",
"html_url": "https://github.com/huggingface/datasets/pull/585",
"diff_url": "https://github.com/huggingface/datasets/pull/585.diff",
"patch_url": "https://github.com/huggingface/datasets/pull/585.patch",
"merged_at": "2020-09-08T07:43:15"
} | 585 | true |
Use github versioning | Right now dataset scripts and metrics are downloaded from S3 which is in sync with master. It means that it's not currently possible to pin the dataset/metric script version.
To fix that I changed the download url from S3 to github, and adding a `version` parameter in `load_dataset` and `load_metric` to pin a certain version of the lib, as in #562 | https://github.com/huggingface/datasets/pull/584 | [
"I noticed that datasets like `cnn_dailymail` need the `version` parameter to be passed to its `config_kwargs`.\r\nShall we rename the `version` paramater in `load_dataset` ? Maybe `repo_version` or `script_version` ?"
] | {
"url": "https://api.github.com/repos/huggingface/datasets/pulls/584",
"html_url": "https://github.com/huggingface/datasets/pull/584",
"diff_url": "https://github.com/huggingface/datasets/pull/584.diff",
"patch_url": "https://github.com/huggingface/datasets/pull/584.patch",
"merged_at": "2020-09-09T13:37:34"
} | 584 | true |
ArrowIndexError on Dataset.select | If the indices table consists in several chunks, then `dataset.select` results in an `ArrowIndexError` error for pyarrow < 1.0.0
Example:
```python
from nlp import load_dataset
mnli = load_dataset("glue", "mnli", split="train")
shuffled = mnli.shuffle(seed=42)
mnli.select(list(range(len(mnli))))
```
raises:
```python
---------------------------------------------------------------------------
ArrowIndexError Traceback (most recent call last)
<ipython-input-64-006a5d38d418> in <module>
----> 1 mnli.shuffle(seed=42).select(list(range(len(mnli))))
~/Desktop/hf/nlp/src/nlp/fingerprint.py in wrapper(*args, **kwargs)
161 # Call actual function
162
--> 163 out = func(self, *args, **kwargs)
164
165 # Update fingerprint of in-place transforms + update in-place history of transforms
~/Desktop/hf/nlp/src/nlp/arrow_dataset.py in select(self, indices, keep_in_memory, indices_cache_file_name, writer_batch_size, new_fingerprint)
1653 if self._indices is not None:
1654 if PYARROW_V0:
-> 1655 indices_array = self._indices.column(0).chunk(0).take(indices_array)
1656 else:
1657 indices_array = self._indices.column(0).take(indices_array)
~/.virtualenvs/hf-datasets/lib/python3.7/site-packages/pyarrow/array.pxi in pyarrow.lib.Array.take()
~/.virtualenvs/hf-datasets/lib/python3.7/site-packages/pyarrow/error.pxi in pyarrow.lib.check_status()
ArrowIndexError: take index out of bounds
```
This is because the `take` method is only done on the first chunk which only contains 1000 elements by default (mnli has ~400 000 elements).
Shall we change that to use
```python
pa.concat_tables(self._indices._indices.slice(i, 1) for i in indices_array)
```
instead of `take` ? @thomwolf | https://github.com/huggingface/datasets/issues/583 | [] | null | 583 | false |
Allow for PathLike objects | Using PathLike objects as input for `load_dataset` does not seem to work. The following will throw an error.
```python
files = list(Path(r"D:\corpora\yourcorpus").glob("*.txt"))
dataset = load_dataset("text", data_files=files)
```
Traceback:
```
Traceback (most recent call last):
File "C:/dev/python/dutch-simplification/main.py", line 7, in <module>
dataset = load_dataset("text", data_files=files)
File "C:\Users\bramv\.virtualenvs\dutch-simplification-nbNdqK9u\lib\site-packages\nlp\load.py", line 548, in load_dataset
builder_instance.download_and_prepare(
File "C:\Users\bramv\.virtualenvs\dutch-simplification-nbNdqK9u\lib\site-packages\nlp\builder.py", line 470, in download_and_prepare
self._save_info()
File "C:\Users\bramv\.virtualenvs\dutch-simplification-nbNdqK9u\lib\site-packages\nlp\builder.py", line 564, in _save_info
self.info.write_to_directory(self._cache_dir)
File "C:\Users\bramv\.virtualenvs\dutch-simplification-nbNdqK9u\lib\site-packages\nlp\info.py", line 149, in write_to_directory
self._dump_info(f)
File "C:\Users\bramv\.virtualenvs\dutch-simplification-nbNdqK9u\lib\site-packages\nlp\info.py", line 156, in _dump_info
file.write(json.dumps(asdict(self)).encode("utf-8"))
File "c:\users\bramv\appdata\local\programs\python\python38\lib\json\__init__.py", line 231, in dumps
return _default_encoder.encode(obj)
File "c:\users\bramv\appdata\local\programs\python\python38\lib\json\encoder.py", line 199, in encode
chunks = self.iterencode(o, _one_shot=True)
File "c:\users\bramv\appdata\local\programs\python\python38\lib\json\encoder.py", line 257, in iterencode
return _iterencode(o, 0)
TypeError: keys must be str, int, float, bool or None, not WindowsPath
```
We have to cast to a string explicitly to make this work. It would be nicer if we could actually use PathLike objects.
```python
files = [str(f) for f in Path(r"D:\corpora\wablieft").glob("*.txt")]
```
| https://github.com/huggingface/datasets/issues/582 | [] | null | 582 | false |
Better error message when input file does not exist | In the following scenario, when `data_files` is an empty list, the stack trace and error message could be improved. This can probably be solved by checking for each file whether it actually exists and/or whether the argument is not false-y.
```python
dataset = load_dataset("text", data_files=[])
```
Example error trace.
```
Using custom data configuration default
Downloading and preparing dataset text/default-d18f9b6611eb8e16 (download: Unknown size, generated: Unknown size, post-processed: Unknown sizetotal: Unknown size) to C:\Users\bramv\.cache\huggingface\datasets\text\default-d18f9b6611eb8e16\0.0.0\3a79870d85f1982d6a2af884fde86a71c771747b4b161fd302d28ad22adf985b...
Traceback (most recent call last):
File "C:\Users\bramv\.virtualenvs\dutch-simplification-nbNdqK9u\lib\site-packages\nlp\builder.py", line 424, in incomplete_dir
yield tmp_dir
File "C:\Users\bramv\.virtualenvs\dutch-simplification-nbNdqK9u\lib\site-packages\nlp\builder.py", line 462, in download_and_prepare
self._download_and_prepare(
File "C:\Users\bramv\.virtualenvs\dutch-simplification-nbNdqK9u\lib\site-packages\nlp\builder.py", line 537, in _download_and_prepare
self._prepare_split(split_generator, **prepare_split_kwargs)
File "C:\Users\bramv\.virtualenvs\dutch-simplification-nbNdqK9u\lib\site-packages\nlp\builder.py", line 813, in _prepare_split
num_examples, num_bytes = writer.finalize()
File "C:\Users\bramv\.virtualenvs\dutch-simplification-nbNdqK9u\lib\site-packages\nlp\arrow_writer.py", line 217, in finalize
self.pa_writer.close()
AttributeError: 'NoneType' object has no attribute 'close'
During handling of the above exception, another exception occurred:
Traceback (most recent call last):
File "C:/dev/python/dutch-simplification/main.py", line 7, in <module>
dataset = load_dataset("text", data_files=files)
File "C:\Users\bramv\.virtualenvs\dutch-simplification-nbNdqK9u\lib\site-packages\nlp\load.py", line 548, in load_dataset
builder_instance.download_and_prepare(
File "C:\Users\bramv\.virtualenvs\dutch-simplification-nbNdqK9u\lib\site-packages\nlp\builder.py", line 470, in download_and_prepare
self._save_info()
File "c:\users\bramv\appdata\local\programs\python\python38\lib\contextlib.py", line 131, in __exit__
self.gen.throw(type, value, traceback)
File "C:\Users\bramv\.virtualenvs\dutch-simplification-nbNdqK9u\lib\site-packages\nlp\builder.py", line 430, in incomplete_dir
shutil.rmtree(tmp_dir)
File "c:\users\bramv\appdata\local\programs\python\python38\lib\shutil.py", line 737, in rmtree
return _rmtree_unsafe(path, onerror)
File "c:\users\bramv\appdata\local\programs\python\python38\lib\shutil.py", line 615, in _rmtree_unsafe
onerror(os.unlink, fullname, sys.exc_info())
File "c:\users\bramv\appdata\local\programs\python\python38\lib\shutil.py", line 613, in _rmtree_unsafe
os.unlink(fullname)
PermissionError: [WinError 32] The process cannot access the file because it is being used by another process: 'C:\\Users\\bramv\\.cache\\huggingface\\datasets\\text\\default-d18f9b6611eb8e16\\0.0.0\\3a79870d85f1982d6a2af884fde86a71c771747b4b161fd302d28ad22adf985b.incomplete\\text-train.arrow'
``` | https://github.com/huggingface/datasets/issues/581 | [] | null | 581 | false |
nlp re-creates already-there caches when using a script, but not within a shell | `nlp` keeps creating new caches for the same file when launching `filter` from a script, and behaves correctly from within the shell.
Example: try running
```
import nlp
hans_easy_data = nlp.load_dataset('hans', split="validation").filter(lambda x: x['label'] == 0)
hans_hard_data = nlp.load_dataset('hans', split="validation").filter(lambda x: x['label'] == 1)
```
twice. If launched from a `file.py` script, the cache will be re-created the second time. If launched as 3 shell/`ipython` commands, `nlp` will correctly re-use the cache.
As observed with @lhoestq. | https://github.com/huggingface/datasets/issues/580 | [
"Couln't reproduce on my side :/ \r\nlet me know if you manage to reproduce on another env (colab for example)",
"Fixed with a clean re-install!"
] | null | 580 | false |
Doc metrics | Adding documentation on metrics loading/using/sharing | https://github.com/huggingface/datasets/pull/579 | [] | {
"url": "https://api.github.com/repos/huggingface/datasets/pulls/579",
"html_url": "https://github.com/huggingface/datasets/pull/579",
"diff_url": "https://github.com/huggingface/datasets/pull/579.diff",
"patch_url": "https://github.com/huggingface/datasets/pull/579.patch",
"merged_at": "2020-09-10T13:06:10"
} | 579 | true |
Add CommonGen Dataset | CC Authors:
@yuchenlin @MichaelZhouwang | https://github.com/huggingface/datasets/pull/578 | [] | {
"url": "https://api.github.com/repos/huggingface/datasets/pulls/578",
"html_url": "https://github.com/huggingface/datasets/pull/578",
"diff_url": "https://github.com/huggingface/datasets/pull/578.diff",
"patch_url": "https://github.com/huggingface/datasets/pull/578.patch",
"merged_at": "2020-09-07T11:49:07"
} | 578 | true |
Some languages in wikipedia dataset are not loading | Hi,
I am working with the `wikipedia` dataset and I have a script that goes over 92 of the available languages in that dataset. So far I have detected that `ar`, `af`, `an` are not loading. Other languages like `fr` and `en` are working fine. Here's how I am loading them:
```
import nlp
langs = ['ar'. 'af', 'an']
for lang in langs:
data = nlp.load_dataset('wikipedia', f'20200501.{lang}', beam_runner='DirectRunner', split='train')
print(lang, len(data))
```
Here's what I see for 'ar' (it gets stuck there):
```
Downloading and preparing dataset wikipedia/20200501.ar (download: Unknown size, generated: Unknown size, post-processed: Unknown sizetotal: Unknown size) to /home/gaguilar/.cache/huggingface/datasets/wikipedia/20200501.ar/1.0.0/7be7f4324255faf70687be8692de57cf79197afdc33ff08d6a04ed602df32d50...
```
Note that those languages are indeed in the list of expected languages. Any suggestions on how to work around this? Thanks! | https://github.com/huggingface/datasets/issues/577 | [
"Some wikipedia languages have already been processed by us and are hosted on our google storage. This is the case for \"fr\" and \"en\" for example.\r\n\r\nFor other smaller languages (in terms of bytes), they are directly downloaded and parsed from the wikipedia dump site.\r\nParsing can take some time for langua... | null | 577 | false |
Fix the code block in doc | https://github.com/huggingface/datasets/pull/576 | [
"thanks :)"
] | {
"url": "https://api.github.com/repos/huggingface/datasets/pulls/576",
"html_url": "https://github.com/huggingface/datasets/pull/576",
"diff_url": "https://github.com/huggingface/datasets/pull/576.diff",
"patch_url": "https://github.com/huggingface/datasets/pull/576.patch",
"merged_at": "2020-09-07T07:37:18"
} | 576 | true | |
Couldn't reach certain URLs and for the ones that can be reached, code just blocks after downloading. | Hi,
I'm following the [quick tour](https://huggingface.co/nlp/quicktour.html) and tried to load the glue dataset:
```
>>> from nlp import load_dataset
>>> dataset = load_dataset('glue', 'mrpc', split='train')
```
However, this ran into a `ConnectionError` saying it could not reach the URL (just pasting the last few lines):
```
/net/vaosl01/opt/NFS/su0/miniconda3/envs/hf/lib/python3.7/site-packages/nlp/utils/file_utils.py in get_from_cache(url, cache_dir, force_download, proxies, etag_timeout, resume_download, user_agent, local_files_only)
354 " to False."
355 )
--> 356 raise ConnectionError("Couldn't reach {}".format(url))
357
358 # From now on, connected is True.
ConnectionError: Couldn't reach https://firebasestorage.googleapis.com/v0/b/mtl-sentence-representations.appspot.com/o/data%2Fmrpc_dev_ids.tsv?alt=media&token=ec5c0836-31d5-48f4-b431-7480817f1adc
```
I tried glue with cola and sst2. I got the same error, just instead of mrpc in the URL, it was replaced with cola and sst2.
Since this was not working, I thought I'll try another dataset. So I tried downloading the imdb dataset:
```
ds = load_dataset('imdb', split='train')
```
This downloads the data, but it just blocks after that:
```
Downloading: 100%|████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 4.56k/4.56k [00:00<00:00, 1.38MB/s]
Downloading: 100%|████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 2.07k/2.07k [00:00<00:00, 1.15MB/s]
Downloading and preparing dataset imdb/plain_text (download: 80.23 MiB, generated: 127.06 MiB, post-processed: Unknown sizetotal: 207.28 MiB) to /net/vaosl01/opt/NFS/su0/huggingface/datasets/imdb/plain_text/1.0.0/76cdbd7249ea3548c928bbf304258dab44d09cd3638d9da8d42480d1d1be3743...
Downloading: 100%|████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 84.1M/84.1M [00:07<00:00, 11.1MB/s]
```
I checked the folder `$HF_HOME/datasets/downloads/extracted/<id>/aclImdb`. This folder is constantly growing in size. When I navigated to the train folder within, there was no file. However, the test folder seemed to be populating. The last time I checked it was 327M. I thought the Imdb dataset was smaller than that. My questions are:
1. Why is it still blocking? Is it still downloading?
2. I specified split as train, so why is the test folder being populated?
3. I read somewhere that after downloading, `nlp` converts the text files into some sort of `arrow` files, which will also take a while. Is this also happening here?
Thanks.
| https://github.com/huggingface/datasets/issues/575 | [
"Update:\r\n\r\nThe imdb download completed after a long time (about 45 mins). Ofcourse once download loading was instantaneous. Also, the loaded object was of type `arrow_dataset`. \r\n\r\nThe urls for glue still doesn't work though.",
"Thanks for the report, I'll give a look!",
"I am also seeing a similar err... | null | 575 | false |
Add modules cache | As discusses in #554 , we should use a module cache directory outside of the python packages directory since we may not have write permissions.
I added a new HF_MODULES_PATH directory that is added to the python path when doing `import nlp`.
In this directory, a module `nlp_modules` is created so that datasets can be added to `nlp_modules.datasets` and metrics to `nlp_modules.metrics`. `nlp_modules` doesn't exist on Pypi.
If someone using cloudpickle still wants to have the downloaded dataset/metrics scripts to be inside the nlp directory, it is still possible to change the environment variable HF_MODULES_CACHE to be a path inside the nlp lib. | https://github.com/huggingface/datasets/pull/574 | [
"All the tests pass on my side. Not sure if it is a cache issue or a pytest issue or a circleci issue.\r\nEDIT: I have the same error on google colab. Trying to fix that",
"I think I fixed it (sorry didn't notice you were on it as well)"
] | {
"url": "https://api.github.com/repos/huggingface/datasets/pulls/574",
"html_url": "https://github.com/huggingface/datasets/pull/574",
"diff_url": "https://github.com/huggingface/datasets/pull/574.diff",
"patch_url": "https://github.com/huggingface/datasets/pull/574.patch",
"merged_at": "2020-09-07T09:01:35"
} | 574 | true |
Faster caching for text dataset | As mentioned in #546 and #548 , hashing `data_files` contents to get the cache directory name for a text dataset can take a long time.
To make it faster I changed the hashing so that it takes into account the `path` and the `last modified timestamp` of each data file, instead of iterating through the content of each file to get a hash. | https://github.com/huggingface/datasets/pull/573 | [] | {
"url": "https://api.github.com/repos/huggingface/datasets/pulls/573",
"html_url": "https://github.com/huggingface/datasets/pull/573",
"diff_url": "https://github.com/huggingface/datasets/pull/573.diff",
"patch_url": "https://github.com/huggingface/datasets/pull/573.patch",
"merged_at": "2020-09-04T12:53:23"
} | 573 | true |
Add CLUE Benchmark (11 datasets) | Add 11 tasks of [CLUE](https://github.com/CLUEbenchmark/CLUE). | https://github.com/huggingface/datasets/pull/572 | [
"Thanks, @lhoestq! I've addressed the comments. \r\nAlso, I have tried to use `ClassLabel` [when possible](https://github.com/huggingface/nlp/pull/572/files#diff-1026ac7d7b78bf029cb0ebe63162c77dR297). Is there still somewhere else we can use `ClassLabel`? ",
"I believe CI failure is unrelated.",
"Great job! "
] | {
"url": "https://api.github.com/repos/huggingface/datasets/pulls/572",
"html_url": "https://github.com/huggingface/datasets/pull/572",
"diff_url": "https://github.com/huggingface/datasets/pull/572.diff",
"patch_url": "https://github.com/huggingface/datasets/pull/572.patch",
"merged_at": "2020-09-07T09:59:10"
} | 572 | true |
Serialization | I added `save` and `load` method to serialize/deserialize a dataset object in a folder.
It moves the arrow files there (or write them if the tables were in memory), and saves the pickle state in a json file `state.json`, except the info that are in a separate file `dataset_info.json`.
Example:
```python
import nlp
squad = nlp.load_dataset("squad", split="train")
squad.save("tmp/squad")
squad = nlp.Dataset.load("tmp/squad")
```
`ls tmp/squad`
```
dataset_info.json squad-train.arrow state.json
```
`cat tmp/squad/state.json`
```json
{
"_data": null,
"_data_files": [
{
"filename": "squad-train.arrow",
"skip": 0,
"take": 87599
}
],
"_fingerprint": "61f452797a686bc1",
"_format_columns": null,
"_format_kwargs": {},
"_format_type": null,
"_indexes": {},
"_indices": null,
"_indices_data_files": [],
"_inplace_history": [
{
"transforms": []
}
],
"_output_all_columns": false,
"_split": "train"
}
```
`cat tmp/squad/dataset_info.json`
```json
{
"builder_name": "squad",
"citation": "@article{2016arXiv160605250R,\n author = {{Rajpurkar}, Pranav and {Zhang}, Jian and {Lopyrev},\n Konstantin and {Liang}, Percy},\n title = \"{SQuAD: 100,000+ Questions for Machine Comprehension of Text}\",\n journal = {arXiv e-prints},\n year = 2016,\n eid = {arXiv:1606.05250},\n pages = {arXiv:1606.05250},\narchivePrefix = {arXiv},\n eprint = {1606.05250},\n}\n",
"config_name": "plain_text",
"dataset_size": 89789763,
"description": "Stanford Question Answering Dataset (SQuAD) is a reading comprehension dataset, consisting of questions posed by crowdworkers on a set of Wikipedia articles, where the answer to every question is a segment of text, or span, from the corresponding reading passage, or the question might be unanswerable.\n",
"download_checksums": {
"https://rajpurkar.github.io/SQuAD-explorer/dataset/dev-v1.1.json": {
"checksum": "95aa6a52d5d6a735563366753ca50492a658031da74f301ac5238b03966972c9",
"num_bytes": 4854279
},
"https://rajpurkar.github.io/SQuAD-explorer/dataset/train-v1.1.json": {
"checksum": "3527663986b8295af4f7fcdff1ba1ff3f72d07d61a20f487cb238a6ef92fd955",
"num_bytes": 30288272
}
},
"download_size": 35142551,
"features": {
"answers": {
"_type": "Sequence",
"feature": {
"answer_start": {
"_type": "Value",
"dtype": "int32",
"id": null
},
"text": {
"_type": "Value",
"dtype": "string",
"id": null
}
},
"id": null,
"length": -1
},
"context": {
"_type": "Value",
"dtype": "string",
"id": null
},
"id": {
"_type": "Value",
"dtype": "string",
"id": null
},
"question": {
"_type": "Value",
"dtype": "string",
"id": null
},
"title": {
"_type": "Value",
"dtype": "string",
"id": null
}
},
"homepage": "https://rajpurkar.github.io/SQuAD-explorer/",
"license": "",
"post_processed": {
"features": null,
"resources_checksums": {
"train": {},
"train[:10%]": {}
}
},
"post_processing_size": 0,
"size_in_bytes": 124932314,
"splits": {
"train": {
"dataset_name": "squad",
"name": "train",
"num_bytes": 79317110,
"num_examples": 87599
},
"validation": {
"dataset_name": "squad",
"name": "validation",
"num_bytes": 10472653,
"num_examples": 10570
}
},
"supervised_keys": null,
"version": {
"description": "New split API (https://tensorflow.org/datasets/splits)",
"major": 1,
"minor": 0,
"nlp_version_to_prepare": null,
"patch": 0,
"version_str": "1.0.0"
}
}
``` | https://github.com/huggingface/datasets/pull/571 | [
"I've added save/load for dataset dicts.\r\n\r\nI agree that in the future we should also have a way to save indexes too, and also the in-place history of transforms.\r\n\r\nAlso I understand that it would be cool to have the load function directly at the root of the library, but I'm not sure this should be inside ... | {
"url": "https://api.github.com/repos/huggingface/datasets/pulls/571",
"html_url": "https://github.com/huggingface/datasets/pull/571",
"diff_url": "https://github.com/huggingface/datasets/pull/571.diff",
"patch_url": "https://github.com/huggingface/datasets/pull/571.patch",
"merged_at": "2020-09-07T07:46:07"
} | 571 | true |
add reuters21578 dataset | Reopen a PR this the merge. | https://github.com/huggingface/datasets/pull/570 | [] | {
"url": "https://api.github.com/repos/huggingface/datasets/pulls/570",
"html_url": "https://github.com/huggingface/datasets/pull/570",
"diff_url": "https://github.com/huggingface/datasets/pull/570.diff",
"patch_url": "https://github.com/huggingface/datasets/pull/570.patch",
"merged_at": "2020-09-03T10:46:51"
} | 570 | true |
Revert "add reuters21578 dataset" | Reverts huggingface/nlp#471 | https://github.com/huggingface/datasets/pull/569 | [] | {
"url": "https://api.github.com/repos/huggingface/datasets/pulls/569",
"html_url": "https://github.com/huggingface/datasets/pull/569",
"diff_url": "https://github.com/huggingface/datasets/pull/569.diff",
"patch_url": "https://github.com/huggingface/datasets/pull/569.patch",
"merged_at": "2020-09-03T10:07:12"
} | 569 | true |
`metric.compute` throws `ArrowInvalid` error | I get the following error with `rouge.compute`. It happens only with distributed training, and it occurs randomly I can't easily reproduce it. This is using `nlp==0.4.0`
```
File "/home/beltagy/trainer.py", line 92, in validation_step
rouge_scores = rouge.compute(predictions=generated_str, references=gold_str, rouge_types=['rouge2', 'rouge1', 'rougeL'])
File "/home/beltagy/miniconda3/envs/allennlp/lib/python3.7/site-packages/nlp/metric.py", line 224, in compute
self.finalize(timeout=timeout)
File "/home/beltagy/miniconda3/envs/allennlp/lib/python3.7/site-packages/nlp/metric.py", line 213, in finalize
self.data = Dataset(**reader.read_files(node_files))
File "/home/beltagy/miniconda3/envs/allennlp/lib/python3.7/site-packages/nlp/arrow_reader.py", line 217, in read_files
dataset_kwargs = self._read_files(files=files, info=self._info, original_instructions=original_instructions)
File "/home/beltagy/miniconda3/envs/allennlp/lib/python3.7/site-packages/nlp/arrow_reader.py", line 162, in _read_files
pa_table: pa.Table = self._get_dataset_from_filename(f_dict)
File "/home/beltagy/miniconda3/envs/allennlp/lib/python3.7/site-packages/nlp/arrow_reader.py", line 276, in _get_dataset_from_filename
f = pa.ipc.open_stream(mmap)
File "/home/beltagy/miniconda3/envs/allennlp/lib/python3.7/site-packages/pyarrow/ipc.py", line 173, in open_stream
return RecordBatchStreamReader(source)
File "/home/beltagy/miniconda3/envs/allennlp/lib/python3.7/site-packages/pyarrow/ipc.py", line 64, in __init__
self._open(source)
File "pyarrow/ipc.pxi", line 469, in pyarrow.lib._RecordBatchStreamReader._open
File "pyarrow/error.pxi", line 122, in pyarrow.lib.pyarrow_internal_check_status
File "pyarrow/error.pxi", line 84, in pyarrow.lib.check_status
pyarrow.lib.ArrowInvalid: Tried reading schema message, was null or length 0
``` | https://github.com/huggingface/datasets/issues/568 | [
"Hmm might be related to what we are solving in #564",
"Could you try to update to `datasets>=1.0.0` (we changed the name of the library) and try again ?\r\nIf is was related to the distributed setup settings it must be fixed.\r\nIf it was related to empty metric inputs it's going to be fixed in #654 ",
"Closin... | null | 568 | false |
Fix BLEURT metrics for backward compatibility | Fix #565 | https://github.com/huggingface/datasets/pull/567 | [] | {
"url": "https://api.github.com/repos/huggingface/datasets/pulls/567",
"html_url": "https://github.com/huggingface/datasets/pull/567",
"diff_url": "https://github.com/huggingface/datasets/pull/567.diff",
"patch_url": "https://github.com/huggingface/datasets/pull/567.patch",
"merged_at": "2020-09-03T07:29:50"
} | 567 | true |
Remove logger pickling to fix gg colab issues | A `logger` objects are not picklable in google colab, contrary to `logger` objects in jupyter notebooks or in python shells.
It creates some issues in google colab right now.
Indeed by calling any `Dataset` method, the fingerprint update pickles the transform function, and as the logger comes with it, it results in an error (full stacktrace [here](http://pastebin.fr/64330)):
```python
/usr/local/lib/python3.6/dist-packages/zmq/backend/cython/socket.cpython-36m-x86_64-linux-gnu.so in zmq.backend.cython.socket.Socket.__reduce_cython__()
TypeError: no default __reduce__ due to non-trivial __cinit__
```
To fix that I no longer dump the transform (`_map_single`, `select`, etc.), but the full name only (`nlp.arrow_dataset.Dataset._map_single`, `nlp.arrow_dataset.Dataset.select`, etc.) | https://github.com/huggingface/datasets/pull/566 | [] | {
"url": "https://api.github.com/repos/huggingface/datasets/pulls/566",
"html_url": "https://github.com/huggingface/datasets/pull/566",
"diff_url": "https://github.com/huggingface/datasets/pull/566.diff",
"patch_url": "https://github.com/huggingface/datasets/pull/566.patch",
"merged_at": "2020-09-03T16:31:52"
} | 566 | true |
No module named 'nlp.logging' | Hi, I am using nlp version 0.4.0. Trying to use bleurt as an eval metric, however, the bleurt script imports nlp.logging which creates the following error. What am I missing?
```
>>> import nlp
2020-09-02 13:47:09.210310: I tensorflow/stream_executor/platform/default/dso_loader.cc:48] Successfully opened dynamic library libcudart.so.10.1
>>> bleurt = nlp.load_metric("bleurt")
Traceback (most recent call last):
File "<stdin>", line 1, in <module>
File "/home/melody/anaconda3/envs/transformers/lib/python3.6/site-packages/nlp/load.py", line 443, in load_metric
metric_cls = import_main_class(module_path, dataset=False)
File "/home/melody/anaconda3/envs/transformers/lib/python3.6/site-packages/nlp/load.py", line 61, in import_main_class
module = importlib.import_module(module_path)
File "/home/melody/anaconda3/envs/transformers/lib/python3.6/importlib/__init__.py", line 126, in import_module
return _bootstrap._gcd_import(name[level:], package, level)
File "<frozen importlib._bootstrap>", line 994, in _gcd_import
File "<frozen importlib._bootstrap>", line 971, in _find_and_load
File "<frozen importlib._bootstrap>", line 955, in _find_and_load_unlocked
File "<frozen importlib._bootstrap>", line 665, in _load_unlocked
File "<frozen importlib._bootstrap_external>", line 678, in exec_module
File "<frozen importlib._bootstrap>", line 219, in _call_with_frames_removed
File "/home/melody/anaconda3/envs/transformers/lib/python3.6/site-packages/nlp/metrics/bleurt/43448cf2959ea81d3ae0e71c5c8ee31dc15eed9932f197f5f50673cbcecff2b5/bleurt.py", line 20, in <module>
from nlp.logging import get_logger
ModuleNotFoundError: No module named 'nlp.logging'
```
Just to show once again that I can't import the logging module:
```
>>> import nlp
2020-09-02 13:48:38.190621: I tensorflow/stream_executor/platform/default/dso_loader.cc:48] Successfully opened dynamic library libcudart.so.10.1
>>> nlp.__version__
'0.4.0'
>>> from nlp.logging import get_logger
Traceback (most recent call last):
File "<stdin>", line 1, in <module>
ModuleNotFoundError: No module named 'nlp.logging'
``` | https://github.com/huggingface/datasets/issues/565 | [
"Thanks for reporting.\r\n\r\nApparently this is a versioning issue: the lib downloaded the `bleurt` script from the master branch where we did this change recently. We'll fix that in a new release this week or early next week. Cc @thomwolf \r\n\r\nUntil that, I'd suggest you to download the right bleurt folder fro... | null | 565 | false |
Wait for writing in distributed metrics | There were CI bugs where a distributed metric would try to read all the files in process 0 while the other processes haven't started writing.
To fix that I added a custom locking mechanism that waits for the file to exist before trying to read it | https://github.com/huggingface/datasets/pull/564 | [
"I agree this fix the problem for the CI where the files are always created in a new and clean temporary directory.\r\n\r\nHowever, in a general setting of a succession of fast distributed operation, the files could already exist from previous metrics runs but one process may still finish before another has even st... | {
"url": "https://api.github.com/repos/huggingface/datasets/pulls/564",
"html_url": "https://github.com/huggingface/datasets/pull/564",
"diff_url": "https://github.com/huggingface/datasets/pull/564.diff",
"patch_url": "https://github.com/huggingface/datasets/pull/564.patch",
"merged_at": "2020-09-09T09:13:22"
} | 564 | true |
[Large datasets] Speed up download and processing | Various improvements to speed-up creation and processing of large scale datasets.
Currently:
- distributed downloads
- remove etag from datafiles hashes to spare a request when restarting a failed download | https://github.com/huggingface/datasets/pull/563 | [
"Looks all good :)\r\nI rebased from master and added a test for parallel `map_nested`",
"you're da best"
] | {
"url": "https://api.github.com/repos/huggingface/datasets/pulls/563",
"html_url": "https://github.com/huggingface/datasets/pull/563",
"diff_url": "https://github.com/huggingface/datasets/pull/563.diff",
"patch_url": "https://github.com/huggingface/datasets/pull/563.patch",
"merged_at": "2020-09-09T09:03:32"
} | 563 | true |
[Reproductibility] Allow to pin versions of datasets/metrics | Repurpose the `version` attribute in datasets and metrics to let the user pin a specific version of datasets and metric scripts:
```
dataset = nlp.load_dataset('squad', version='1.0.0')
metric = nlp.load_metric('squad', version='1.0.0')
```
Notes:
- version number are the release version of the library
- currently only possible for canonical datasets/metrics, ie. integrated in the GitHub repo of the library | https://github.com/huggingface/datasets/pull/562 | [
"Closing this one in favor of #584 "
] | {
"url": "https://api.github.com/repos/huggingface/datasets/pulls/562",
"html_url": "https://github.com/huggingface/datasets/pull/562",
"diff_url": "https://github.com/huggingface/datasets/pull/562.diff",
"patch_url": "https://github.com/huggingface/datasets/pull/562.patch",
"merged_at": null
} | 562 | true |
Made `share_dataset` more readable | https://github.com/huggingface/datasets/pull/561 | [] | {
"url": "https://api.github.com/repos/huggingface/datasets/pulls/561",
"html_url": "https://github.com/huggingface/datasets/pull/561",
"diff_url": "https://github.com/huggingface/datasets/pull/561.diff",
"patch_url": "https://github.com/huggingface/datasets/pull/561.patch",
"merged_at": "2020-09-03T09:00:29"
} | 561 | true | |
Using custom DownloadConfig results in an error | ## Version / Environment
Ubuntu 18.04
Python 3.6.8
nlp 0.4.0
## Description
Loading `imdb` dataset works fine when when I don't specify any `download_config` argument. When I create a custom `DownloadConfig` object and pass it to the `nlp.load_dataset` function, this results in an error.
## How to reproduce
### Example without DownloadConfig --> works
```python
import os
os.environ["HF_HOME"] = "/data/hf-test-without-dl-config-01/"
import logging
import nlp
logging.basicConfig(level=logging.INFO)
if __name__ == "__main__":
imdb = nlp.load_dataset(path="imdb")
```
### Example with DownloadConfig --> doesn't work
```python
import os
os.environ["HF_HOME"] = "/data/hf-test-with-dl-config-01/"
import logging
import nlp
from nlp.utils import DownloadConfig
logging.basicConfig(level=logging.INFO)
if __name__ == "__main__":
download_config = DownloadConfig()
imdb = nlp.load_dataset(path="imdb", download_config=download_config)
```
Error traceback:
```
Traceback (most recent call last):
File "/.../example_with_dl_config.py", line 13, in <module>
imdb = nlp.load_dataset(path="imdb", download_config=download_config)
File "/.../python3.6/python3.6/site-packages/nlp/load.py", line 549, in load_dataset
download_config=download_config, download_mode=download_mode, ignore_verifications=ignore_verifications,
File "/.../python3.6/python3.6/site-packages/nlp/builder.py", line 463, in download_and_prepare
dl_manager=dl_manager, verify_infos=verify_infos, **download_and_prepare_kwargs
File "/.../python3.6/python3.6/site-packages/nlp/builder.py", line 518, in _download_and_prepare
split_generators = self._split_generators(dl_manager, **split_generators_kwargs)
File "/.../python3.6/python3.6/site-packages/nlp/datasets/imdb/76cdbd7249ea3548c928bbf304258dab44d09cd3638d9da8d42480d1d1be3743/imdb.py", line 86, in _split_generators
arch_path = dl_manager.download_and_extract(_DOWNLOAD_URL)
File "/.../python3.6/python3.6/site-packages/nlp/utils/download_manager.py", line 220, in download_and_extract
return self.extract(self.download(url_or_urls))
File "/.../python3.6/python3.6/site-packages/nlp/utils/download_manager.py", line 158, in download
self._record_sizes_checksums(url_or_urls, downloaded_path_or_paths)
File "/.../python3.6/python3.6/site-packages/nlp/utils/download_manager.py", line 108, in _record_sizes_checksums
self._recorded_sizes_checksums[url] = get_size_checksum_dict(path)
File "/.../python3.6/python3.6/site-packages/nlp/utils/info_utils.py", line 79, in get_size_checksum_dict
with open(path, "rb") as f:
IsADirectoryError: [Errno 21] Is a directory: '/data/hf-test-with-dl-config-01/datasets/extracted/b6802c5b61824b2c1f7dbf7cda6696b5f2e22214e18d171ce1ed3be90c931ce5'
```
| https://github.com/huggingface/datasets/issues/560 | [
"From my limited understanding, part of the issue seems related to the `prepare_module` and `download_and_prepare` functions each handling the case where no config is passed. For example, `prepare_module` does mutate the object passed and forces the flags `extract_compressed_file` and `force_extract` to `True`.\r\... | null | 560 | false |
Adding the KILT knowledge source and tasks | This adds Wikipedia pre-processed for KILT, as well as the task data. Only the question IDs are provided for TriviaQA, but they can easily be mapped back with:
```
import nlp
kilt_wikipedia = nlp.load_dataset('kilt_wikipedia')
kilt_tasks = nlp.load_dataset('kilt_tasks')
triviaqa = nlp.load_dataset('trivia_qa', 'unfiltered.nocontext')
triviaqa_map = {}
for k in ['train', 'validation', 'test']:
triviaqa_map = dict([(q_id, i) for i, q_id in enumerate(triviaqa[k]['question_id'])])
kilt_tasks[k + '_triviaqa'] = kilt_tasks[k + '_triviaqa'].filter(lambda x: x['id'] in triviaqa_map)
kilt_tasks[k + '_triviaqa'].map(lambda x: {'input': triviaqa[split][triviaqa_map[x['id']]]['question']})
```
It would be great to have the dataset by Monday, which is when the paper should land on Arxiv and @fabiopetroni is planning on tweeting about the paper and `facebookresearch` repository for the datasett | https://github.com/huggingface/datasets/pull/559 | [
"Feel free to merge when you are happy with it @yjernite :-)"
] | {
"url": "https://api.github.com/repos/huggingface/datasets/pulls/559",
"html_url": "https://github.com/huggingface/datasets/pull/559",
"diff_url": "https://github.com/huggingface/datasets/pull/559.diff",
"patch_url": "https://github.com/huggingface/datasets/pull/559.patch",
"merged_at": "2020-09-04T18:05:47"
} | 559 | true |
Rerun pip install -e | Hopefully it fixes the github actions | https://github.com/huggingface/datasets/pull/558 | [] | {
"url": "https://api.github.com/repos/huggingface/datasets/pulls/558",
"html_url": "https://github.com/huggingface/datasets/pull/558",
"diff_url": "https://github.com/huggingface/datasets/pull/558.diff",
"patch_url": "https://github.com/huggingface/datasets/pull/558.patch",
"merged_at": "2020-09-01T17:24:50"
} | 558 | true |
Fix a few typos | https://github.com/huggingface/datasets/pull/557 | [] | {
"url": "https://api.github.com/repos/huggingface/datasets/pulls/557",
"html_url": "https://github.com/huggingface/datasets/pull/557",
"diff_url": "https://github.com/huggingface/datasets/pull/557.diff",
"patch_url": "https://github.com/huggingface/datasets/pull/557.patch",
"merged_at": "2020-09-02T07:39:06"
} | 557 | true | |
Add DailyDialog | http://yanran.li/dailydialog.html
https://arxiv.org/pdf/1710.03957.pdf
| https://github.com/huggingface/datasets/pull/556 | [] | {
"url": "https://api.github.com/repos/huggingface/datasets/pulls/556",
"html_url": "https://github.com/huggingface/datasets/pull/556",
"diff_url": "https://github.com/huggingface/datasets/pull/556.diff",
"patch_url": "https://github.com/huggingface/datasets/pull/556.patch",
"merged_at": "2020-09-03T15:38:39"
} | 556 | true |
Upgrade pip in benchmark github action | It looks like it fixes the `import nlp` issue we have | https://github.com/huggingface/datasets/pull/555 | [] | {
"url": "https://api.github.com/repos/huggingface/datasets/pulls/555",
"html_url": "https://github.com/huggingface/datasets/pull/555",
"diff_url": "https://github.com/huggingface/datasets/pull/555.diff",
"patch_url": "https://github.com/huggingface/datasets/pull/555.patch",
"merged_at": "2020-09-01T15:26:15"
} | 555 | true |
nlp downloads to its module path | I am trying to package `nlp` for Nix, because it is now an optional dependency for `transformers`. The problem that I encounter is that the `nlp` library downloads to the module path, which is typically not writable in most package management systems:
```>>> import nlp
>>> squad_dataset = nlp.load_dataset('squad')
Traceback (most recent call last):
File "<stdin>", line 1, in <module>
File "/nix/store/2yhik0hhqayksmkkfb0ylqp8cf5wa5wp-python3-3.8.5-env/lib/python3.8/site-packages/nlp/load.py", line 530, in load_dataset
module_path, hash = prepare_module(path, download_config=download_config, dataset=True)
File "/nix/store/2yhik0hhqayksmkkfb0ylqp8cf5wa5wp-python3-3.8.5-env/lib/python3.8/site-packages/nlp/load.py", line 329, in prepare_module
os.makedirs(main_folder_path, exist_ok=True)
File "/nix/store/685kq8pyhrvajah1hdsfn4q7gm3j4yd4-python3-3.8.5/lib/python3.8/os.py", line 223, in makedirs
mkdir(name, mode)
OSError: [Errno 30] Read-only file system: '/nix/store/2yhik0hhqayksmkkfb0ylqp8cf5wa5wp-python3-3.8.5-env/lib/python3.8/site-packages/nlp/datasets/squad'
```
Do you have any suggested workaround for this issue?
Perhaps overriding the default value for `force_local_path` of `prepare_module`? | https://github.com/huggingface/datasets/issues/554 | [
"Indeed this is a known issue arising from the fact that we try to be compatible with cloupickle.\r\n\r\nDoes this also happen if you are installing in a virtual environment?",
"> Indeed this is a know issue with the fact that we try to be compatible with cloupickle.\r\n> \r\n> Does this also happen if you are in... | null | 554 | false |
[Fix GitHub Actions] test adding tmate | https://github.com/huggingface/datasets/pull/553 | [] | {
"url": "https://api.github.com/repos/huggingface/datasets/pulls/553",
"html_url": "https://github.com/huggingface/datasets/pull/553",
"diff_url": "https://github.com/huggingface/datasets/pull/553.diff",
"patch_url": "https://github.com/huggingface/datasets/pull/553.patch",
"merged_at": null
} | 553 | true | |
Add multiprocessing | Adding multiprocessing to `.map`
It works in 3 steps:
- shard the dataset in `num_proc` shards
- spawn one process per shard and call `map` on them
- concatenate the resulting datasets
Example of usage:
```python
from nlp import load_dataset
dataset = load_dataset("squad", split="train")
def function(x):
return {"lowered": x.lower()}
processed = d.map(
function,
input_columns=["context"],
num_proc=4,
cache_file_name="playground/tmp.arrow",
load_from_cache_file=False
)
```
Here it writes 4 files depending on the process rank:
- `playground/tmp_00000_of_00004.arrow`
- `playground/tmp_00001_of_00004.arrow`
- `playground/tmp_00002_of_00004.arrow`
- `playground/tmp_00003_of_00004.arrow`
The suffix format can be specified by the user.
If the `cache_file_name` is not specified, it writes into separated files depending on the fingerprint, as usual.
I still need to:
- write tests for this
- try to improve the logging (currently it shows 4 progress bars, but if one finishes before the others, then the following messages are written over the progress bars)
| https://github.com/huggingface/datasets/pull/552 | [
"Logging looks like\r\n\r\n```\r\nDone writing 21900 indices in 3854400 bytes .\r\nProcess #0 will write at playground/tmp_00000_of_00004.arrow\r\nDone writing 21900 indices in 3854400 bytes .\r\nProcess #1 will write at playground/tmp_00001_of_00004.arrow\r\nDone writing 21900 indices in 3854400 bytes .\r\nProcess... | {
"url": "https://api.github.com/repos/huggingface/datasets/pulls/552",
"html_url": "https://github.com/huggingface/datasets/pull/552",
"diff_url": "https://github.com/huggingface/datasets/pull/552.diff",
"patch_url": "https://github.com/huggingface/datasets/pull/552.patch",
"merged_at": "2020-09-02T10:01:25"
} | 552 | true |
added HANS dataset | Adds the [HANS](https://github.com/tommccoy1/hans) dataset to evaluate NLI systems. | https://github.com/huggingface/datasets/pull/551 | [] | {
"url": "https://api.github.com/repos/huggingface/datasets/pulls/551",
"html_url": "https://github.com/huggingface/datasets/pull/551",
"diff_url": "https://github.com/huggingface/datasets/pull/551.diff",
"patch_url": "https://github.com/huggingface/datasets/pull/551.patch",
"merged_at": "2020-09-01T12:17:10"
} | 551 | true |
[BUGFIX] Solving mismatched checksum issue for the LinCE dataset (#539) | Hi,
I have added the updated `dataset_infos.json` file for the LinCE benchmark. This update is to fix the mismatched checksum bug #539 for one of the datasets in the LinCE benchmark. To update the file, I run this command from the nlp root directory:
```
python nlp-cli test ./datasets/lince --save_infos --all_configs
```
**NOTE**: I needed to change [this line](https://github.com/huggingface/nlp/blob/master/src/nlp/commands/dummy_data.py#L8) from: `from .utils.logging import get_logger` to `from nlp.utils.logging import get_logger`, otherwise the script was not able to import `get_logger`. However, I did not include that in this PR since that could have been just my environment (and another PR could be fixing this already if it is actually an issue). | https://github.com/huggingface/datasets/pull/550 | [
"Thanks a lot for that!\r\nThe line you are mentioning is a bug indeed, do you mind fixing it at the same time?",
"No worries! \r\n\r\nI pushed right away the fix, but then I realized that the master branch already had it, so I ended up merging the master branch with lince locally and then overwriting the previou... | {
"url": "https://api.github.com/repos/huggingface/datasets/pulls/550",
"html_url": "https://github.com/huggingface/datasets/pull/550",
"diff_url": "https://github.com/huggingface/datasets/pull/550.diff",
"patch_url": "https://github.com/huggingface/datasets/pull/550.patch",
"merged_at": "2020-09-03T09:06:01"
} | 550 | true |
Fix bleurt logging import | Bleurt started throwing an error in some code we have.
This looks like the fix but...
It's also unnerving that even a prebuilt docker image with pinned versions can be working 1 day and then fail the next (especially for production systems).
Any way for us to pin your metrics code so that they are guaranteed not to to change and possibly fail on repository changes?
Thanks (and also for your continued work on the lib...) | https://github.com/huggingface/datasets/pull/549 | [
"That’s a good point that we started to discuss internally as well. We should pin the dataset en metrics code by default indeed.\r\nLet’s update this in the coming release.",
"Ok closed this with #567 and we are working on a more general solution to pin dataset version in #562 (should be in the coming release)."
... | {
"url": "https://api.github.com/repos/huggingface/datasets/pulls/549",
"html_url": "https://github.com/huggingface/datasets/pull/549",
"diff_url": "https://github.com/huggingface/datasets/pull/549.diff",
"patch_url": "https://github.com/huggingface/datasets/pull/549.patch",
"merged_at": null
} | 549 | true |
[Breaking] Switch text loading to multi-threaded PyArrow loading | Test if we can get better performances for large-scale text datasets by using multi-threaded text file loading based on Apache Arrow multi-threaded CSV loader.
If it works ok, it would fix #546.
**Breaking change**:
The text lines now do not include final line-breaks anymore. | https://github.com/huggingface/datasets/pull/548 | [
"Awesome !\r\nAlso I was wondering if we should try to make the hashing of the `data_files` faster (it is used to build the cache directory of datasets like `text` or `json`). Right now it reads each file and hashes all of its data. We could simply hash the path and some metadata including the `time last modified` ... | {
"url": "https://api.github.com/repos/huggingface/datasets/pulls/548",
"html_url": "https://github.com/huggingface/datasets/pull/548",
"diff_url": "https://github.com/huggingface/datasets/pull/548.diff",
"patch_url": "https://github.com/huggingface/datasets/pull/548.patch",
"merged_at": "2020-09-08T10:19:57"
} | 548 | true |
[Distributed] Making loading distributed datasets a bit safer | Add some file-locks during dataset loading | https://github.com/huggingface/datasets/pull/547 | [] | {
"url": "https://api.github.com/repos/huggingface/datasets/pulls/547",
"html_url": "https://github.com/huggingface/datasets/pull/547",
"diff_url": "https://github.com/huggingface/datasets/pull/547.diff",
"patch_url": "https://github.com/huggingface/datasets/pull/547.patch",
"merged_at": "2020-08-31T15:16:29"
} | 547 | true |
Very slow data loading on large dataset | I made a simple python script to check the NLP library speed, which loads 1.1 TB of textual data.
It has been 8 hours and still, it is on the loading steps.
It does work when the text dataset size is small about 1 GB, but it doesn't scale.
It also uses a single thread during the data loading step.
```
train_files = glob.glob("xxx/*.txt",recursive=True)
random.shuffle(train_files)
print(train_files)
dataset = nlp.load_dataset('text',
data_files=train_files,
name="customDataset",
version="1.0.0",
cache_dir="xxx/nlp")
```
Is there something that I am missing ? | https://github.com/huggingface/datasets/issues/546 | [
"When you load a text file for the first time with `nlp`, the file is converted into Apache Arrow format. Arrow allows to use memory-mapping, which means that you can load an arbitrary large dataset.\r\n\r\nNote that as soon as the conversion has been done once, the next time you'll load the dataset it will be much... | null | 546 | false |
New release coming up for this library | Hi all,
A few words on the roadmap for this library.
The next release will be a big one and is planed at the end of this week.
In addition to the support for indexed datasets (useful for non-parametric models like REALM, RAG, DPR, knn-LM and many other fast dataset retrieval technics), it will:
- have support for multi-modal datasets
- include various significant improvements on speed for standard processing (map, shuffling, ...)
- have a better support for metrics (better caching, and a robust API) and a bigger focus on reproductibility
- change the name to the final name (voted by the community): `datasets`
- be the 1.0.0 release as we think the API will be mostly stabilized from now on | https://github.com/huggingface/datasets/issues/545 | [
"Update: release is planed mid-next week."
] | null | 545 | false |
[Distributed] Fix load_dataset error when multiprocessing + add test | Fix #543 + add test | https://github.com/huggingface/datasets/pull/544 | [] | {
"url": "https://api.github.com/repos/huggingface/datasets/pulls/544",
"html_url": "https://github.com/huggingface/datasets/pull/544",
"diff_url": "https://github.com/huggingface/datasets/pull/544.diff",
"patch_url": "https://github.com/huggingface/datasets/pull/544.patch",
"merged_at": "2020-08-31T11:15:10"
} | 544 | true |
nlp.load_dataset is not safe for multi processes when loading from local files | Loading from local files, e.g., `dataset = nlp.load_dataset('csv', data_files=['file_1.csv', 'file_2.csv'])`
concurrently from multiple processes, will raise `FileExistsError` from builder's line 430, https://github.com/huggingface/nlp/blob/6655008c738cb613c522deb3bd18e35a67b2a7e5/src/nlp/builder.py#L423-L438
Likely because multiple processes step into download_and_prepare, https://github.com/huggingface/nlp/blob/6655008c738cb613c522deb3bd18e35a67b2a7e5/src/nlp/load.py#L550-L554
This can happen when launching distributed training with commands like `python -m torch.distributed.launch --nproc_per_node 4` on a new collection of files never loaded before.
I can create a PR that puts in some file locks. It would be helpful if I can be informed of the convention for naming and placement of the lock. | https://github.com/huggingface/datasets/issues/543 | [
"I'll take a look!"
] | null | 543 | false |
Add TensorFlow example | Update the Quick Tour documentation in order to add the TensorFlow equivalent source code for the classification example. Now it is possible to select either the code in PyTorch or in TensorFlow in the Quick tour. | https://github.com/huggingface/datasets/pull/542 | [] | {
"url": "https://api.github.com/repos/huggingface/datasets/pulls/542",
"html_url": "https://github.com/huggingface/datasets/pull/542",
"diff_url": "https://github.com/huggingface/datasets/pull/542.diff",
"patch_url": "https://github.com/huggingface/datasets/pull/542.patch",
"merged_at": "2020-08-31T09:49:19"
} | 542 | true |
Best practices for training tokenizers with nlp | Hi, thank you for developing this library.
What do you think are the best practices for training tokenizers using `nlp`? In the document and examples, I could only find pre-trained tokenizers used. | https://github.com/huggingface/datasets/issues/541 | [
"Docs that explain how to train a tokenizer with `datasets` are available here: https://huggingface.co/docs/tokenizers/training_from_memory#using-the-datasets-library"
] | null | 541 | false |
[BUGFIX] Fix Race Dataset Checksum bug | In #537 I noticed that there was a bug in checksum checking when I have tried to download the race dataset. The reason for this is that the current preprocessing was just considering the `high school` data and it was ignoring the `middle` one. This PR just fixes it :)
Moreover, I have added some descriptions. | https://github.com/huggingface/datasets/pull/540 | [
"I'm not sure this would fix #537 .\r\nHowever your point about the missing `middle` data is right and we probably want to include these data as well.\r\nDo you think it would we worth having different configurations for this dataset for users who want to only load part of it (`high school` or `middle` or `all`) ?"... | {
"url": "https://api.github.com/repos/huggingface/datasets/pulls/540",
"html_url": "https://github.com/huggingface/datasets/pull/540",
"diff_url": "https://github.com/huggingface/datasets/pull/540.diff",
"patch_url": "https://github.com/huggingface/datasets/pull/540.patch",
"merged_at": "2020-09-18T11:42:20"
} | 540 | true |
[Dataset] `NonMatchingChecksumError` due to an update in the LinCE benchmark data | Hi,
There is a `NonMatchingChecksumError` error for the `lid_msaea` (language identification for Modern Standard Arabic - Egyptian Arabic) dataset from the LinCE benchmark due to a minor update on that dataset.
How can I update the checksum of the library to solve this issue? The error is below and it also appears in the [nlp viewer](https://huggingface.co/nlp/viewer/?dataset=lince&config=lid_msaea):
```python
import nlp
nlp.load_dataset('lince', 'lid_msaea')
```
Output:
```
NonMatchingChecksumError: ['https://ritual.uh.edu/lince/libaccess/eyJ1c2VybmFtZSI6ICJodWdnaW5nZmFjZSBubHAiLCAidXNlcl9pZCI6IDExMSwgImVtYWlsIjogImR1bW15QGVtYWlsLmNvbSJ9/lid_msaea.zip']
Traceback:
File "/home/sasha/streamlit/lib/streamlit/ScriptRunner.py", line 322, in _run_script
exec(code, module.__dict__)
File "/home/sasha/nlp-viewer/run.py", line 196, in <module>
dts, fail = get(str(option.id), str(conf_option.name) if conf_option else None)
File "/home/sasha/streamlit/lib/streamlit/caching.py", line 591, in wrapped_func
return get_or_create_cached_value()
File "/home/sasha/streamlit/lib/streamlit/caching.py", line 575, in get_or_create_cached_value
return_value = func(*args, **kwargs)
File "/home/sasha/nlp-viewer/run.py", line 150, in get
builder_instance.download_and_prepare()
File "/home/sasha/.local/share/virtualenvs/lib-ogGKnCK_/lib/python3.7/site-packages/nlp/builder.py", line 432, in download_and_prepare
download_config.force_download = download_mode == FORCE_REDOWNLOAD
File "/home/sasha/.local/share/virtualenvs/lib-ogGKnCK_/lib/python3.7/site-packages/nlp/builder.py", line 469, in _download_and_prepare
File "/home/sasha/.local/share/virtualenvs/lib-ogGKnCK_/lib/python3.7/site-packages/nlp/utils/info_utils.py", line 36, in verify_checksums
raise NonMatchingChecksumError(str(bad_urls))
```
Thank you in advance!
@lhoestq | https://github.com/huggingface/datasets/issues/539 | [
"Hi @gaguilar \r\n\r\nIf you want to take care of this, it very simple, you just need to regenerate the `dataset_infos.json` file as indicated [in the doc](https://huggingface.co/nlp/share_dataset.html#adding-metadata) by [installing from source](https://huggingface.co/nlp/installation.html#installing-from-source) ... | null | 539 | false |
[logging] Add centralized logging - Bump-up cache loads to warnings | Add a `nlp.logging` module to set the global logging level easily. The verbosity level also controls the tqdm bars (disabled when set higher than INFO).
You can use:
```
nlp.logging.set_verbosity(verbosity: int)
nlp.logging.set_verbosity_info()
nlp.logging.set_verbosity_warning()
nlp.logging.set_verbosity_debug()
nlp.logging.set_verbosity_error()
nlp.logging.get_verbosity() -> int
```
And use the levels:
```
nlp.logging.CRITICAL
nlp.logging.DEBUG
nlp.logging.ERROR
nlp.logging.FATAL
nlp.logging.INFO
nlp.logging.NOTSET
nlp.logging.WARN
nlp.logging.WARNING
``` | https://github.com/huggingface/datasets/pull/538 | [] | {
"url": "https://api.github.com/repos/huggingface/datasets/pulls/538",
"html_url": "https://github.com/huggingface/datasets/pull/538",
"diff_url": "https://github.com/huggingface/datasets/pull/538.diff",
"patch_url": "https://github.com/huggingface/datasets/pull/538.patch",
"merged_at": "2020-08-31T11:42:50"
} | 538 | true |
[Dataset] RACE dataset Checksums error | Hi there, I just would like to use this awesome lib to perform a dataset fine-tuning on RACE dataset. I have performed the following steps:
```
dataset = nlp.load_dataset("race")
len(dataset["train"]), len(dataset["validation"])
```
But then I got the following error:
```
---------------------------------------------------------------------------
NonMatchingChecksumError Traceback (most recent call last)
<ipython-input-15-8bf7603ce0ed> in <module>
----> 1 dataset = nlp.load_dataset("race")
2 len(dataset["train"]), len(dataset["validation"])
~/miniconda3/envs/masters/lib/python3.8/site-packages/nlp/load.py in load_dataset(path, name, version, data_dir, data_files, split, cache_dir, features, download_config, download_mode, ignore_verifications, save_infos, **config_kwargs)
546
547 # Download and prepare data
--> 548 builder_instance.download_and_prepare(
549 download_config=download_config, download_mode=download_mode, ignore_verifications=ignore_verifications,
550 )
~/miniconda3/envs/masters/lib/python3.8/site-packages/nlp/builder.py in download_and_prepare(self, download_config, download_mode, ignore_verifications, try_from_hf_gcs, dl_manager, **download_and_prepare_kwargs)
460 logger.info("Dataset not on Hf google storage. Downloading and preparing it from source")
461 if not downloaded_from_gcs:
--> 462 self._download_and_prepare(
463 dl_manager=dl_manager, verify_infos=verify_infos, **download_and_prepare_kwargs
464 )
~/miniconda3/envs/masters/lib/python3.8/site-packages/nlp/builder.py in _download_and_prepare(self, dl_manager, verify_infos, **prepare_split_kwargs)
519 # Checksums verification
520 if verify_infos:
--> 521 verify_checksums(
522 self.info.download_checksums, dl_manager.get_recorded_sizes_checksums(), "dataset source files"
523 )
~/miniconda3/envs/masters/lib/python3.8/site-packages/nlp/utils/info_utils.py in verify_checksums(expected_checksums, recorded_checksums, verification_name)
36 if len(bad_urls) > 0:
37 error_msg = "Checksums didn't match" + for_verification_name + ":\n"
---> 38 raise NonMatchingChecksumError(error_msg + str(bad_urls))
39 logger.info("All the checksums matched successfully" + for_verification_name)
40
NonMatchingChecksumError: Checksums didn't match for dataset source files:
['http://www.cs.cmu.edu/~glai1/data/race/RACE.tar.gz']
``` | https://github.com/huggingface/datasets/issues/537 | [
"`NonMatchingChecksumError` means that the checksum of the downloaded file is not the expected one.\r\nEither the file you downloaded was corrupted along the way, or the host updated the file.\r\nCould you try to clear your cache and run `load_dataset` again ? If the error is still there, it means that there was an... | null | 537 | false |
Fingerprint | This PR is a continuation of #513 , in which many in-place functions were introduced or updated (cast_, flatten_) etc.
However the caching didn't handle these changes. Indeed the caching took into account only the previous cache file name of the table, and not the possible in-place transforms of the table.
To fix that, I added the concept of dataset fingerprint, that is updated after each transform (in place or not), and stored inside the table metadata.
When a dataset is created, an initial fingerprint is computed. If the dataset is memory-mapped, then the fingerprint generator doesn't read the table and only looks at the filename. However if the table is in-memory, then the fingerprint generator reads the content of the table using a batched non-crypto hashing.
I added a utility class to compute hashes of arbitrary python objects in `fingerprint.py` : `Hasher`. The API is close to standard hashing tools (`.update`, `.hexdigest`). It also supports custom hashing functions depending on object types using a registry like pickle. I added a custom hashing function to hash a `pa.Table` in a batched way, and also for `nlp.DatasetInfo` to leverage its json serialization feature.
Note about this PR:
This is a draft PR because #513 needs to be merged first.
The diff that is shown is for branches fingerprint -> indices (and not master, for now) | https://github.com/huggingface/datasets/pull/536 | [
"I changed the way I implemented fingerprint updates to use decorator functions.\r\n\r\nI also added a new attribute called `_inplace_history` that stores the in-place history of transforms (like cast_, rename_columns, etc.). This history is useful to replay the changes that were done in-place when unpickling a dat... | {
"url": "https://api.github.com/repos/huggingface/datasets/pulls/536",
"html_url": "https://github.com/huggingface/datasets/pull/536",
"diff_url": "https://github.com/huggingface/datasets/pull/536.diff",
"patch_url": "https://github.com/huggingface/datasets/pull/536.patch",
"merged_at": "2020-08-31T14:20:39"
} | 536 | true |
Benchmarks | Adding some benchmarks with DVC/CML
To add a new tracked benchmark:
- create a new python benchmarking script in `./benchmarks/`. The script can use the utilities in `./benchmarks/utils.py` and should output a JSON file with results in `./benchmarks/results/`.
- add a new pipeline stage in [dvc.yaml](./dvc.yaml) with the name of your new benchmark.
That's it | https://github.com/huggingface/datasets/pull/535 | [] | {
"url": "https://api.github.com/repos/huggingface/datasets/pulls/535",
"html_url": "https://github.com/huggingface/datasets/pull/535",
"diff_url": "https://github.com/huggingface/datasets/pull/535.diff",
"patch_url": "https://github.com/huggingface/datasets/pull/535.patch",
"merged_at": "2020-08-27T08:39:59"
} | 535 | true |
`list_datasets()` is broken. | version = '0.4.0'
`list_datasets()` is broken. It results in the following error :
```
In [3]: nlp.list_datasets()
Out[3]: ---------------------------------------------------------------------------
AttributeError Traceback (most recent call last)
~/.virtualenvs/san-lgUCsFg_/lib/python3.8/site-packages/IPython/core/formatters.py in __call__(self, obj)
700 type_pprinters=self.type_printers,
701 deferred_pprinters=self.deferred_printers)
--> 702 printer.pretty(obj)
703 printer.flush()
704 return stream.getvalue()
~/.virtualenvs/san-lgUCsFg_/lib/python3.8/site-packages/IPython/lib/pretty.py in pretty(self, obj)
375 if cls in self.type_pprinters:
376 # printer registered in self.type_pprinters
--> 377 return self.type_pprinters[cls](obj, self, cycle)
378 else:
379 # deferred printer
~/.virtualenvs/san-lgUCsFg_/lib/python3.8/site-packages/IPython/lib/pretty.py in inner(obj, p, cycle)
553 p.text(',')
554 p.breakable()
--> 555 p.pretty(x)
556 if len(obj) == 1 and type(obj) is tuple:
557 # Special case for 1-item tuples.
~/.virtualenvs/san-lgUCsFg_/lib/python3.8/site-packages/IPython/lib/pretty.py in pretty(self, obj)
392 if cls is not object \
393 and callable(cls.__dict__.get('__repr__')):
--> 394 return _repr_pprint(obj, self, cycle)
395
396 return _default_pprint(obj, self, cycle)
~/.virtualenvs/san-lgUCsFg_/lib/python3.8/site-packages/IPython/lib/pretty.py in _repr_pprint(obj, p, cycle)
698 """A pprint that just redirects to the normal repr function."""
699 # Find newlines and replace them with p.break_()
--> 700 output = repr(obj)
701 lines = output.splitlines()
702 with p.group():
~/.virtualenvs/san-lgUCsFg_/lib/python3.8/site-packages/nlp/hf_api.py in __repr__(self)
110
111 def __repr__(self):
--> 112 single_line_description = self.description.replace("\n", "")
113 return f"nlp.ObjectInfo(id='{self.id}', description='{single_line_description}', files={self.siblings})"
114
AttributeError: 'NoneType' object has no attribute 'replace'
``` | https://github.com/huggingface/datasets/issues/534 | [
"Thanks for reporting !\r\nThis has been fixed in #475 and the fix will be available in the next release",
"What you can do instead to get the list of the datasets is call\r\n\r\n```python\r\nprint([dataset.id for dataset in nlp.list_datasets()])\r\n```",
"Thanks @lhoestq . "
] | null | 534 | false |
Fix ArrayXD for pyarrow 0.17.1 by using non fixed length list arrays | It should fix the CI problems in #513 | https://github.com/huggingface/datasets/pull/533 | [] | {
"url": "https://api.github.com/repos/huggingface/datasets/pulls/533",
"html_url": "https://github.com/huggingface/datasets/pull/533",
"diff_url": "https://github.com/huggingface/datasets/pull/533.diff",
"patch_url": "https://github.com/huggingface/datasets/pull/533.patch",
"merged_at": "2020-08-26T08:02:23"
} | 533 | true |
File exists error when used with TPU | Hi,
I'm getting a "File exists" error when I use [text dataset](https://github.com/huggingface/nlp/tree/master/datasets/text) for pre-training a RoBERTa model using `transformers` (3.0.2) and `nlp`(0.4.0) on a VM with TPU (v3-8).
I modified [line 131 in the original `run_language_modeling.py`](https://github.com/huggingface/transformers/blob/master/examples/language-modeling/run_language_modeling.py#L131) as follows:
```python
# line 131: return LineByLineTextDataset(tokenizer=tokenizer, file_path=file_path, block_size=args.block_size)
dataset = load_dataset("text", data_files=file_path, split="train")
dataset = dataset.map(lambda ex: tokenizer(ex["text"], add_special_tokens=True,
truncation=True, max_length=args.block_size), batched=True)
dataset.set_format(type='torch', columns=['input_ids'])
return dataset
```
When I run this with [`xla_spawn.py`](https://github.com/huggingface/transformers/blob/master/examples/xla_spawn.py), I get the following error (it produces one message per core in TPU, which I believe is fine).
It seems the current version doesn't take into account distributed training processes as in [this example](https://github.com/huggingface/transformers/blob/a573777901e662ec2e565be312ffaeedef6effec/src/transformers/data/datasets/language_modeling.py#L35-L38)?
```
08/25/2020 13:59:41 - WARNING - nlp.builder - Using custom data configuration default
08/25/2020 13:59:43 - INFO - nlp.builder - Generating dataset text (/home/*****/.cache/huggingface/datasets/text/default-b0932b2bdbb63283/0.0.0/447f2bcfa2a721a37bc8fdf23800eade1523cf07f7eada6fe661fe4d070d380d)
08/25/2020 13:59:43 - INFO - nlp.builder - Generating dataset text (/home/*****/.cache/huggingface/datasets/text/default-b0932b2bdbb63283/0.0.0/447f2bcfa2a721a37bc8fdf23800eade1523cf07f7eada6fe661fe4d070d380d)
08/25/2020 13:59:43 - INFO - nlp.builder - Generating dataset text (/home/*****/.cache/huggingface/datasets/text/default-b0932b2bdbb63283/0.0.0/447f2bcfa2a721a37bc8fdf23800eade1523cf07f7eada6fe661fe4d070d380d)
08/25/2020 13:59:43 - INFO - nlp.builder - Generating dataset text (/home/*****/.cache/huggingface/datasets/text/default-b0932b2bdbb63283/0.0.0/447f2bcfa2a721a37bc8fdf23800eade1523cf07f7eada6fe661fe4d070d380d)
08/25/2020 13:59:43 - INFO - nlp.builder - Generating dataset text (/home/*****/.cache/huggingface/datasets/text/default-b0932b2bdbb63283/0.0.0/447f2bcfa2a721a37bc8fdf23800eade1523cf07f7eada6fe661fe4d070d380d)
08/25/2020 13:59:43 - INFO - nlp.builder - Generating dataset text (/home/*****/.cache/huggingface/datasets/text/default-b0932b2bdbb63283/0.0.0/447f2bcfa2a721a37bc8fdf23800eade1523cf07f7eada6fe661fe4d070d380d)
08/25/2020 13:59:43 - INFO - nlp.builder - Generating dataset text (/home/*****/.cache/huggingface/datasets/text/default-b0932b2bdbb63283/0.0.0/447f2bcfa2a721a37bc8fdf23800eade1523cf07f7eada6fe661fe4d070d380d)
08/25/2020 13:59:43 - INFO - nlp.builder - Generating dataset text (/home/*****/.cache/huggingface/datasets/text/default-b0932b2bdbb63283/0.0.0/447f2bcfa2a721a37bc8fdf23800eade1523cf07f7eada6fe661fe4d070d380d)
Downloading and preparing dataset text/default-b0932b2bdbb63283 (download: Unknown size, generated: Unknown size, post-processed: Unknown size, total: Unknown size) to /home/*****/.cache/huggingface/datasets/text/default-b0932b2bdbb63283/0.0.0/
447f2bcfa2a721a37bc8fdf23800eade1523cf07f7eada6fe661fe4d070d380d...
Downloading and preparing dataset text/default-b0932b2bdbb63283 (download: Unknown size, generated: Unknown size, post-processed: Unknown size, total: Unknown size) to /home/*****/.cache/huggingface/datasets/text/default-b0932b2bdbb63283/0.0.0/
447f2bcfa2a721a37bc8fdf23800eade1523cf07f7eada6fe661fe4d070d380d...
Downloading and preparing dataset text/default-b0932b2bdbb63283 (download: Unknown size, generated: Unknown size, post-processed: Unknown size, total: Unknown size) to /home/*****/.cache/huggingface/datasets/text/default-b0932b2bdbb63283/0.0.0/
447f2bcfa2a721a37bc8fdf23800eade1523cf07f7eada6fe661fe4d070d380d...
Downloading and preparing dataset text/default-b0932b2bdbb63283 (download: Unknown size, generated: Unknown size, post-processed: Unknown size, total: Unknown size) to /home/*****/.cache/huggingface/datasets/text/default-b0932b2bdbb63283/0.0.0/
447f2bcfa2a721a37bc8fdf23800eade1523cf07f7eada6fe661fe4d070d380d...
Downloading and preparing dataset text/default-b0932b2bdbb63283 (download: Unknown size, generated: Unknown size, post-processed: Unknown size, total: Unknown size) to /home/*****/.cache/huggingface/datasets/text/default-b0932b2bdbb63283/0.0.0/
447f2bcfa2a721a37bc8fdf23800eade1523cf07f7eada6fe661fe4d070d380d...
Downloading and preparing dataset text/default-b0932b2bdbb63283 (download: Unknown size, generated: Unknown size, post-processed: Unknown size, total: Unknown size) to /home/*****/.cache/huggingface/datasets/text/default-b0932b2bdbb63283/0.0.0/
447f2bcfa2a721a37bc8fdf23800eade1523cf07f7eada6fe661fe4d070d380d...
Exception in device=TPU:6: [Errno 17] File exists: '/home/*****/.cache/huggingface/datasets/text/default-b0932b2bdbb63283/0.0.0/447f2bcfa2a721a37bc8fdf23800eade1523cf07f7eada6fe661fe4d070d380d.incomplete'
Exception in device=TPU:4: [Errno 17] File exists: '/home/*****/.cache/huggingface/datasets/text/default-b0932b2bdbb63283/0.0.0/447f2bcfa2a721a37bc8fdf23800eade1523cf07f7eada6fe661fe4d070d380d.incomplete'
Exception in device=TPU:1: [Errno 17] File exists: '/home/*****/.cache/huggingface/datasets/text/default-b0932b2bdbb63283/0.0.0/447f2bcfa2a721a37bc8fdf23800eade1523cf07f7eada6fe661fe4d070d380d.incomplete'
Downloading and preparing dataset text/default-b0932b2bdbb63283 (download: Unknown size, generated: Unknown size, post-processed: Unknown size, total: Unknown size) to /home/*****/.cache/huggingface/datasets/text/default-b0932b2bdbb63283/0.0.0/
447f2bcfa2a721a37bc8fdf23800eade1523cf07f7eada6fe661fe4d070d380d...
Exception in device=TPU:7: [Errno 17] File exists: '/home/*****/.cache/huggingface/datasets/text/default-b0932b2bdbb63283/0.0.0/447f2bcfa2a721a37bc8fdf23800eade1523cf07f7eada6fe661fe4d070d380d.incomplete'
Exception in device=TPU:3: [Errno 17] File exists: '/home/*****/.cache/huggingface/datasets/text/default-b0932b2bdbb63283/0.0.0/447f2bcfa2a721a37bc8fdf23800eade1523cf07f7eada6fe661fe4d070d380d.incomplete'
Downloading and preparing dataset text/default-b0932b2bdbb63283 (download: Unknown size, generated: Unknown size, post-processed: Unknown size, total: Unknown size) to /home/*****/.cache/huggingface/datasets/text/default-b0932b2bdbb63283/0.0.0/
447f2bcfa2a721a37bc8fdf23800eade1523cf07f7eada6fe661fe4d070d380d...
Exception in device=TPU:2: [Errno 17] File exists: '/home/*****/.cache/huggingface/datasets/text/default-b0932b2bdbb63283/0.0.0/447f2bcfa2a721a37bc8fdf23800eade1523cf07f7eada6fe661fe4d070d380d.incomplete'
Exception in device=TPU:0: [Errno 17] File exists: '/home/*****/.cache/huggingface/datasets/text/default-b0932b2bdbb63283/0.0.0/447f2bcfa2a721a37bc8fdf23800eade1523cf07f7eada6fe661fe4d070d380d.incomplete'
Traceback (most recent call last):
File "/anaconda3/envs/torch-xla-1.6/lib/python3.6/site-packages/torch_xla/distributed/xla_multiprocessing.py", line 231, in _start_fn
fn(gindex, *args)
File "/anaconda3/envs/torch-xla-1.6/lib/python3.6/site-packages/torch_xla/distributed/xla_multiprocessing.py", line 231, in _start_fn
fn(gindex, *args)
File "/anaconda3/envs/torch-xla-1.6/lib/python3.6/site-packages/torch_xla/distributed/xla_multiprocessing.py", line 231, in _start_fn
fn(gindex, *args)
File "/home/*****/huggingface_roberta/run_language_modeling.py", line 300, in _mp_fn
main()
File "/home/*****/huggingface_roberta/run_language_modeling.py", line 300, in _mp_fn
main()
File "/home/*****/huggingface_roberta/run_language_modeling.py", line 300, in _mp_fn
main()
File "/home/*****/huggingface_roberta/run_language_modeling.py", line 240, in main
train_dataset = get_dataset(data_args, tokenizer=tokenizer) if training_args.do_train else None
File "/home/*****/huggingface_roberta/run_language_modeling.py", line 240, in main
train_dataset = get_dataset(data_args, tokenizer=tokenizer) if training_args.do_train else None
File "/home/*****/huggingface_roberta/run_language_modeling.py", line 240, in main
train_dataset = get_dataset(data_args, tokenizer=tokenizer) if training_args.do_train else None
File "/home/*****/huggingface_roberta/run_language_modeling.py", line 134, in get_dataset
dataset = load_dataset("text", data_files=file_path, split="train")
File "/anaconda3/envs/torch-xla-1.6/lib/python3.6/site-packages/nlp/load.py", line 546, in load_dataset
download_config=download_config, download_mode=download_mode, ignore_verifications=ignore_verifications,
File "/home/*****/huggingface_roberta/run_language_modeling.py", line 134, in get_dataset
dataset = load_dataset("text", data_files=file_path, split="train")
File "/home/*****/huggingface_roberta/run_language_modeling.py", line 134, in get_dataset
dataset = load_dataset("text", data_files=file_path, split="train")
File "/anaconda3/envs/torch-xla-1.6/lib/python3.6/site-packages/nlp/builder.py", line 450, in download_and_prepare
with incomplete_dir(self._cache_dir) as tmp_data_dir:
Traceback (most recent call last):
File "/anaconda3/envs/torch-xla-1.6/lib/python3.6/site-packages/nlp/load.py", line 546, in load_dataset
download_config=download_config, download_mode=download_mode, ignore_verifications=ignore_verifications,
File "/anaconda3/envs/torch-xla-1.6/lib/python3.6/contextlib.py", line 81, in __enter__
return next(self.gen)
File "/anaconda3/envs/torch-xla-1.6/lib/python3.6/site-packages/nlp/load.py", line 546, in load_dataset
download_config=download_config, download_mode=download_mode, ignore_verifications=ignore_verifications,
File "/anaconda3/envs/torch-xla-1.6/lib/python3.6/site-packages/torch_xla/distributed/xla_multiprocessing.py", line 231, in _start_fn
fn(gindex, *args)
File "/anaconda3/envs/torch-xla-1.6/lib/python3.6/site-packages/nlp/builder.py", line 450, in download_and_prepare
with incomplete_dir(self._cache_dir) as tmp_data_dir:
File "/anaconda3/envs/torch-xla-1.6/lib/python3.6/site-packages/nlp/builder.py", line 422, in incomplete_dir
os.makedirs(tmp_dir)
File "/anaconda3/envs/torch-xla-1.6/lib/python3.6/site-packages/nlp/builder.py", line 450, in download_and_prepare
with incomplete_dir(self._cache_dir) as tmp_data_dir:
File "/home/*****/huggingface_roberta/run_language_modeling.py", line 300, in _mp_fn
main()
File "/anaconda3/envs/torch-xla-1.6/lib/python3.6/contextlib.py", line 81, in __enter__
return next(self.gen)
File "/anaconda3/envs/torch-xla-1.6/lib/python3.6/os.py", line 220, in makedirs
mkdir(name, mode)
File "/anaconda3/envs/torch-xla-1.6/lib/python3.6/contextlib.py", line 81, in __enter__
return next(self.gen)
File "/home/*****/huggingface_roberta/run_language_modeling.py", line 240, in main
train_dataset = get_dataset(data_args, tokenizer=tokenizer) if training_args.do_train else None
File "/anaconda3/envs/torch-xla-1.6/lib/python3.6/site-packages/nlp/builder.py", line 422, in incomplete_dir
os.makedirs(tmp_dir)
File "/anaconda3/envs/torch-xla-1.6/lib/python3.6/site-packages/torch_xla/distributed/xla_multiprocessing.py", line 231, in _start_fn
fn(gindex, *args)
File "/anaconda3/envs/torch-xla-1.6/lib/python3.6/site-packages/nlp/builder.py", line 422, in incomplete_dir
os.makedirs(tmp_dir)
File "/home/*****/huggingface_roberta/run_language_modeling.py", line 134, in get_dataset
dataset = load_dataset("text", data_files=file_path, split="train")
File "/anaconda3/envs/torch-xla-1.6/lib/python3.6/os.py", line 220, in makedirs
mkdir(name, mode)
File "/anaconda3/envs/torch-xla-1.6/lib/python3.6/site-packages/nlp/load.py", line 546, in load_dataset
download_config=download_config, download_mode=download_mode, ignore_verifications=ignore_verifications,
FileExistsError: [Errno 17] File exists: '/home/*****/.cache/huggingface/datasets/text/default-b0932b2bdbb63283/0.0.0/447f2bcfa2a721a37bc8fdf23800eade1523cf07f7eada6fe661fe4d070d380d.incomplete'
File "/home/*****/huggingface_roberta/run_language_modeling.py", line 300, in _mp_fn
main()
File "/anaconda3/envs/torch-xla-1.6/lib/python3.6/site-packages/nlp/builder.py", line 450, in download_and_prepare
with incomplete_dir(self._cache_dir) as tmp_data_dir:
File "/anaconda3/envs/torch-xla-1.6/lib/python3.6/os.py", line 220, in makedirs
mkdir(name, mode)
FileExistsError: [Errno 17] File exists: '/home/*****/.cache/huggingface/datasets/text/default-b0932b2bdbb63283/0.0.0/447f2bcfa2a721a37bc8fdf23800eade1523cf07f7eada6fe661fe4d070d380d.incomplete'
File "/home/*****/huggingface_roberta/run_language_modeling.py", line 240, in main
train_dataset = get_dataset(data_args, tokenizer=tokenizer) if training_args.do_train else None
File "/anaconda3/envs/torch-xla-1.6/lib/python3.6/contextlib.py", line 81, in __enter__
return next(self.gen)
FileExistsError: [Errno 17] File exists: '/home/*****/.cache/huggingface/datasets/text/default-b0932b2bdbb63283/0.0.0/447f2bcfa2a721a37bc8fdf23800eade1523cf07f7eada6fe661fe4d070d380d.incomplete'
File "/home/*****/huggingface_roberta/run_language_modeling.py", line 134, in get_dataset
dataset = load_dataset("text", data_files=file_path, split="train")
File "/anaconda3/envs/torch-xla-1.6/lib/python3.6/site-packages/nlp/builder.py", line 422, in incomplete_dir
os.makedirs(tmp_dir)
File "/anaconda3/envs/torch-xla-1.6/lib/python3.6/site-packages/nlp/load.py", line 546, in load_dataset
download_config=download_config, download_mode=download_mode, ignore_verifications=ignore_verifications,
File "/anaconda3/envs/torch-xla-1.6/lib/python3.6/os.py", line 220, in makedirs
mkdir(name, mode)
File "/anaconda3/envs/torch-xla-1.6/lib/python3.6/site-packages/nlp/builder.py", line 450, in download_and_prepare
with incomplete_dir(self._cache_dir) as tmp_data_dir:
File "/anaconda3/envs/torch-xla-1.6/lib/python3.6/contextlib.py", line 81, in __enter__
return next(self.gen)
FileExistsError: [Errno 17] File exists: '/home/*****/.cache/huggingface/datasets/text/default-b0932b2bdbb63283/0.0.0/447f2bcfa2a721a37bc8fdf23800eade1523cf07f7eada6fe661fe4d070d380d.incomplete'
File "/anaconda3/envs/torch-xla-1.6/lib/python3.6/site-packages/nlp/builder.py", line 422, in incomplete_dir
os.makedirs(tmp_dir)
File "/anaconda3/envs/torch-xla-1.6/lib/python3.6/os.py", line 220, in makedirs
mkdir(name, mode)
Traceback (most recent call last):
FileExistsError: [Errno 17] File exists: '/home/*****/.cache/huggingface/datasets/text/default-b0932b2bdbb63283/0.0.0/447f2bcfa2a721a37bc8fdf23800eade1523cf07f7eada6fe661fe4d070d380d.incomplete'
Traceback (most recent call last):
File "/anaconda3/envs/torch-xla-1.6/lib/python3.6/site-packages/torch_xla/distributed/xla_multiprocessing.py", line 231, in _start_fn
fn(gindex, *args)
File "/home/*****/huggingface_roberta/run_language_modeling.py", line 300, in _mp_fn
main()
File "/home/*****/huggingface_roberta/run_language_modeling.py", line 240, in main
train_dataset = get_dataset(data_args, tokenizer=tokenizer) if training_args.do_train else None
File "/home/*****/huggingface_roberta/run_language_modeling.py", line 134, in get_dataset
dataset = load_dataset("text", data_files=file_path, split="train")
File "/anaconda3/envs/torch-xla-1.6/lib/python3.6/site-packages/torch_xla/distributed/xla_multiprocessing.py", line 231, in _start_fn
fn(gindex, *args)
File "/anaconda3/envs/torch-xla-1.6/lib/python3.6/site-packages/nlp/load.py", line 546, in load_dataset
download_config=download_config, download_mode=download_mode, ignore_verifications=ignore_verifications,
File "/anaconda3/envs/torch-xla-1.6/lib/python3.6/site-packages/nlp/builder.py", line 450, in download_and_prepare
with incomplete_dir(self._cache_dir) as tmp_data_dir:
File "/home/*****/huggingface_roberta/run_language_modeling.py", line 300, in _mp_fn
main()
File "/home/*****/huggingface_roberta/run_language_modeling.py", line 240, in main
train_dataset = get_dataset(data_args, tokenizer=tokenizer) if training_args.do_train else None
File "/home/*****/huggingface_roberta/run_language_modeling.py", line 134, in get_dataset
dataset = load_dataset("text", data_files=file_path, split="train")
File "/anaconda3/envs/torch-xla-1.6/lib/python3.6/site-packages/nlp/load.py", line 546, in load_dataset
download_config=download_config, download_mode=download_mode, ignore_verifications=ignore_verifications,
File "/anaconda3/envs/torch-xla-1.6/lib/python3.6/site-packages/nlp/builder.py", line 450, in download_and_prepare
with incomplete_dir(self._cache_dir) as tmp_data_dir:
File "/anaconda3/envs/torch-xla-1.6/lib/python3.6/contextlib.py", line 81, in __enter__
return next(self.gen)
File "/anaconda3/envs/torch-xla-1.6/lib/python3.6/site-packages/nlp/builder.py", line 422, in incomplete_dir
os.makedirs(tmp_dir)
File "/anaconda3/envs/torch-xla-1.6/lib/python3.6/os.py", line 220, in makedirs
mkdir(name, mode)
FileExistsError: [Errno 17] File exists: '/home/*****/.cache/huggingface/datasets/text/default-b0932b2bdbb63283/0.0.0/447f2bcfa2a721a37bc8fdf23800eade1523cf07f7eada6fe661fe4d070d380d.incomplete'
```
| https://github.com/huggingface/datasets/issues/532 | [
"I am facing probably facing similar issues with \r\n\r\n`wiki40b_en_100_0`",
"Could you try to run `dataset = load_dataset(\"text\", data_files=file_path, split=\"train\")` once before calling the script ?\r\n\r\nIt looks like several processes try to create the dataset in arrow format at the same time. If the d... | null | 532 | false |
add concatenate_datasets to the docs | https://github.com/huggingface/datasets/pull/531 | [] | {
"url": "https://api.github.com/repos/huggingface/datasets/pulls/531",
"html_url": "https://github.com/huggingface/datasets/pull/531",
"diff_url": "https://github.com/huggingface/datasets/pull/531.diff",
"patch_url": "https://github.com/huggingface/datasets/pull/531.patch",
"merged_at": "2020-08-25T09:02:19"
} | 531 | true | |
use ragged tensor by default | I think it's better if it's clear whether the returned tensor is ragged or not when the type is set to tensorflow.
Previously it was a tensor (not ragged) if numpy could stack the output (which can change depending on the batch of example you take), which make things difficult to handle, as it may sometimes return a ragged tensor and sometimes not.
Therefore I reverted this behavior to always return a ragged tensor as we used to do. | https://github.com/huggingface/datasets/pull/530 | [
"Yes I agree. Maybe something that lets specify different format depending on the column ? Especially to better control dtype and shape (and ragged for tf)\r\n\r\nOh and I forgot: this one should also fix the second issue found in #477 for the next release",
"I am running into the same issue with the error messag... | {
"url": "https://api.github.com/repos/huggingface/datasets/pulls/530",
"html_url": "https://github.com/huggingface/datasets/pull/530",
"diff_url": "https://github.com/huggingface/datasets/pull/530.diff",
"patch_url": "https://github.com/huggingface/datasets/pull/530.patch",
"merged_at": "2020-08-24T19:22:25"
} | 530 | true |
Add MLSUM | Hello (again :) !),
So, I started a new branch because of a [rebase issue](https://github.com/huggingface/nlp/pull/463), sorry for the mess.
However, the command `pytest tests/test_dataset_common.py::LocalDatasetTest::test_load_real_dataset_mlsum` still fails because there is no default language dataset : the script throws an error as a specific config language is necessary.
I think that setting a default language would be a bad workaround for this so I kept it as it is. Putting all the train files across languages together would also be a bad idea because of the size.
Thanks for your help,
Rachel
| https://github.com/huggingface/datasets/pull/529 | [
"Could you test to run the test using the changes in #527 and let me know if it fixes the issue ? If so I'll merge it and we'll be good to go :)",
"Hello, it does work on the fixing real dataset branch. Merci Quentin :)",
"Nice, glad to hear that :)\r\nde rien !"
] | {
"url": "https://api.github.com/repos/huggingface/datasets/pulls/529",
"html_url": "https://github.com/huggingface/datasets/pull/529",
"diff_url": "https://github.com/huggingface/datasets/pull/529.diff",
"patch_url": "https://github.com/huggingface/datasets/pull/529.patch",
"merged_at": "2020-08-26T08:04:10"
} | 529 | true |
fix missing variable names in docs | fix #524 | https://github.com/huggingface/datasets/pull/528 | [
"The problem came from `default: ` that is rendered differently and hides the parameter names. I changed `default: ...` to `defaults to ...`"
] | {
"url": "https://api.github.com/repos/huggingface/datasets/pulls/528",
"html_url": "https://github.com/huggingface/datasets/pull/528",
"diff_url": "https://github.com/huggingface/datasets/pull/528.diff",
"patch_url": "https://github.com/huggingface/datasets/pull/528.patch",
"merged_at": "2020-08-25T09:04:03"
} | 528 | true |
Fix config used for slow test on real dataset | As noticed in #470, #474, #476, #504 , the slow test `test_load_real_dataset` couldn't run on datasets that require config parameters.
To fix that I replaced it with one test with the first config of BUILDER_CONFIGS `test_load_real_dataset`, and another test that runs all of the configs in BUILDER_CONFIGS `test_load_real_dataset_all_configs` | https://github.com/huggingface/datasets/pull/527 | [] | {
"url": "https://api.github.com/repos/huggingface/datasets/pulls/527",
"html_url": "https://github.com/huggingface/datasets/pull/527",
"diff_url": "https://github.com/huggingface/datasets/pull/527.diff",
"patch_url": "https://github.com/huggingface/datasets/pull/527.patch",
"merged_at": "2020-08-25T09:20:44"
} | 527 | true |
Returning None instead of "python" if dataset is unformatted | Following the discussion on Slack, this small fix ensures that calling `dataset.set_format(type=dataset.format["type"])` works properly. Slightly breaking as calling `dataset.format` when the dataset is unformatted will return `None` instead of `python`. | https://github.com/huggingface/datasets/pull/526 | [
"We have to change the tests to expect `None` instead of `python` then",
"Merging!"
] | {
"url": "https://api.github.com/repos/huggingface/datasets/pulls/526",
"html_url": "https://github.com/huggingface/datasets/pull/526",
"diff_url": "https://github.com/huggingface/datasets/pull/526.diff",
"patch_url": "https://github.com/huggingface/datasets/pull/526.patch",
"merged_at": "2020-08-24T12:50:42"
} | 526 | true |
wmt download speed example | Continuing from the slack 1.0 roadmap thread w @lhoestq , I realized the slow downloads is only a thing sometimes. Here are a few examples, I suspect there are multiple issues. All commands were run from the same gcp us-central-1f machine.
```
import nlp
nlp.load_dataset('wmt16', 'de-en')
```
Downloads at 49.1 KB/S
Whereas
```
pip install gdown # download from google drive
!gdown https://drive.google.com/uc?id=1iO7um-HWoNoRKDtw27YUSgyeubn9uXqj
```
Downloads at 127 MB/s. (The file is a copy of wmt-en-de raw).
```
nlp.load_dataset('wmt16', 'ro-en')
```
goes at 27 MB/s, much faster.
if we wget the same data from s3 is the same download speed, but ¼ the file size:
```
wget https://s3.amazonaws.com/datasets.huggingface.co/translation/wmt_en_ro_packed_200_rand.tgz
```
Finally,
```
nlp.load_dataset('wmt19', 'zh-en')
```
Starts fast, but broken. (duplicate of #493 )
| https://github.com/huggingface/datasets/issues/525 | [
"Thanks for creating the issue :)\r\nThe download link for wmt-en-de raw looks like a mirror. We should use that instead of the current url.\r\nIs this mirror official ?\r\n\r\nAlso it looks like for `ro-en` it tried to download other languages. If we manage to only download the one that is asked it'd be cool\r\n\r... | null | 525 | false |
Some docs are missing parameter names | See https://huggingface.co/nlp/master/package_reference/main_classes.html#nlp.Dataset.map. I believe this is because the parameter names are enclosed in backticks in the docstrings, maybe it's an old docstring format that doesn't work with the current Sphinx version. | https://github.com/huggingface/datasets/issues/524 | [
"Indeed, good catch!"
] | null | 524 | false |
Speed up Tokenization by optimizing cast_to_python_objects | I changed how `cast_to_python_objects` works to make it faster.
It is used to cast numpy/pytorch/tensorflow/pandas objects to python lists, and it works recursively.
To avoid iterating over possibly long lists, it first checks if the first element that is not None has to be casted.
If the first element needs to be casted, then all the elements of the list will be casted, otherwise they'll stay the same.
This trick allows to cast objects that contain tokenizers outputs without iterating over every single token for example.
Speed improvement:
```python
import transformers
import nlp
tok = transformers.BertTokenizerFast.from_pretrained("bert-base-uncased")
txt = ["a " * 512] * 1000
dataset = nlp.Dataset.from_dict({"txt": txt})
# Tokenization using .map is now faster. Previously it was taking 3.5s
%time _ = dataset.map(lambda x: tok(x["txt"]), batched=True, load_from_cache_file=False)
# 450ms
# for comparison
%time _ = tok(txt)
# 280ms
``` | https://github.com/huggingface/datasets/pull/523 | [
"I took your comments into account and added tests for `cast_to_python_objects`"
] | {
"url": "https://api.github.com/repos/huggingface/datasets/pulls/523",
"html_url": "https://github.com/huggingface/datasets/pull/523",
"diff_url": "https://github.com/huggingface/datasets/pull/523.diff",
"patch_url": "https://github.com/huggingface/datasets/pull/523.patch",
"merged_at": "2020-08-24T08:54:14"
} | 523 | true |
dictionnary typo in docs | Many places dictionary is spelled dictionnary, not sure if its on purpose or not.
Fixed in this pr:
https://github.com/huggingface/nlp/pull/521 | https://github.com/huggingface/datasets/issues/522 | [
"Thanks!"
] | null | 522 | false |
Fix dictionnary (dictionary) typo | This error happens many times I'm thinking maybe its spelled like this on purpose? | https://github.com/huggingface/datasets/pull/521 | [
"Hahah thanks Yonatan. It was not on purpose, we are just not very good at spelling :)"
] | {
"url": "https://api.github.com/repos/huggingface/datasets/pulls/521",
"html_url": "https://github.com/huggingface/datasets/pull/521",
"diff_url": "https://github.com/huggingface/datasets/pull/521.diff",
"patch_url": "https://github.com/huggingface/datasets/pull/521.patch",
"merged_at": "2020-08-20T07:52:04"
} | 521 | true |
Transform references for sacrebleu | Currently it is impossible to use sacrebleu when len(predictions) != the number of references per prediction (very uncommon), due to a strange format expected by sacrebleu. If one passes in the data to `nlp.metric.compute()` in sacrebleu format, `nlp` throws an error due to mismatching lengths between predictions and references. If one uses a more standard format where predictions and references are lists of the same length, sacrebleu throws an error.
This PR transforms reference data in a more standard format into the [unusual format](https://github.com/mjpost/sacreBLEU#using-sacrebleu-from-python) expected by sacrebleu. | https://github.com/huggingface/datasets/pull/520 | [
"I think I agree @lhoestq so I pushed a change.\r\nThanks for your work on the library!"
] | {
"url": "https://api.github.com/repos/huggingface/datasets/pulls/520",
"html_url": "https://github.com/huggingface/datasets/pull/520",
"diff_url": "https://github.com/huggingface/datasets/pull/520.diff",
"patch_url": "https://github.com/huggingface/datasets/pull/520.patch",
"merged_at": "2020-08-20T09:30:53"
} | 520 | true |
[BUG] Metrics throwing new error on master since 0.4.0 | The following error occurs when passing in references of type `List[List[str]]` to metrics like bleu.
Wasn't happening on 0.4.0 but happening now on master.
```
File "/usr/local/lib/python3.7/site-packages/nlp/metric.py", line 226, in compute
self.add_batch(predictions=predictions, references=references)
File "/usr/local/lib/python3.7/site-packages/nlp/metric.py", line 242, in add_batch
batch = self.info.features.encode_batch(batch)
File "/usr/local/lib/python3.7/site-packages/nlp/features.py", line 527, in encode_batch
encoded_batch[key] = [encode_nested_example(self[key], cast_to_python_objects(obj)) for obj in column]
File "/usr/local/lib/python3.7/site-packages/nlp/features.py", line 527, in <listcomp>
encoded_batch[key] = [encode_nested_example(self[key], cast_to_python_objects(obj)) for obj in column]
File "/usr/local/lib/python3.7/site-packages/nlp/features.py", line 456, in encode_nested_example
raise ValueError("Got a string but expected a list instead: '{}'".format(obj))
``` | https://github.com/huggingface/datasets/issues/519 | [
"Update - maybe this is only failing on bleu because I was not tokenizing inputs to the metric",
"Closing - seems to be just forgetting to tokenize. And found the helpful discussion in huggingface/evaluate#105 "
] | null | 519 | false |
[METRICS, breaking] Refactor caching behavior, pickle/cloudpickle metrics and dataset, add tests on metrics | Move the acquisition of the filelock at a later stage during metrics processing so it can be pickled/cloudpickled after instantiation.
Also add some tests on pickling, concurrent but separate metric instances and concurrent and distributed metric instances.
Changes significantly the caching behavior for the metrics:
- if the metric is used in a non-distributed setup (most common case) we try to find a free cache file using UUID instead of asking for an `experiment_id` if we can't lock the cache file this allows to use several instances of the same metrics in parallel.
- if the metrics is used in a distributed setup we ask for an `experiment_id` if we can't lock the cache file (because all the nodes need to have related cache file names for the final sync.
- after the computation, we free the locks and delete all the cache files.
Breaking: Some arguments for Metrics initialization have been removed for simplicity (`version`...) and some have been renamed for consistency with the rest of the library (`in_memory` => `keep_in_memory`).
Also remove the `_has_transformers` detection in utils to avoid importing transformers everytime during loading. | https://github.com/huggingface/datasets/pull/518 | [
"(test failure is unrelated)",
"As discussed with @thomwolf merging since the hyperparameter-search has been merged in transformers."
] | {
"url": "https://api.github.com/repos/huggingface/datasets/pulls/518",
"html_url": "https://github.com/huggingface/datasets/pull/518",
"diff_url": "https://github.com/huggingface/datasets/pull/518.diff",
"patch_url": "https://github.com/huggingface/datasets/pull/518.patch",
"merged_at": "2020-08-24T16:01:39"
} | 518 | true |
add MLDoc dataset | Hi,
I am recommending that someone add MLDoc, a multilingual news topic classification dataset.
- Here's a link to the Github: https://github.com/facebookresearch/MLDoc
- and the paper: http://www.lrec-conf.org/proceedings/lrec2018/pdf/658.pdf
Looks like the dataset contains news stories in multiple languages that can be classified into four hierarchical groups: CCAT (Corporate/Industrial), ECAT (Economics), GCAT (Government/Social) and MCAT (Markets). There are 13 languages: Dutch, French, German, Chinese, Japanese, Russian, Portuguese, Spanish, Latin American Spanish, Italian, Danish, Norwegian, and Swedish | https://github.com/huggingface/datasets/issues/517 | [
"Any updates on this?",
"This request is still an open issue waiting to be addressed by any community member, @GuillemGSubies."
] | null | 517 | false |
[Breaking] Rename formated to formatted | `formated` is not correct but `formatted` is | https://github.com/huggingface/datasets/pull/516 | [] | {
"url": "https://api.github.com/repos/huggingface/datasets/pulls/516",
"html_url": "https://github.com/huggingface/datasets/pull/516",
"diff_url": "https://github.com/huggingface/datasets/pull/516.diff",
"patch_url": "https://github.com/huggingface/datasets/pull/516.patch",
"merged_at": "2020-08-20T08:41:16"
} | 516 | true |
Fix batched map for formatted dataset | If you had a dataset formatted as numpy for example, and tried to do a batched map, then it would crash because one of the elements from the inputs was missing for unchanged columns (ex: batch of length 999 instead of 1000).
The happened during the creation of the `pa.Table`, since columns had different lengths. | https://github.com/huggingface/datasets/pull/515 | [] | {
"url": "https://api.github.com/repos/huggingface/datasets/pulls/515",
"html_url": "https://github.com/huggingface/datasets/pull/515",
"diff_url": "https://github.com/huggingface/datasets/pull/515.diff",
"patch_url": "https://github.com/huggingface/datasets/pull/515.patch",
"merged_at": "2020-08-20T20:30:42"
} | 515 | true |
dataset.shuffle(keep_in_memory=True) is never allowed | As of commit ef4aac2, the usage of the parameter `keep_in_memory=True` is never possible: `dataset.select(keep_in_memory=True)`
The commit added the lines
```python
# lines 994-996 in src/nlp/arrow_dataset.py
assert (
not keep_in_memory or cache_file_name is None
), "Please use either `keep_in_memory` or `cache_file_name` but not both."
```
This affects both `shuffle()` as `select()` is a sub-routine, and `map()` that has the same check.
I'd love to fix this myself, but unsure what the intention of the assert is given the rest of the logic in the function concerning `ccache_file_name` and `keep_in_memory`. | https://github.com/huggingface/datasets/issues/514 | [
"This seems to be fixed in #513 for the filter function, replacing `cache_file_name` with `indices_cache_file_name` in the assert. Although not for the `map()` function @thomwolf ",
"Maybe I'm a bit tired but I fail to see the issue here.\r\n\r\nSince `cache_file_name` is `None` by default, if you set `keep_in_me... | null | 514 | false |
[speedup] Use indices mappings instead of deepcopy for all the samples reordering methods | Use an indices mapping instead of rewriting the dataset for all the samples re-ordering/selection methods (`select`, `sort`, `shuffle`, `shard`, `train_test_split`).
Added a `flatten_indices` method which copy the dataset to a new table to remove the indices mapping with tests.
All the samples re-ordering/selection methods should be a lot faster. The downside is that iterating on very large batch of the dataset might be a little slower when we have changed the order of the samples since with in these case we use `pyarrow.Table.take` instead of `pyarrow.Table.slice`. There is no free lunch but the speed of iterating over the dataset is rarely the bottleneck.
*Backward breaking change*: the `cache_file_name` argument in all the samples re-ordering/selection methods (`select`, `sort`, `shuffle`, `shard`, `train_test_split`) is now called `indices_cache_file_name` on purpose to make it explicit to the user that this caching file is used for caching the indices mapping and not the dataset itself. | https://github.com/huggingface/datasets/pull/513 | [
"Ok I fixed `concatenate_datasets` and added tests\r\nFeel free to merge if it's good for you @thomwolf ",
"Ok, adding some benchmarks for map/filters and then I'll merge",
"Warning from pytorch that we should maybe consider at some point @lhoestq:\r\n```\r\n/__w/nlp/nlp/src/nlp/arrow_dataset.py:648: UserWarnin... | {
"url": "https://api.github.com/repos/huggingface/datasets/pulls/513",
"html_url": "https://github.com/huggingface/datasets/pull/513",
"diff_url": "https://github.com/huggingface/datasets/pull/513.diff",
"patch_url": "https://github.com/huggingface/datasets/pull/513.patch",
"merged_at": "2020-08-28T08:41:50"
} | 513 | true |
Delete CONTRIBUTING.md | https://github.com/huggingface/datasets/pull/512 | [
"😱",
"Yeah, this is spammy behavior. I've reported the user handle."
] | {
"url": "https://api.github.com/repos/huggingface/datasets/pulls/512",
"html_url": "https://github.com/huggingface/datasets/pull/512",
"diff_url": "https://github.com/huggingface/datasets/pull/512.diff",
"patch_url": "https://github.com/huggingface/datasets/pull/512.patch",
"merged_at": null
} | 512 | true | |
dataset.shuffle() and select() resets format. Intended? | Calling `dataset.shuffle()` or `dataset.select()` on a dataset resets its format set by `dataset.set_format()`. Is this intended or an oversight?
When working on quite large datasets that require a lot of preprocessing I find it convenient to save the processed dataset to file using `torch.save("dataset.pt")`. Later loading the dataset object using `torch.load("dataset.pt")`, which conserves the defined format before saving.
I do shuffling and selecting (for controlling dataset size) after loading the data from .pt-file, as it's convenient whenever you train multiple models with varying sizes of the same dataset.
The obvious workaround for this is to set the format again after using `dataset.select()` or `dataset.shuffle()`.
_I guess this is more of a discussion on the design philosophy of the functions. Please let me know if this is not the right channel for these kinds of discussions or if they are not wanted at all!_
#### How to reproduce:
```python
import nlp
from transformers import T5Tokenizer
tokenizer = T5Tokenizer.from_pretrained("t5-base")
def create_features(batch):
context_encoding = tokenizer.batch_encode_plus(batch["context"])
return {"input_ids": context_encoding["input_ids"]}
dataset = nlp.load_dataset("cosmos_qa", split="train")
dataset = dataset.map(create_features, batched=True)
dataset.set_format(type="torch", columns=["input_ids"])
dataset[0]
# {'input_ids': tensor([ 1804, 3525, 1602, ... 0, 0])}
dataset = dataset.shuffle()
dataset[0]
# {'id': '3Q9(...)20', 'context': "Good Old War an (...) play ?', 'answer0': 'None of the above choices .', 'answer1': 'This person likes music and likes to see the show , they will see other bands play .', (...) 'input_ids': [1804, 3525, 1602, ... , 0, 0]}
``` | https://github.com/huggingface/datasets/issues/511 | [
"Hi @vegarab yes feel free to open a discussion here.\r\n\r\nThis design choice was not very much thought about.\r\n\r\nSince `dataset.select()` (like all the method without a trailing underscore) is non-destructive and returns a new dataset it has most of its properties initialized from scratch (except the table a... | null | 511 | false |
Version of numpy to use the library | Thank you so much for your excellent work! I would like to use nlp library in my project. While importing nlp, I am receiving the following error `AttributeError: module 'numpy.random' has no attribute 'Generator'` Numpy version in my project is 1.16.0. May I learn which numpy version is used for the nlp library.
Thanks in advance. | https://github.com/huggingface/datasets/issues/510 | [
"Seems like this method was added in 1.17. I'll add a requirement on this.",
"Thank you so much. After upgrading the numpy library, it worked."
] | null | 510 | false |
Converting TensorFlow dataset example | Hi,
I want to use TensorFlow datasets with this repo, I noticed you made some conversion script,
can you give a simple example of using it?
Thanks
| https://github.com/huggingface/datasets/issues/509 | [
"Do you want to convert a dataset script to the tfds format ?\r\nIf so, we currently have a comversion script nlp/commands/convert.py but it is a conversion script that goes from tfds to nlp.\r\nI think it shouldn't be too hard to do the changes in reverse (at some manual adjustments).\r\nIf you manage to make it w... | null | 509 | false |
TypeError: Receiver() takes no arguments | I am trying to load a wikipedia data set
```
import nlp
from nlp import load_dataset
dataset = load_dataset("wikipedia", "20200501.en", split="train", cache_dir=data_path, beam_runner='DirectRunner')
#dataset = load_dataset('wikipedia', '20200501.sv', cache_dir=data_path, beam_runner='DirectRunner')
```
This fails in the apache beam runner.
```
Traceback (most recent call last):
File "D:/ML/wikiembedding/gpt2_sv.py", line 36, in <module>
dataset = load_dataset("wikipedia", "20200501.en", split="train", cache_dir=my_cache_dir, beam_runner='DirectRunner')
File "C:\Users\seto\AppData\Local\Programs\Python\Python38\lib\site-packages\nlp\load.py", line 548, in load_dataset
builder_instance.download_and_prepare(
File "C:\Users\seto\AppData\Local\Programs\Python\Python38\lib\site-packages\nlp\builder.py", line 462, in download_and_prepare
self._download_and_prepare(
File "C:\Users\seto\AppData\Local\Programs\Python\Python38\lib\site-packages\nlp\builder.py", line 969, in _download_and_prepare
pipeline_results = pipeline.run()
File "C:\Users\seto\AppData\Local\Programs\Python\Python38\lib\site-packages\apache_beam\pipeline.py", line 534, in run
return self.runner.run_pipeline(self, self._options)
....
File "C:\Users\seto\AppData\Local\Programs\Python\Python38\lib\site-packages\apache_beam\runners\worker\bundle_processor.py", line 218, in process_encoded
self.output(decoded_value)
File "C:\Users\seto\AppData\Local\Programs\Python\Python38\lib\site-packages\apache_beam\runners\worker\operations.py", line 332, in output
cython.cast(Receiver, self.receivers[output_index]).receive(windowed_value)
File "C:\Users\seto\AppData\Local\Programs\Python\Python38\lib\site-packages\Cython\Shadow.py", line 167, in cast
return type(*args)
TypeError: Receiver() takes no arguments
```
This is run on a Windows 10 machine with python 3.8. I get the same error loading the swedish wikipedia dump. | https://github.com/huggingface/datasets/issues/508 | [
"Which version of Apache Beam do you have (can you copy your full environment info here)?",
"apache-beam==2.23.0\r\nnlp==0.4.0\r\n\r\nFor me this was resolved by running the same python script on Linux (or really WSL). ",
"Do you manage to run a dummy beam pipeline with python on windows ? \r\nYou can test a du... | null | 508 | false |
Errors when I use | I tried the following example code from https://huggingface.co/deepset/roberta-base-squad2 and got errors
I am using **transformers 3.0.2** code .
from transformers.pipelines import pipeline
from transformers.modeling_auto import AutoModelForQuestionAnswering
from transformers.tokenization_auto import AutoTokenizer
model_name = "deepset/roberta-base-squad2"
nlp = pipeline('question-answering', model=model_name, tokenizer=model_name)
QA_input = {
'question': 'Why is model conversion important?',
'context': 'The option to convert models between FARM and transformers gives freedom to the user and let people easily switch between frameworks.'
}
res = nlp(QA_input)
The errors are :
res = nlp(QA_input)
File ".local/lib/python3.6/site-packages/transformers/pipelines.py", line 1316, in __call__
for s, e, score in zip(starts, ends, scores)
File ".local/lib/python3.6/site-packages/transformers/pipelines.py", line 1316, in <listcomp>
for s, e, score in zip(starts, ends, scores)
KeyError: 0
| https://github.com/huggingface/datasets/issues/507 | [
"Looks like an issue with 3.0.2 transformers version. Works fine when I use \"master\" version of transformers."
] | null | 507 | false |
fix dataset.map for function without outputs | As noticed in #505 , giving a function that doesn't return anything in `.map` raises an error because of an unreferenced variable.
I fixed that and added tests.
Thanks @avloss for reporting | https://github.com/huggingface/datasets/pull/506 | [] | {
"url": "https://api.github.com/repos/huggingface/datasets/pulls/506",
"html_url": "https://github.com/huggingface/datasets/pull/506",
"diff_url": "https://github.com/huggingface/datasets/pull/506.diff",
"patch_url": "https://github.com/huggingface/datasets/pull/506.patch",
"merged_at": "2020-08-17T11:24:38"
} | 506 | true |
tmp_file referenced before assignment | Just learning about this library - so might've not set up all the flags correctly, but was getting this error about "tmp_file". | https://github.com/huggingface/datasets/pull/505 | [
"Thanks for reporting the issue ! I'm creating a new PR to fix it and add tests.\r\n(I'm doing a new PR because I know there's some other place where it needs to be fixed)",
"I'm closing this one as I created the other PR."
] | {
"url": "https://api.github.com/repos/huggingface/datasets/pulls/505",
"html_url": "https://github.com/huggingface/datasets/pull/505",
"diff_url": "https://github.com/huggingface/datasets/pull/505.diff",
"patch_url": "https://github.com/huggingface/datasets/pull/505.patch",
"merged_at": null
} | 505 | true |
Added downloading to Hyperpartisan news detection | Following the discussion on Slack and #349, I've updated the hyperpartisan dataset to pull directly from Zenodo rather than manual install, which should make this dataset much more accessible. Many thanks to @johanneskiesel !
Currently doesn't pass `test_load_real_dataset` - I'm using `self.config.name` which is `default` in this test. Might be related to #474 | https://github.com/huggingface/datasets/pull/504 | [
"Thank you @ghomasHudson for making our dataset available! This is great!",
"The test passes since #527 :)"
] | {
"url": "https://api.github.com/repos/huggingface/datasets/pulls/504",
"html_url": "https://github.com/huggingface/datasets/pull/504",
"diff_url": "https://github.com/huggingface/datasets/pull/504.diff",
"patch_url": "https://github.com/huggingface/datasets/pull/504.patch",
"merged_at": "2020-08-27T08:18:41"
} | 504 | true |
CompGuessWhat?! 0.2.0 | We updated some metadata information associated with the dataset. In addition, we've updated the `create_dummy_data.py` script to generate data samples for the dataset. | https://github.com/huggingface/datasets/pull/503 | [
"I don't see any significant change in the dataset script (except the version value update), can you check that again please ?",
"Hi @aleSuglia , can you check that all the changes you wanted to do are in the dataset script ?",
"Hey sorry but I'm in the middle of a conference deadline. I'll let you know asap!",... | {
"url": "https://api.github.com/repos/huggingface/datasets/pulls/503",
"html_url": "https://github.com/huggingface/datasets/pull/503",
"diff_url": "https://github.com/huggingface/datasets/pull/503.diff",
"patch_url": "https://github.com/huggingface/datasets/pull/503.patch",
"merged_at": null
} | 503 | true |
Fix tokenizers caching | I've found some cases where the caching didn't work properly for tokenizers:
1. if a tokenizer has a regex pattern, then the caching would be inconsistent across sessions
2. if a tokenizer has a cache attribute that changes after some calls, the the caching would not work after cache updates
3. if a tokenizer is used inside a function, the caching of this function would result in the same cache file for different tokenizers
4. if `unique_no_split_tokens`'s attribute is not the same across sessions (after loading a tokenizer) then the caching could be inconsistent
To fix that, this is what I did:
1. register a specific `save_regex` function for pickle that makes regex dumps deterministic
2. ignore cache attribute of some tokenizers before dumping
3. enable recursive dump by default for all dumps
4. make `unique_no_split_tokens` deterministic in https://github.com/huggingface/transformers/pull/6461
I also added tests to make sure that tokenizers hashing works as expected.
In the future we should find a way to test if hashing also works across session (maybe using two CI jobs ? or by hardcoding a tokenizer's hash ?) | https://github.com/huggingface/datasets/pull/502 | [
"This should fix #501 and also the issue you sent me on slack @sgugger ."
] | {
"url": "https://api.github.com/repos/huggingface/datasets/pulls/502",
"html_url": "https://github.com/huggingface/datasets/pull/502",
"diff_url": "https://github.com/huggingface/datasets/pull/502.diff",
"patch_url": "https://github.com/huggingface/datasets/pull/502.patch",
"merged_at": "2020-08-19T13:37:17"
} | 502 | true |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.