Datasets:

---

calvpang

/

github-issues

like 0

[]

[]

[]

[]

[]

[ "Hello, thanks. I had a talk with the dataset authors, found out that the data now is obsolete and they'll get a stable version soon. So temporality closing the PR.\r\n Although I have a question, What should _id_ be in the return statement? Should that be something like a start index (or) the type of split will do? Thanks. ", "@reshinthadithyan we should hold off on this for a couple of weeks till NeurIPS concludes. The [NLC2CMD](http://nlc2cmd.us-east.mybluemix.net/) data will be out then; which includes a cleaner version of this NL2Bash data. The older data is sort of obsolete now. ", "Ah nvm you already commented 😆 " ]

[]

[ { "login": "lhoestq", "id": 42851186, "node_id": "MDQ6VXNlcjQyODUxMTg2", "avatar_url": "https://avatars.githubusercontent.com/u/42851186?v=4", "gravatar_id": "", "url": "https://api.github.com/users/lhoestq", "html_url": "https://github.com/lhoestq", "followers_url": "https://api.github.com/users/lhoestq/followers", "following_url": "https://api.github.com/users/lhoestq/following{/other_user}", "gists_url": "https://api.github.com/users/lhoestq/gists{/gist_id}", "starred_url": "https://api.github.com/users/lhoestq/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/lhoestq/subscriptions", "organizations_url": "https://api.github.com/users/lhoestq/orgs", "repos_url": "https://api.github.com/users/lhoestq/repos", "events_url": "https://api.github.com/users/lhoestq/events{/privacy}", "received_events_url": "https://api.github.com/users/lhoestq/received_events", "type": "User", "site_admin": false } ]

[ "Thanks for reporting ! I'm looking into it." ]

[]

[]

[]

[]

[]

[ "This dataset seems to have around 1000 configs. Therefore when creating the dummy data we end up with hundreds of MB of dummy data which we don't want to add in the repo.\r\nLet's make this PR on hold for now and find a solution after the sprint of next week", "Closing in favor of #1566 " ]

[ { "id": 2107841032, "node_id": "MDU6TGFiZWwyMTA3ODQxMDMy", "url": "https://api.github.com/repos/huggingface/datasets/labels/nlp-viewer", "name": "nlp-viewer", "color": "94203D", "default": false, "description": "" } ]

[]

[ "Thanks for reporting !\r\ncc @srush for the empty feature list issue and the encoding issue\r\ncc @julien-c maybe we can update the url and just have a redirection from the old url to the new one ?", "Ok, I redirected on our side to a new url. ⚠️ @srush: if you update the Streamlit config too to `/datasets/viewer`, let me know because I'll need to change our nginx config at the same time", "9", "‏⠀‏‏‏⠀‏‏‏⠀ ‏⠀ ", "‏⠀‏‏‏⠀‏‏‏⠀ ‏⠀ " ]

[]

[]

[]

[]

[]

[]

[]

[]

[ "Closing since we don't need to update the tags of those three datasets (for each one of them there is only one tag set)" ]

[]

[]

[ "@lhoestq removed prints and added the dataset card. ", "@lhoestq, I want to add other datasets as well. I am not sure if it is possible to do so with the same branch. ", "Hi @zaidalyafeai, really excited to get more Arabic coverage in the lib, thanks for your contribution!\r\n\r\nCouple of last comments:\r\n- this PR seems to modify some files that are unrelated to your dataset. Could you rebase from master? It should take care of that.\r\n- The dataset card is a good start! Can you describe the task in a few words and add more information in the Data Structure part, including listing and describing the fields? Also, if you don't know how to fill out a paragraph, or if you have some information but think more would be beneficial, please leave `[More Information Needed]` instead of `[N/A]`", "> Hi @zaidalyafeai, really excited to get more Arabic coverage in the lib, thanks for your contribution!\r\n> \r\n> Couple of last comments:\r\n> \r\n> * this PR seems to modify some files that are unrelated to your dataset. Could you rebase from master? It should take care of that.\r\n> * The dataset card is a good start! Can you describe the task in a few words and add more information in the Data Structure part, including listing and describing the fields? Also, if you don't know how to fill out a paragraph, or if you have some information but think more would be beneficial, please leave `[More Information Needed]` instead of `[N/A]`\r\n\r\nI have no idea how some other files changed. I tried to rebase and push but this created some errors. I had to run the command \r\n`git push -u --force origin add-metrec-dataset` which might cause some problems. ", "Feel free to create another branch/another PR without all the other changes", "@yjernite can you explain which other files are changed because of the PR ? https://github.com/huggingface/datasets/pull/893/files only shows files related to the dataset. ", "Right ! github is nice with us today :)", "Looks like this one is ready to merge, thanks @zaidalyafeai !", "@lhoestq thanks for the merge. I am not a GitHub geek. I already have another dataset to add. I'm not sure how to add another given my forked repo. Do I follow the same steps with a different checkout name ?", "If you've followed the instructions in here : https://github.com/huggingface/datasets/blob/master/ADD_NEW_DATASET.md#start-by-preparing-your-environment\r\n\r\n(especially point 2. and the command `git remote add upstream ....`)\r\n\r\nThen you can try\r\n```\r\ngit checkout master\r\ngit fetch upstream\r\ngit rebase upstream/master\r\ngit checkout -b add-<my-new-dataset-name>\r\n```" ]

[]

[]

[ "Looks good to me. Do we also support TSV in this helper (explain if it should be text or CSV) and in the dummy-data creator?", "snli is basically based on tsv files (but named as .txt) and it is in the list of datasets of reference.\r\nThe dummy data creator supports tsv", "merging this one.\r\nIf you think of other datasets of reference to add we can still add them later" ]

[]

[]

[]

[]

[]

[ "Thanks for the comments. I addressed them and pushed again.\r\nWhen I run \"make quality\" I get the following error but I don't know how to resolve it or what the problem ist respectively:\r\nwould reformat /Users/joelniklaus/NextCloud/PhDJoelNiklaus/Code/datasets/datasets/ler/ler.py\r\nOh no! 💥 💔 💥\r\n1 file would be reformatted, 257 files would be left unchanged.\r\nmake: *** [quality] Error 1\r\n", "Awesome thanks :)\r\nTo automatically format the python files you can run `make style`", "I did that now. But still getting the following error:\r\nblack --check --line-length 119 --target-version py36 tests src benchmarks datasets metrics\r\nAll done! ✨ 🍰 ✨\r\n258 files would be left unchanged.\r\nisort --check-only tests src benchmarks datasets metrics\r\nflake8 tests src benchmarks datasets metrics\r\ndatasets/ler/ler.py:46:96: W291 trailing whitespace\r\ndatasets/ler/ler.py:47:68: W291 trailing whitespace\r\ndatasets/ler/ler.py:48:102: W291 trailing whitespace\r\ndatasets/ler/ler.py:49:112: W291 trailing whitespace\r\ndatasets/ler/ler.py:50:92: W291 trailing whitespace\r\ndatasets/ler/ler.py:51:116: W291 trailing whitespace\r\ndatasets/ler/ler.py:52:84: W291 trailing whitespace\r\nmake: *** [quality] Error 1\r\n\r\nHowever: When I look at the file I don't see any trailing whitespace", "maybe a bug with flake8 ? could you try to update it ? which version do you have ?", "This is my flake8 version: 3.7.9 (mccabe: 0.6.1, pycodestyle: 2.5.0, pyflakes: 2.1.1) CPython 3.8.5 on Darwin\r\n", "Now I updated to: 3.8.4 (mccabe: 0.6.1, pycodestyle: 2.6.0, pyflakes: 2.2.0) CPython 3.8.5 on Darwin\r\n\r\nAnd now I even get additional errors:\r\nblack --check --line-length 119 --target-version py36 tests src benchmarks datasets metrics\r\nAll done! ✨ 🍰 ✨\r\n258 files would be left unchanged.\r\nisort --check-only tests src benchmarks datasets metrics\r\nflake8 tests src benchmarks datasets metrics\r\ndatasets/polyglot_ner/polyglot_ner.py:123:64: F541 f-string is missing placeholders\r\ndatasets/ler/ler.py:46:96: W291 trailing whitespace\r\ndatasets/ler/ler.py:47:68: W291 trailing whitespace\r\ndatasets/ler/ler.py:48:102: W291 trailing whitespace\r\ndatasets/ler/ler.py:49:112: W291 trailing whitespace\r\ndatasets/ler/ler.py:50:92: W291 trailing whitespace\r\ndatasets/ler/ler.py:51:116: W291 trailing whitespace\r\ndatasets/ler/ler.py:52:84: W291 trailing whitespace\r\ndatasets/math_dataset/math_dataset.py:233:25: E741 ambiguous variable name 'l'\r\nmetrics/coval/coval.py:236:31: F541 f-string is missing placeholders\r\nmake: *** [quality] Error 1\r\n\r\nI do this on macOS Catalina 10.15.7 in case this matters", "Code quality test now passes, thanks :) \r\n\r\nTo fix the other tests failing I think you can just rebase from master.\r\nAlso make sure that the dummy data test passes with\r\n```python\r\nRUN_SLOW=1 pytest tests/test_dataset_common.py::LocalDatasetTest::test_load_dataset_all_configs_ler\r\n```", "I will close this PR because abishek did the same better (https://github.com/huggingface/datasets/pull/944)", "Sorry you had to close your PR ! It looks like this week's sprint doesn't always make it easy to see what's being added/what's already added. \r\nThank you for contributing to the library. You did a great job on adding LER so feel free to add other ones that you would like to see in the library, it will be a pleasure to review" ]

[]

[]

[ "I like the idea ! And the approach is right imo\r\n\r\nNote that by changing this we will have to add a way for users to get the config lists of a dataset. In the current user workflow, the user could see the list of the config when the missing config error is raised but now it won't be the case because of the default config.", "Maybe let's add a test in the test_builder.py test script ?", "@lhoestq Okay great, I added a test as well as two new inspect functions: `get_dataset_config_names` and `get_dataset_infos` (the latter is something I've been wanting anyway). As a quick hack, you can also just pass a random config name (e.g. an empty string) to `load_dataset` to get the config names in the error msg as before. Also added a couple paragraphs to the adding new datasets doc.\r\n\r\nI'll send a separate PR incorporating this in existing datasets so we can get this merged before our sprint on Monday.\r\n\r\nAny ideas on the failing tests? I'm having trouble making sense of it. **Edit**: nvm, it was master." ]

[]

[]

[ "Yes following the Tensorflow Datasets convention, objects with type `Sequence of a Dict` are actually stored as a `dictionary of lists`.\r\nSee the [documentation](https://huggingface.co/docs/datasets/features.html?highlight=features) for more details", "Thanks.\r\nThis is a bit (very) confusing, but I guess if its intended, I'll just work with it as if its how my data was originally structured :) \r\n" ]

[ { "id": 1935892857, "node_id": "MDU6TGFiZWwxOTM1ODkyODU3", "url": "https://api.github.com/repos/huggingface/datasets/labels/bug", "name": "bug", "color": "d73a4a", "default": true, "description": "Something isn't working" } ]

[]

[ "Yes right now `ArrayXD` can only be used as a column feature type, not a subtype.\r\nWith the current Arrow limitations I don't think we'll be able to make it work as a subtype, however it should be possible to allow dimensions of dynamic sizes (`Array3D(shape=(None, 137, 2), dtype=\"float32\")` for example since the [underlying arrow type](https://github.com/huggingface/datasets/blob/master/src/datasets/features.py#L236) allows dynamic sizes.\r\n\r\nFor now I'd suggest the use of nested `Sequence` types. Once we have the dynamic sizes you can update the dataset.\r\nWhat do you think ?", "> Yes right now ArrayXD can only be used as a column feature type, not a subtype. \r\n\r\nMeaning it can't be nested under `Sequence`?\r\nIf so, for now I'll just make it a python list and make it with the nested `Sequence` type you suggested.", "Yea unfortunately..\r\nThat's a current limitation with Arrow ExtensionTypes that can't be used in the default Arrow Array objects.\r\nWe already have an ExtensionArray that allows us to use them as column types but not for subtypes.\r\nMaybe we can extend it, I haven't experimented with that yet", "Cool\r\nSo please consider this issue as a feature request for:\r\n```\r\nArray3D(shape=(None, 137, 2), dtype=\"float32\")\r\n```\r\n\r\nits a way to represent videos, poses, and other cool sequences", "@lhoestq well, so sequence of sequences doesn't work either...\r\n\r\n```\r\npyarrow.lib.ArrowCapacityError: List array cannot contain more than 2147483646 child elements, have 2147483648\r\n```\r\n\r\n\r\n", "Working with Arrow can be quite fun sometimes.\r\nYou can fix this issue by trying to reduce the writer batch size (same trick than the one used to reduce the RAM usage in https://github.com/huggingface/datasets/issues/741).\r\n\r\nLet me know if it works.\r\nI haven't investigated yet on https://github.com/huggingface/datasets/issues/741 since I was preparing this week's sprint to add datasets but this is in my priority list for early next week.", "The batch size fix doesn't work... not for #741 and not for this dataset I'm trying (DGS corpus)\r\nLoading the DGS corpus takes 400GB of RAM, which is fine with me as my machine is large enough\r\n", "Sorry it doesn't work. Will let you know once I fixed it", "Hi @lhoestq , any update on dynamic sized arrays?\r\n(`Array3D(shape=(None, 137, 2), dtype=\"float32\")`)", "Not yet, I've been pretty busy with the dataset sprint lately but this is something that's been asked several times already. So I'll definitely work on this as soon as I'm done with the sprint and with the RAM issue you reported.", "Hi @lhoestq,\r\nAny chance you have some updates on the supporting `ArrayXD` as a subtype or support of dynamic sized arrays?\r\n\r\ne.g.:\r\n`datasets.features.Sequence(datasets.features.Array2D(shape=(137, 2), dtype=\"float32\"))`\r\n`Array3D(shape=(None, 137, 2), dtype=\"float32\")`", "Hi ! We haven't worked in this lately and it's not in our very short-term roadmap since it requires a bit a work to make it work with arrow. Though this will definitely be added at one point.", "@lhoestq, thanks for the update.\r\n\r\nI actually tried to modify some piece of code to make it work. Can you please tell if I missing anything here?\r\nI think that for vast majority of cases it's enough to make first dimension of the array dynamic i.e. `shape=(None, 100, 100)`. For that, it's enough to modify class [ArrayExtensionArray](https://github.com/huggingface/datasets/blob/9ca24250ea44e7611c4dabd01ecf9415a7f0be6c/src/datasets/features.py#L397) to output list of arrays of different sizes instead of list of arrays of same sizes (current version)\r\nBelow are my modifications of this class.\r\n\r\n```\r\nclass ArrayExtensionArray(pa.ExtensionArray):\r\n def __array__(self):\r\n zero_copy_only = _is_zero_copy_only(self.storage.type)\r\n return self.to_numpy(zero_copy_only=zero_copy_only)\r\n\r\n def __getitem__(self, i):\r\n return self.storage[i]\r\n\r\n def to_numpy(self, zero_copy_only=True):\r\n storage: pa.ListArray = self.storage\r\n size = 1\r\n for i in range(self.type.ndims):\r\n size *= self.type.shape[i]\r\n storage = storage.flatten()\r\n numpy_arr = storage.to_numpy(zero_copy_only=zero_copy_only)\r\n numpy_arr = numpy_arr.reshape(len(self), *self.type.shape)\r\n return numpy_arr\r\n\r\n def to_list_of_numpy(self, zero_copy_only=True):\r\n storage: pa.ListArray = self.storage\r\n shape = self.type.shape\r\n arrays = []\r\n for dim in range(1, self.type.ndims):\r\n assert shape[dim] is not None, f\"Support only dynamic size on first dimension. Got: {shape}\"\r\n\r\n first_dim_offsets = np.array([off.as_py() for off in storage.offsets])\r\n for i in range(len(storage)):\r\n storage_el = storage[i:i+1]\r\n first_dim = first_dim_offsets[i+1] - first_dim_offsets[i]\r\n # flatten storage\r\n for dim in range(self.type.ndims):\r\n storage_el = storage_el.flatten()\r\n\r\n numpy_arr = storage_el.to_numpy(zero_copy_only=zero_copy_only)\r\n arrays.append(numpy_arr.reshape(first_dim, *shape[1:]))\r\n\r\n return arrays\r\n\r\n def to_pylist(self):\r\n zero_copy_only = _is_zero_copy_only(self.storage.type)\r\n if self.type.shape[0] is None:\r\n return self.to_list_of_numpy(zero_copy_only=zero_copy_only)\r\n else:\r\n return self.to_numpy(zero_copy_only=zero_copy_only).tolist()\r\n```\r\n\r\nI ran few tests and it works as expected. Let me know what you think.", "Thanks for diving into this !\r\n\r\nIndeed focusing on making the first dimensions dynamic make total sense (and users could still re-order their dimensions to match this constraint).\r\nYour code looks great :) I think it can even be extended to support several dynamic dimensions if we want to.\r\n\r\nFeel free to open a PR to include these changes, then we can update our test suite to make sure it works in all use cases.\r\nIn particular I think we might need a few tweaks to allow it to be converted to pandas (though I haven't tested yet):\r\n\r\n```python\r\nfrom datasets import Dataset, Features, Array3D\r\n\r\n# this works\r\nmatrix = [[1, 0], [0, 1]]\r\nfeatures = Features({\"a\": Array3D(dtype=\"int32\", shape=(1, 2, 2))})\r\nd = Dataset.from_dict({\"a\": [[matrix], [matrix]]})\r\nprint(d.to_pandas())\r\n\r\n# this should work as well\r\nmatrix = [[1, 0], [0, 1]]\r\nfeatures = Features({\"a\": Array3D(dtype=\"int32\", shape=(None, 2, 2))})\r\nd = Dataset.from_dict({\"a\": [[matrix], [matrix] * 2]})\r\nprint(d.to_pandas())\r\n```\r\n\r\nI'll be happy to help you on this :)" ]

[]

[]

[ "I think this issue is still not resolve yet. Please check my comment in the following issue, thanks.\r\n[#577](https://github.com/huggingface/datasets/issues/577#issuecomment-868122769)" ]

[ { "id": 2067376369, "node_id": "MDU6TGFiZWwyMDY3Mzc2MzY5", "url": "https://api.github.com/repos/huggingface/datasets/labels/dataset%20request", "name": "dataset request", "color": "e99695", "default": false, "description": "Requesting to add a new dataset" } ]

[]

[ "Good point!", "Yes indeed. We can probably improve that by using lazy imports", "#1690 added fast start-up of the library " ]

[]

[]

[ "I took your comments into account.\r\nAlso now after compressing the dummy_data.zip file it runs a dummy data test (=make sure each split has at least 1 example using the dummy data)", "I just tested the tool with some datasets and found out that it's not working for datasets that download files using `download_and_extract(file_url)` (where file_url is a `str`). That's because in that case the dummy_data.zip is not a folder but a single zipped file.\r\n\r\nI think we have to fix that or we can have unexpected behavior when a scripts calls `download_and_extract(file_url)` several times, since it would always point to the same dummy data file.\r\n\r\nSo I decided to change that to have a folder containing the dummy files instead but it breaks around 90 tests so I need to update 90 dummy data files to follow this scheme. I'll probably fix them tomorrow morning.\r\n\r\nWhat do you guys think ? Also cc @patrickvonplaten to make sure I understand things correctly", "Ok I changed to use the dummy_data.zip content to be a folder even for single url calls to `dl_manager.download_and_extract`. Therefore the automatic dummy data generation tool works for most datasets now.\r\n\r\nTo avoid having to change all the old dummy_data.zip files I added backward compatiblity. \r\n\r\nThe only test failing is `tests/test_dataset_common.py::RemoteDatasetTest::test_load_dataset_xcopa`\r\nIt is expected to fail since I had modify its dummy data structure that was wrong. It was causing issue with backward compatibility. It will be fixed as soon as this PR is merged" ]

[ { "id": 1935892871, "node_id": "MDU6TGFiZWwxOTM1ODkyODcx", "url": "https://api.github.com/repos/huggingface/datasets/labels/enhancement", "name": "enhancement", "color": "a2eeef", "default": true, "description": "New feature or request" }, { "id": 1935892912, "node_id": "MDU6TGFiZWwxOTM1ODkyOTEy", "url": "https://api.github.com/repos/huggingface/datasets/labels/question", "name": "question", "color": "d876e3", "default": true, "description": "Further information is requested" } ]

[]

[ "Not at the moment but we could likely support this feature.", "?", "I think it would be a very helpful feature, because sometimes one only wants to evaluate models on the dev set, and the whole training data may be many times bigger.\r\nThis makes the task impossible with limited memory resources." ]

[]

[]

[]

[]

[]

[]

[ { "id": 2067376369, "node_id": "MDU6TGFiZWwyMDY3Mzc2MzY5", "url": "https://api.github.com/repos/huggingface/datasets/labels/dataset%20request", "name": "dataset request", "color": "e99695", "default": false, "description": "Requesting to add a new dataset" } ]

[]

[ "I’ll take this one to test the workflow for the sprint next week cc @yjernite @lhoestq ", "@thomwolf here's a slightly adapted version of the code from the [official Tapas repository](https://github.com/google-research/tapas/blob/master/tapas/utils/interaction_utils.py) that is used to turn the `answer_coordinates` and `answer_texts` columns into true Python lists of tuples/strings:\r\n\r\n```\r\nimport pandas as pd\r\nimport ast\r\n\r\ndata = pd.read_csv(\"/content/sqa_data/random-split-1-dev.tsv\", sep='\\t')\r\n\r\ndef _parse_answer_coordinates(answer_coordinate_str):\r\n \"\"\"Parses the answer_coordinates of a question.\r\n Args:\r\n answer_coordinate_str: A string representation of a Python list of tuple\r\n strings.\r\n For example: \"['(1, 4)','(1, 3)', ...]\"\r\n \"\"\"\r\n\r\n try:\r\n answer_coordinates = []\r\n # make a list of strings\r\n coords = ast.literal_eval(answer_coordinate_str)\r\n # parse each string as a tuple\r\n for row_index, column_index in sorted(\r\n ast.literal_eval(coord) for coord in coords):\r\n answer_coordinates.append((row_index, column_index))\r\n except SyntaxError:\r\n raise ValueError('Unable to evaluate %s' % answer_coordinate_str)\r\n \r\n return answer_coordinates\r\n\r\n\r\ndef _parse_answer_text(answer_text):\r\n \"\"\"Populates the answer_texts field of `answer` by parsing `answer_text`.\r\n Args:\r\n answer_text: A string representation of a Python list of strings.\r\n For example: \"[u'test', u'hello', ...]\"\r\n \"\"\"\r\n try:\r\n answer = []\r\n for value in ast.literal_eval(answer_text):\r\n answer.append(value)\r\n except SyntaxError:\r\n raise ValueError('Unable to evaluate %s' % answer_text)\r\n\r\n return answer\r\n\r\ndata['answer_coordinates'] = data['answer_coordinates'].apply(lambda coords_str: _parse_answer_coordinates(coords_str))\r\ndata['answer_text'] = data['answer_text'].apply(lambda txt: _parse_answer_text(txt))\r\n```\r\n\r\nHere I'm using Pandas to read in one of the TSV files (the dev set). \r\n\r\n", "Closing since SQA was added in #1566 " ]

[ { "id": 2067388877, "node_id": "MDU6TGFiZWwyMDY3Mzg4ODc3", "url": "https://api.github.com/repos/huggingface/datasets/labels/dataset%20bug", "name": "dataset bug", "color": "2edb81", "default": false, "description": "A bug in a dataset script provided in the library" } ]

[]

[ "Hi ! It runs on my side without issues. I tried\r\n```python\r\nfrom datasets import load_dataset\r\nload_dataset(\"boolq\")\r\n```\r\n\r\nWhat version of datasets and tensorflow are your runnning ?\r\nAlso if you manage to get a minimal reproducible script (on google colab for example) that would be useful.", "hey\ni do the exact same commands. for me it fails i guess might be issues with\ncaching maybe?\nthanks\nbest\nrabeeh\n\nOn Tue, Nov 24, 2020, 10:24 AM Quentin Lhoest <[email protected]>\nwrote:\n\n> Hi ! It runs on my side without issues. I tried\n>\n> from datasets import load_datasetload_dataset(\"boolq\")\n>\n> What version of datasets and tensorflow are your runnning ?\n> Also if you manage to get a minimal reproducible script (on google colab\n> for example) that would be useful.\n>\n> —\n> You are receiving this because you authored the thread.\n> Reply to this email directly, view it on GitHub\n> <https://github.com/huggingface/datasets/issues/879#issuecomment-732769114>,\n> or unsubscribe\n> <https://github.com/notifications/unsubscribe-auth/ABP4ZCGGDR2FUMRKZTIY5CTSRN3VXANCNFSM4T7R3U6A>\n> .\n>\n", "Could you check if it works on the master branch ?\r\nYou can use `load_dataset(\"boolq\", script_version=\"master\")` to do so.\r\nWe did some changes recently in boolq to remove the TF dependency and we changed the way the data files are downloaded in https://github.com/huggingface/datasets/pull/881" ]

[ { "id": 1935892871, "node_id": "MDU6TGFiZWwxOTM1ODkyODcx", "url": "https://api.github.com/repos/huggingface/datasets/labels/enhancement", "name": "enhancement", "color": "a2eeef", "default": true, "description": "New feature or request" }, { "id": 1935892912, "node_id": "MDU6TGFiZWwxOTM1ODkyOTEy", "url": "https://api.github.com/repos/huggingface/datasets/labels/question", "name": "question", "color": "d876e3", "default": true, "description": "Further information is requested" } ]

[]

[ "This would be a neat feature", "> neat feature\r\n\r\nI dint get these clearly, can you please elaborate like how to work on these ", "It could maybe work almost out of the box just by using `cached_path` in the text/csv/json scripts, no?", "Thanks thomwolf and julien-c\r\n\r\nI'm still confusion on what you guys said, \r\n\r\nI have solved the problem as follows:\r\n\r\n1. read the csv file using pandas from s3 \r\n2. Convert to dictionary key as column name and values as list column data\r\n3. convert it to Dataset using \r\n`from datasets import Dataset`\r\n`train_dataset = Dataset.from_dict(train_dict)`", "We were brainstorming around your use-case.\r\n\r\nLet's keep the issue open for now, I think this is an interesting question to think about.", "> We were brainstorming around your use-case.\r\n> \r\n> Let's keep the issue open for now, I think this is an interesting question to think about.\r\n\r\nSure thomwolf, Thanks for your concern ", "I agree it would be cool to have that feature. Also that's good to know that pandas supports this.\r\nFor the moment I'd suggest to first download the files locally as thom suggested and then load the dataset by providing paths to the local files", "Don't get\n", "Any updates on this issue?\r\nI face a similar issue. I have many parquet files in S3 and I would like to train on them. \r\nTo be honest I even face issues with only getting the last layer embedding out of them.", "Hi dorlavie, \r\nYou can find one solution that i have mentioned above, that can help you. \r\nAnd there is one more solution also which is downloading files locally\r\n", "> Hi dorlavie,\r\n> You can find one solution that i have mentioned above, that can help you.\r\n> And there is one more solution also which is downloading files locally\r\n\r\nmahesh1amour, thanks for the fast reply\r\n\r\nUnfortunately, in my case I can not read with pandas. The dataset is too big (50GB). \r\nIn addition, due to security concerns I am not allowed to save the data locally", "@dorlavie could use `boto3` to download the data to your local machine and then load it with `dataset`\r\n\r\nboto3 example [documentation](https://boto3.amazonaws.com/v1/documentation/api/latest/guide/s3-example-download-file.html)\r\n```python\r\nimport boto3\r\n\r\ns3 = boto3.client('s3')\r\ns3.download_file('BUCKET_NAME', 'OBJECT_NAME', 'FILE_NAME')\r\n```\r\n\r\ndatasets example [documentation](https://huggingface.co/docs/datasets/loading_datasets.html)\r\n\r\n```python\r\nfrom datasets import load_dataset\r\ndataset = load_dataset('csv', data_files=['my_file_1.csv', 'my_file_2.csv', 'my_file_3.csv'])\r\n```\r\n", "Thanks @philschmid for the suggestion.\r\nAs I mentioned in the previous comment, due to security issues I can not save the data locally.\r\nI need to read it from S3 and process it directly.\r\n\r\nI guess that many other people try to train / fit those models on huge datasets (e.g entire Wiki), what is the best practice in those cases?", "If I understand correctly you're not allowed to write data on disk that you downloaded from S3 for example ?\r\nOr is it the use of the `boto3` library that is not allowed in your case ?", "@lhoestq yes you are correct.\r\nI am not allowed to save the \"raw text\" locally - The \"raw text\" must be saved only on S3.\r\nI am allowed to save the output of any model locally. \r\nIt doesn't matter how I do it boto3/pandas/pyarrow, it is forbidden", "@dorlavie are you using sagemaker for training too? Then you could use S3 URI, for example `s3://my-bucket/my-training-data` and pass it within the `.fit()` function when you start the sagemaker training job. Sagemaker would then download the data from s3 into the training runtime and you could load it from disk\r\n\r\n**sagemaker start training job**\r\n```python\r\npytorch_estimator.fit({'train':'s3://my-bucket/my-training-data','eval':'s3://my-bucket/my-evaluation-data'})\r\n```\r\n\r\n**in the train.py script**\r\n```python\r\nfrom datasets import load_from_disk\r\n\r\ntrain_dataset = load_from_disk(os.environ['SM_CHANNEL_TRAIN'])\r\n```\r\n\r\nI have created an example of how to use transformers and datasets with sagemaker. \r\nhttps://github.com/philschmid/huggingface-sagemaker-example/tree/main/03_huggingface_sagemaker_trainer_with_data_from_s3\r\n\r\nThe example contains a jupyter notebook `sagemaker-example.ipynb` and an `src/` folder. The sagemaker-example is a jupyter notebook that is used to create the training job on AWS Sagemaker. The `src/` folder contains the `train.py`, our training script, and `requirements.txt` for additional dependencies.\r\n\r\n" ]

[]

[]

[ "Hi ! Thanks for reporting.\r\nDo you have the same slowdown when you iterate through the raw dataset object as well ? (no dataloader)\r\nIt would be nice to know whether it comes from the dataloader or not", "> Hi ! Thanks for reporting.\r\n> Do you have the same slowdown when you iterate through the raw dataset object as well ? (no dataloader)\r\n> It would be nice to know whether it comes from the dataloader or not\r\n\r\nI did not iter data from raw dataset, maybe I will test later. Now I iter all files directly from `open(file)`, around 20000it/s." ]

[]

[]

[ "It looks like there was an issue while building the imdb dataset.\r\nCould you provide more information about your OS and the version of python and `datasets` ?\r\n\r\nAlso could you try again with \r\n```python\r\ndataset = datasets.load_dataset(\"imdb\", split=\"train\", download_mode=\"force_redownload\")\r\n```\r\nto make sure it's not a corrupted file issue ?", "I was using version 1.1.2 and this resolved with version 1.1.3, thanks. ", "Hello,\r\nI have the same pb with 1.8.0", "Hi ! I just tried in 1.8.0 and it worked fine. Can you try again ? Maybe the dataset host had some issues that are fixed now", "Hello,\r\nIt works fine now :) !\r\nThanks !" ]

[]

[]

[ "I just opened a PR to fix this.\r\nThanks for reporting !" ]

[]

[]

[ "This was fixed in #740 \r\nCould you try to update `datasets` and try again ?", "This has been fixed in datasets 1.1.3" ]

[]

[]

[ "I get the same error. It was fixed some days ago, but again it appears", "Hi @mrm8488 it's working again today without any fix so I am closing this issue.", "I see the issue happening again today - \r\n\r\n[nltk_data] Downloading package stopwords to /root/nltk_data...\r\n[nltk_data] Package stopwords is already up-to-date!\r\nDownloading and preparing dataset cnn_dailymail/3.0.0 (download: 558.32 MiB, generated: 1.28 GiB, post-processed: Unknown size, total: 1.82 GiB) to /root/.cache/huggingface/datasets/cnn_dailymail/3.0.0/3.0.0/0128610a44e10f25b4af6689441c72af86205282d26399642f7db38fa7535602...\r\n\r\n---------------------------------------------------------------------------\r\n\r\nNotADirectoryError Traceback (most recent call last)\r\n\r\n<ipython-input-9-cd4bf8bea840> in <module>()\r\n 22 \r\n 23 \r\n---> 24 train = load_dataset('cnn_dailymail', '3.0.0', split='train')\r\n 25 validation = load_dataset('cnn_dailymail', '3.0.0', split='validation')\r\n 26 test = load_dataset('cnn_dailymail', '3.0.0', split='test')\r\n\r\n5 frames\r\n\r\n/root/.cache/huggingface/modules/datasets_modules/datasets/cnn_dailymail/0128610a44e10f25b4af6689441c72af86205282d26399642f7db38fa7535602/cnn_dailymail.py in _find_files(dl_paths, publisher, url_dict)\r\n 132 else:\r\n 133 logging.fatal(\"Unsupported publisher: %s\", publisher)\r\n--> 134 files = sorted(os.listdir(top_dir))\r\n 135 \r\n 136 ret_files = []\r\n\r\nNotADirectoryError: [Errno 20] Not a directory: '/root/.cache/huggingface/datasets/downloads/1bc05d24fa6dda2468e83a73cf6dc207226e01e3c48a507ea716dc0421da583b/cnn/stories'\r\n\r\nCan someone please take a look ?", "Sometimes happens. Try in a while", "It is working now, thank you. " ]

[]

[]

[ "thanks ! merging now" ]

[]

[]

[ "Loading the iwslt2017-en-nl config of iwslt2017 works fine on my side. \r\nMaybe you can open an issue on transformers as well ? And also add more details about your environment (OS, python version, version of transformers and datasets etc.)", "closing now, figured out this is because the max length of decoder was set smaller than the input_dimensions. thanks " ]

[ { "id": 1935892871, "node_id": "MDU6TGFiZWwxOTM1ODkyODcx", "url": "https://api.github.com/repos/huggingface/datasets/labels/enhancement", "name": "enhancement", "color": "a2eeef", "default": true, "description": "New feature or request" } ]

[]

[ "Hi ! Thanks for your message.\r\nIndeed it's a free feature we can add and that can be useful.\r\nIf you want to contribute, feel free to open a PR to add it to the text dataset script :)" ]

[]

[]

[ ":+1: Thanks for fixing it!" ]

[]

[]

[ "I keep this PR in stand-by for next week's datasets sprint. If the next release is 2.0.0 then we can include it given that it's breaking for many metrics" ]

[]

[]

[]

[ { "id": 2067376369, "node_id": "MDU6TGFiZWwyMDY3Mzc2MzY5", "url": "https://api.github.com/repos/huggingface/datasets/labels/dataset%20request", "name": "dataset request", "color": "e99695", "default": false, "description": "Requesting to add a new dataset" } ]

[]

[ "PR is already open here : #348 \r\nThe only thing remaining is to compute the metadata of each subdataset (one per language + shuffled/unshuffled).\r\nAs soon as #863 is merged we can start computing them. This will take a bit of time though", "Grand, thanks for this!" ]

[]

[]

[ "I'm sorry, this was a problem with my environment.\r\nNow that I have identified the cause of environmental dependency, I would like to fix it and try it.\r\nExcuse me for making a noise." ]

[ { "id": 2067388877, "node_id": "MDU6TGFiZWwyMDY3Mzg4ODc3", "url": "https://api.github.com/repos/huggingface/datasets/labels/dataset%20bug", "name": "dataset bug", "color": "2edb81", "default": false, "description": "A bug in a dataset script provided in the library" } ]

[ { "login": "lhoestq", "id": 42851186, "node_id": "MDQ6VXNlcjQyODUxMTg2", "avatar_url": "https://avatars.githubusercontent.com/u/42851186?v=4", "gravatar_id": "", "url": "https://api.github.com/users/lhoestq", "html_url": "https://github.com/lhoestq", "followers_url": "https://api.github.com/users/lhoestq/followers", "following_url": "https://api.github.com/users/lhoestq/following{/other_user}", "gists_url": "https://api.github.com/users/lhoestq/gists{/gist_id}", "starred_url": "https://api.github.com/users/lhoestq/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/lhoestq/subscriptions", "organizations_url": "https://api.github.com/users/lhoestq/orgs", "repos_url": "https://api.github.com/users/lhoestq/repos", "events_url": "https://api.github.com/users/lhoestq/events{/privacy}", "received_events_url": "https://api.github.com/users/lhoestq/received_events", "type": "User", "site_admin": false } ]

[ "Same error here!\r\n", "Same here! My kaggle notebook stopped working like yesterday. It's strange because I have fixed version of datasets==1.1.2", "I'm looking at it right now", "I couldn't reproduce unfortunately. I tried\r\n```python\r\nfrom datasets import load_dataset\r\n\r\nload_dataset(\"cnn_dailymail\", \"3.0.0\", download_mode=\"force_redownload\")\r\n```\r\nand it worked fine on both my env (python 3.7.2) and colab (python 3.6.9)\r\n\r\nMaybe there was an issue with the google drive download link of the dataset ?\r\nAre you still having the issue ? If so could your give me more info about your python and requests version ?", "No, It's working fine now. Very strange. Here are my python and request versions\r\n\r\nrequests 2.24.0\r\nPython 3.8.2", "It's working as expected. Closing the issue \r\n\r\nThanks everybody." ]

[]

[]

[]

[]

[]

[]

[ { "id": 1935892871, "node_id": "MDU6TGFiZWwxOTM1ODkyODcx", "url": "https://api.github.com/repos/huggingface/datasets/labels/enhancement", "name": "enhancement", "color": "a2eeef", "default": true, "description": "New feature or request" }, { "id": 1935892912, "node_id": "MDU6TGFiZWwxOTM1ODkyOTEy", "url": "https://api.github.com/repos/huggingface/datasets/labels/question", "name": "question", "color": "d876e3", "default": true, "description": "Further information is requested" } ]

[ { "login": "albertvillanova", "id": 8515462, "node_id": "MDQ6VXNlcjg1MTU0NjI=", "avatar_url": "https://avatars.githubusercontent.com/u/8515462?v=4", "gravatar_id": "", "url": "https://api.github.com/users/albertvillanova", "html_url": "https://github.com/albertvillanova", "followers_url": "https://api.github.com/users/albertvillanova/followers", "following_url": "https://api.github.com/users/albertvillanova/following{/other_user}", "gists_url": "https://api.github.com/users/albertvillanova/gists{/gist_id}", "starred_url": "https://api.github.com/users/albertvillanova/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/albertvillanova/subscriptions", "organizations_url": "https://api.github.com/users/albertvillanova/orgs", "repos_url": "https://api.github.com/users/albertvillanova/repos", "events_url": "https://api.github.com/users/albertvillanova/events{/privacy}", "received_events_url": "https://api.github.com/users/albertvillanova/received_events", "type": "User", "site_admin": false } ]

[ "The preprocessing tokenizes the input text. Tokenization outputs `input_ids`, `attention_mask`, `token_type_ids` and `special_tokens_mask`. All those are of length`max_seq_length` because of padding. Therefore for each sample it generate 4 *`max_seq_length` integers. Currently they're all saved as int64. This is why the tokenization takes so much space.\r\n\r\nI'm sure we can optimize that though\r\nWhat do you think @sgugger ?", "First I think we should disable padding in the dataset processing and let the data collator do it.\r\n\r\nThen I'm wondering if you need attention_mask and token_type_ids at this point ?\r\n\r\nFinally we can also specify the output feature types at this line https://github.com/huggingface/transformers/blob/master/examples/language-modeling/run_mlm.py#L280 to use more optimized integer precisions for the output. Maybe something like:\r\n- input_ids: uint16 or uint32\r\n- token_type_ids: uint8 or bool\r\n- attention_mask: bool\r\n- special_tokens_mask: bool\r\n\r\nAlso IMO these changes are all on the `transformers` side. Maybe we should discuss on the `transformers` repo", "> First I think we should disable padding in the dataset processing and let the data collator do it.\r\n\r\nNo, you can't do that on TPUs as dynamic shapes will result in a very slow training. The script can however be tweaked to use the `PaddingDataCollator` with a fixed max length instead of dynamic batching.\r\n\r\nFor the other optimizations, they can be done by changing the script directly for each user's use case. Not sure we can find something that is general enough to be in transformers or the examples script.", "Oh yes right..\r\nDo you think that a lazy map feature on the `datasets` side could help to avoid storing padded tokenized texts then ?", "I think I can do the tweak mentioned above with the data collator as short fix (but fully focused on v4 right now so that will be for later this week, beginning of next week :-) ).\r\nIf it doesn't hurt performance to tokenize on the fly, that would clearly be the long-term solution however!", "> Hey guys,\r\n> \r\n> I was trying to create a new bert model from scratch via _huggingface transformers + tokenizers + dataets_ (actually using this example script by your team: https://github.com/huggingface/transformers/blob/master/examples/language-modeling/run_mlm.py). It was supposed to be a first test with a small 5 GB raw text file but I can't even end the preprocessing handled by datasets because this tiny 5 GB text file becomes more than 1 TB when processing. My system was running out of space and crashed prematurely.\r\n> \r\n> I've done training from scratch via Google's bert repo in the past and I can remember that the resulting pretraining data can become quite big. But 5 GB becoming 1 TB was never the case. Is this considered normal or is it a bug?\r\n> \r\n> I've used the following CMD:\r\n> `python xla_spawn.py --num_cores=8 run_mlm.py --model_type bert --config_name config.json --tokenizer_name tokenizer.json --train_file dataset_full.txt --do_train --output_dir out --max_steps 500000 --save_steps 2500 --save_total_limit 2 --prediction_loss_only --line_by_line --max_seq_length 128 --pad_to_max_length --preprocessing_num_workers 16 --per_device_train_batch_size 128 --overwrite_output_dir --debug`\r\n\r\nIt's actually because of the parameter 'preprocessing_num_worker' when using TPU. \r\nI am also planning to have my model trained on the google TPU with a 11gb text corpus. With x8 cores enabled, each TPU core has its own dataset. When not using distributed training, the preprocessed file is about 77gb. On the opposite, if enable xla, the file produced will easily consume all my free space(more than 220gb, I think it will be, in the end, around 600gb ). \r\nSo I think that's maybe where the problem came from. \r\n\r\nIs there any possibility that all of the cores share the same preprocess dataset?\r\n\r\n@sgugger @RammMaschine ", "Hi @NebelAI, we have optimized Datasets' disk usage in the latest release v1.5.\r\n\r\nFeel free to update your Datasets version\r\n```shell\r\npip install -U datasets\r\n```\r\nand see if it better suits your needs." ]

[ { "id": 2067388877, "node_id": "MDU6TGFiZWwyMDY3Mzg4ODc3", "url": "https://api.github.com/repos/huggingface/datasets/labels/dataset%20bug", "name": "dataset bug", "color": "2edb81", "default": false, "description": "A bug in a dataset script provided in the library" } ]

[]

[]

[]

[]

[]

[]

[]

[ "Added dummy data and encoding to open(). Now everything should be fine, hopefully :)" ]

[]

[]

[]

[]

[]

[ "@lhoestq I fixed issues except for the dummy_data zip file. But I think I know why is it happening. So when unzipping dummy_data.zip it gets save in /tmp directory where glob doesn't pick it up. For regular downloads, the archive gets unzipped in ~/.cache/huggingface. Could that be a reason?", "Nice thanks :)\r\n\r\nWhen testing with the dummy data, the `download_manager.download_and_extract()` call returns the path to the unzipped dummy_data.zip archive. Therefore glob should be able to find your dummy .epub.txt file", "@lhoestq I understand but for some reason, it is not happening. I added logs to see where dummy_data.zip gets unzipped in /tmp but I suppose when the test process finishes that tmp is gone. I also tried to glob anything in _generate_examples from that directory using /* instead of **/*.epub.txt and nothing is being returned. Always an empty array. ", "Ok weird ! I can take a look tomorrow if you want", "Please do, I will take a fresh look as well. ", "In _generate_examples_ I wrote the following:\r\n```\r\nglob_target = os.path.join(directory, \"**/*.epub.txt\")\r\nprint(f\"Glob target {glob_target }\")\r\n```\r\n\r\nAnd here is the test failure:\r\n\r\n\r\n========================================================================================== FAILURES ===========================================================================================\r\n________________________________________________________________ LocalDatasetTest.test_load_dataset_all_configs_bookcorpusopen ________________________________________________________________\r\n\r\nself = <tests.test_dataset_common.LocalDatasetTest testMethod=test_load_dataset_all_configs_bookcorpusopen>, dataset_name = 'bookcorpusopen'\r\n\r\n @slow\r\n def test_load_dataset_all_configs(self, dataset_name):\r\n configs = self.dataset_tester.load_all_configs(dataset_name, is_local=True)\r\n> self.dataset_tester.check_load_dataset(dataset_name, configs, is_local=True)\r\n\r\ntests/test_dataset_common.py:232: \r\n_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _\r\ntests/test_dataset_common.py:193: in check_load_dataset\r\n self.parent.assertTrue(len(dataset[split]) > 0)\r\nE AssertionError: False is not true\r\n------------------------------------------------------------------------------------ Captured stdout call -------------------------------------------------------------------------------------\r\nDownloading and preparing dataset book_corpus_open/plain_text (download: 1.00 MiB, generated: 1.00 MiB, post-processed: Unknown size, total: 2.00 MiB) to /var/folders/y_/6k6zhblx0k9dsdz5nd_z9x5c0000gp/T/tmpmuu0_ln2/book_corpus_open/plain_text/1.0.0...\r\nGlob target /var/folders/y_/6k6zhblx0k9dsdz5nd_z9x5c0000gp/T/tmpm6tpvb3f/extracted/d953b414cceb4fe3985eeaf68aec2f4435f166b2edf66863d805e3825b7d336b/dummy_data/**/*.epub.txt\r\nDataset book_corpus_open downloaded and prepared to /var/folders/y_/6k6zhblx0k9dsdz5nd_z9x5c0000gp/T/tmpmuu0_ln2/book_corpus_open/plain_text/1.0.0. Subsequent calls will reuse this data.\r\n------------------------------------------------------------------------------------ Captured stderr call -------------------------------------------------------------------------------------\r\n \r\n", "And when I do os.listdir on the given directory I get:\r\n\r\n glob_target = os.path.join(directory, \"**/*.epub.txt\")\r\n print(f\"Glob target {glob_target }\")\r\n> print(os.listdir(path=directory))\r\nE FileNotFoundError: [Errno 2] No such file or directory: '/var/folders/y_/6k6zhblx0k9dsdz5nd_z9x5c0000gp/T/tmpbu_aom5q/extracted/d953b414cceb4fe3985eeaf68aec2f4435f166b2edf66863d805e3825b7d336b/dummy_data'\r\n", "Thanks for the info, I'm looking at it right now", "Ok found the issue !\r\n\r\nThe dummy_data.zip file must be an archive of a folder named dummy_data. Currently the dummy_data.zip is an archive of a folder named book1. In order to have a valid dummy_data.zip file you must first take the dummy book1 folder, place it inside a folder named dummy_data and then compress the dummy_data folder to get dummy_data.zip", "Excellent, I am on it @lhoestq ", "> Awesome thank you so much for adding it :)\r\n\r\nYou're welcome, ok all tests are green now! I needed it asap as well. Thanks for your help @lhoestq .", "I just wanted to say thank you to everyone involved in making this happen! I was certain that I would have to add bookcorpusnew myself, but then @vblagoje came along and did it, and @lhoestq gave some great support in a timely fashion.\r\n\r\nBy the way @vblagoje, are you on Twitter? I'm https://twitter.com/theshawwn if you'd like to DM and say hello. Once again, thanks for doing this!\r\n\r\nI'll mention over at https://github.com/soskek/bookcorpus/issues/27 that this was merged.", "Thank you Shawn. You did all the heavy lifting ;-)", "@vblagoje Would you be interested in adding books3 as well? https://twitter.com/theshawwn/status/1320282149329784833\r\n\r\nHuggingface is interested and asked me to add it, but I had a bit of trouble during setup (https://github.com/huggingface/datasets/issues/790) and never got around to it. At this point you have much more experience than I do with the datasets lib.\r\n\r\nIt *seems* like it might simply be a matter of copy-pasting this PR, changing books1 to books3, and possibly trimming off the leading paths -- each book is at e.g. the-eye/Books/Bibliotok/J/Jurassic Park.epub.txt, which is rather lengthy compared to just the filename -- but the full path is probably fine, so feel free to do the least amount of work that gets the job done. Otherwise I suppose I'll get around to it eventually; thanks again!", "@shawwn I'll take a look as soon as I clear my work queue. TBH, I would likely work on making sure HF datasets has all the datasets used to train https://github.com/alexa/bort/ and these are: Wikipedia, Wiktionary, OpenWebText (Gokaslan and Cohen, 2019), UrbanDictionary, Onel Billion Words (Chelba et al., 2014), the news subset of Common Crawl (Nagel, 2016)10, and Bookcorpus. cc @lhoestq " ]

[]

[]

[]

[ { "id": 2067388877, "node_id": "MDU6TGFiZWwyMDY3Mzg4ODc3", "url": "https://api.github.com/repos/huggingface/datasets/labels/dataset%20bug", "name": "dataset bug", "color": "2edb81", "default": false, "description": "A bug in a dataset script provided in the library" } ]

[]

[ "Hi,I also posted it to the forum, but this is a bug, perhaps it needs to be reported here? thanks ", "It looks like the official OPUS server for WMT16 doesn't provide the data files anymore (503 error).\r\nI searched a bit and couldn't find a mirror except maybe http://nlp.ffzg.hr/resources/corpora/setimes/ (the data are a cleaned version of the original ones though)\r\nShould we consider replacing the old urls with these ones even though it's not the exact same data ?", "The data storage is down at the moment. Sorry. Hopefully, it will come back soon. Apologies for the inconvenience ...", "Dear great huggingface team, this is not working yet, I really appreciate some temporary fix on this, I need this for my project and this is time sensitive and I will be grateful for your help on this. ", "We have reached out to the OPUS team which is currently working on making the data available again. Cc @jorgtied ", "thank you @thomwolf and HuggingFace team for the help. ", "OPUS is still down - hopefully back tomorrow.", "Hi, this is still down, I would be really grateful if you could ping them one more time. thank you so much. ", "Hi\r\nI am trying with multiple setting of wmt datasets and all failed so far, I need to have at least one dataset working for testing somecodes, and this is really time sensitive, I greatly appreciate letting me know of one translation datasets currently working. thanks ", "It is still down, unfortunately. I'm sorry for that. It should come up again later today or tomorrow at the latest if no additional complications will happen.", "Hi all, \r\nI pulled a request that fix this issue by replacing urls. \r\n\r\nhttps://github.com/huggingface/datasets/pull/1901\r\n\r\nThanks!\r\n", "It's still down for the wmt." ]

[ { "id": 1935892871, "node_id": "MDU6TGFiZWwxOTM1ODkyODcx", "url": "https://api.github.com/repos/huggingface/datasets/labels/enhancement", "name": "enhancement", "color": "a2eeef", "default": true, "description": "New feature or request" }, { "id": 1935892884, "node_id": "MDU6TGFiZWwxOTM1ODkyODg0", "url": "https://api.github.com/repos/huggingface/datasets/labels/help%20wanted", "name": "help wanted", "color": "008672", "default": true, "description": "Extra attention is needed" }, { "id": 1935892912, "node_id": "MDU6TGFiZWwxOTM1ODkyOTEy", "url": "https://api.github.com/repos/huggingface/datasets/labels/question", "name": "question", "color": "d876e3", "default": true, "description": "Further information is requested" } ]

[ { "login": "albertvillanova", "id": 8515462, "node_id": "MDQ6VXNlcjg1MTU0NjI=", "avatar_url": "https://avatars.githubusercontent.com/u/8515462?v=4", "gravatar_id": "", "url": "https://api.github.com/users/albertvillanova", "html_url": "https://github.com/albertvillanova", "followers_url": "https://api.github.com/users/albertvillanova/followers", "following_url": "https://api.github.com/users/albertvillanova/following{/other_user}", "gists_url": "https://api.github.com/users/albertvillanova/gists{/gist_id}", "starred_url": "https://api.github.com/users/albertvillanova/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/albertvillanova/subscriptions", "organizations_url": "https://api.github.com/users/albertvillanova/orgs", "repos_url": "https://api.github.com/users/albertvillanova/repos", "events_url": "https://api.github.com/users/albertvillanova/events{/privacy}", "received_events_url": "https://api.github.com/users/albertvillanova/received_events", "type": "User", "site_admin": false } ]

[ "Unfortunately `concatenate_datasets` only supports concatenating the rows, while what you want to achieve is concatenate the columns.\r\nCurrently to add more columns to a dataset, one must use `map`.\r\nWhat you can do is somehting like this:\r\n```python\r\n# suppose you have datasets d1, d2, d3\r\ndef add_columns(example, index):\r\n example.update(d2[index])\r\n example.update(d3[index])\r\n return example\r\n\r\nfull_dataset = d1.map(add_columns, with_indices=True)\r\n```", "Closing this one, feel free to re-open if you have other questions about this issue", "That's not really difficult to add, though, no?\r\nI think it can be done without copy.\r\nMaybe let's add it to the roadmap?", "Actually it's doable but requires to update the `Dataset._data_files` schema to support this.\r\nI'm re-opening this since we may want to add this in the future", "Hi @lhoestq, I would love to help and add this feature if still needed. My plan is to add an axis variable in the `concatenate_datasets` function in `arrow_dataset.py` and when that is set to 1 concatenate columns instead of rows. ", "Hi ! I would love to see this feature implemented as well :) Thank you for proposing your help !\r\n\r\nHere is a few things about the current implementation:\r\n- A dataset object is a wrapper of one `pyarrow.Table` that contains the data\r\n- Pyarrow offers an API that allows to transform Table objects. For example there are functions like `concat_tables`, `Table.rename_columns`, `Table.add_column` etc.\r\n\r\nTherefore adding columns from another dataset is possible thanks to the pyarrow API and in particular `Table.add_column` :) \r\n\r\nHowever this breaks some features we have regarding pickle. A dataset object can be pickled and unpickled without loading all the data in memory. It is useful for multiprocessing for example. Pickling a dataset object is possible thanks to the `Dataset._data_files` which defines the list of arrow files that will be used to form the final Table (basically all the data from each files are concatenated on axis 0).\r\n\r\nTherefore to be able to add columns to a Dataset and still be able to work with it in a multiprocessing setup, we need to extend this last aspect to be able to reconstruct a Table object from multiple arrow files that are combined in both axis 0 and 1. Currently this reconstruction mechanism only supports axis 0.\r\n\r\nI'm sure we can figure something out that enables users to add columns from another dataset while keeping the multiprocessing support.", "@lhoestq, we have two Pull Requests to implement:\r\n- Dataset.add_item: #1870\r\n- Dataset.add_column: #2145\r\nwhich add a single row or column, repectively.\r\n\r\nThe request here is to implement the concatenation of *multiple* rows/columns. Am I right?\r\n\r\nWe should agree on the API:\r\n- `concatenate_datasets` with `axis`?\r\n- other Dataset method name?", "For the API, I like `concatenate_datasets` with `axis` personally :)\r\nFrom a list of `Dataset` objects, it would concatenate them to a new `Dataset` object backed by a `ConcatenationTable`, that is the concatenation of the tables of each input dataset. The concatenation is either on axis=0 (append rows) or on axis=1 (append columns).\r\n\r\nRegarding what we need to implement:\r\nThe axis=0 is already supported and is the current behavior of `concatenate_datasets`.\r\nAlso `add_item` is not needed to implement axis=1 (though it's an awesome addition to this library).\r\n\r\nTo implement axis=1, we either need `add_column` or a `ConcatenationTable` constructor to concatenate tables horizontally.\r\nI have a preference for using a `ConcatenationTable` constructor because this way we can end up with a `ConcatenationTable` with only 1 additional block per table, while `add_column` would add 1 block per new column.\r\n\r\nMaybe we can simply have an equivalent of `ConcatenationTable.from_tables` but for axis=1 ?\r\n`axis` could also be an argument of `ConcatenationTable.from_tables`", "@lhoestq I think I guessed your suggestions in advance... 😉 #2151", "Cool ! Sorry I missed this one ^^\r\nI'm taking a look ;)" ]

[ { "id": 2067376369, "node_id": "MDU6TGFiZWwyMDY3Mzc2MzY5", "url": "https://api.github.com/repos/huggingface/datasets/labels/dataset%20request", "name": "dataset request", "color": "e99695", "default": false, "description": "Requesting to add a new dataset" } ]

[]

[]

[ { "id": 2067400959, "node_id": "MDU6TGFiZWwyMDY3NDAwOTU5", "url": "https://api.github.com/repos/huggingface/datasets/labels/Metric%20discussion", "name": "Metric discussion", "color": "d722e8", "default": false, "description": "Discussions on the metrics" } ]

[]

[ "@alexyalunin \r\n\r\nI did something similar for others languages.\r\n\r\n[Repo: rouge-metric](https://github.com/m3hrdadfi/rouge-metric)" ]

[]

[]

[ "@lhoestq Better?" ]

[]

[]

[ "Thanks for reporting !\r\nWe plan to show information about the different configs of the datasets on the website, with the corresponding `load_dataset` calls.\r\n\r\nAlso I think the bullet points formatting has been fixed" ]

[]

[]

[ "As you can see in the error the test checks if `indices_mappings_in_memory` is True or not, which is different from the test you do in your script. In a dataset, both the data and the indices mapping can be either on disk or in memory.\r\n\r\nThe indices mapping correspond to a mapping on top of the data table that is used to re-order/select a sample of the original data table. For example if you do `dataset.train_test_split`, then the resulting train and test datasets will have both an indices mapping to tell which examples are in train and which ones in test.\r\n\r\nBefore saving your datasets on disk, you should call `dataset.flatten_indices()` to remove the indices mapping. It should fix your issue. Under the hood it will create a new data table using the indices mapping. The new data table is going to be a subset of the old one (for example taking only the test set examples), and since the indices mapping will be gone you'll be able to concatenate your datasets.\r\n", "> As you can see in the error the test checks if `indices_mappings_in_memory` is True or not, which is different from the test you do in your script. In a dataset, both the data and the indices mapping can be either on disk or in memory.\r\n> \r\n> The indices mapping correspond to a mapping on top of the data table that is used to re-order/select a sample of the original data table. For example if you do `dataset.train_test_split`, then the resulting train and test datasets will have both an indices mapping to tell which examples are in train and which ones in test.\r\n> \r\n> Before saving your datasets on disk, you should call `dataset.flatten_indices()` to remove the indices mapping. It should fix your issue. Under the hood it will create a new data table using the indices mapping. The new data table is going to be a subset of the old one (for example taking only the test set examples), and since the indices mapping will be gone you'll be able to concatenate your datasets.\r\n\r\n`dataset.flatten_indices()` solved my problem, thanks so much!", "@lhoestq we can add a mention of `dataset.flatten_indices()` in the error message (no rush, just put it on your TODO list or I can do it when I come at it)", "Yup I agree ! And in the docs as well" ]

[]

[]

[ "It looks like an issue with wandb/tqdm here.\r\nWe're using the `multiprocess` library instead of the `multiprocessing` builtin python package to support various types of mapping functions. Maybe there's some sort of incompatibility.\r\n\r\nCould you make a minimal script to reproduce or a google colab ?", "hi facing the same issue here - \r\n\r\n`AssertionError: Caught AssertionError in DataLoader worker process 0.\r\nOriginal Traceback (most recent call last):\r\n File \"/usr/lib/python3.6/logging/__init__.py\", line 996, in emit\r\n stream.write(msg)\r\n File \"/usr/local/lib/python3.6/dist-packages/wandb/sdk/lib/redirect.py\", line 100, in new_write\r\n cb(name, data)\r\n File \"/usr/local/lib/python3.6/dist-packages/wandb/sdk/wandb_run.py\", line 723, in _console_callback\r\n self._backend.interface.publish_output(name, data)\r\n File \"/usr/local/lib/python3.6/dist-packages/wandb/sdk/interface/interface.py\", line 153, in publish_output\r\n self._publish_output(o)\r\n File \"/usr/local/lib/python3.6/dist-packages/wandb/sdk/interface/interface.py\", line 158, in _publish_output\r\n self._publish(rec)\r\n File \"/usr/local/lib/python3.6/dist-packages/wandb/sdk/interface/interface.py\", line 456, in _publish\r\n if self._process and not self._process.is_alive():\r\n File \"/usr/lib/python3.6/multiprocessing/process.py\", line 134, in is_alive\r\n assert self._parent_pid == os.getpid(), 'can only test a child process'\r\nAssertionError: can only test a child process\r\n\r\nDuring handling of the above exception, another exception occurred:\r\n\r\nTraceback (most recent call last):\r\n File \"/usr/local/lib/python3.6/dist-packages/torch/utils/data/_utils/worker.py\", line 198, in _worker_loop\r\n data = fetcher.fetch(index)\r\n File \"/usr/local/lib/python3.6/dist-packages/torch/utils/data/_utils/fetch.py\", line 44, in fetch\r\n data = [self.dataset[idx] for idx in possibly_batched_index]\r\n File \"/usr/local/lib/python3.6/dist-packages/torch/utils/data/_utils/fetch.py\", line 44, in <listcomp>\r\n data = [self.dataset[idx] for idx in possibly_batched_index]\r\n File \"<ipython-input-8-a4d9a08d114e>\", line 20, in __getitem__\r\n return_token_type_ids=True\r\n File \"/usr/local/lib/python3.6/dist-packages/transformers/tokenization_utils_base.py\", line 2405, in encode_plus\r\n **kwargs,\r\n File \"/usr/local/lib/python3.6/dist-packages/transformers/tokenization_utils_base.py\", line 2125, in _get_padding_truncation_strategies\r\n \"Truncation was not explicitly activated but `max_length` is provided a specific value, \"\r\n File \"/usr/lib/python3.6/logging/__init__.py\", line 1320, in warning\r\n self._log(WARNING, msg, args, **kwargs)\r\n File \"/usr/lib/python3.6/logging/__init__.py\", line 1444, in _log\r\n self.handle(record)\r\n File \"/usr/lib/python3.6/logging/__init__.py\", line 1454, in handle\r\n self.callHandlers(record)\r\n File \"/usr/lib/python3.6/logging/__init__.py\", line 1516, in callHandlers\r\n hdlr.handle(record)\r\n File \"/usr/lib/python3.6/logging/__init__.py\", line 865, in handle\r\n self.emit(record)\r\n File \"/usr/lib/python3.6/logging/__init__.py\", line 1000, in emit\r\n self.handleError(record)\r\n File \"/usr/lib/python3.6/logging/__init__.py\", line 917, in handleError\r\n sys.stderr.write('--- Logging error ---\\n')\r\n File \"/usr/local/lib/python3.6/dist-packages/wandb/sdk/lib/redirect.py\", line 100, in new_write\r\n cb(name, data)\r\n File \"/usr/local/lib/python3.6/dist-packages/wandb/sdk/wandb_run.py\", line 723, in _console_callback\r\n self._backend.interface.publish_output(name, data)\r\n File \"/usr/local/lib/python3.6/dist-packages/wandb/sdk/interface/interface.py\", line 153, in publish_output\r\n self._publish_output(o)\r\n File \"/usr/local/lib/python3.6/dist-packages/wandb/sdk/interface/interface.py\", line 158, in _publish_output\r\n self._publish(rec)\r\n File \"/usr/local/lib/python3.6/dist-packages/wandb/sdk/interface/interface.py\", line 456, in _publish\r\n if self._process and not self._process.is_alive():\r\n File \"/usr/lib/python3.6/multiprocessing/process.py\", line 134, in is_alive\r\n assert self._parent_pid == os.getpid(), 'can only test a child process'\r\nAssertionError: can only test a child process`\r\n", "It looks like this warning : \r\n\"Truncation was not explicitly activated but max_length is provided a specific value, \"\r\nis not handled well by wandb.\r\n\r\nThe error occurs when calling the tokenizer.\r\nMaybe you can try to specify `truncation=True` when calling the tokenizer to remove the warning ?\r\nOtherwise I don't know why wandb would fail on a warning. Maybe one of its logging handlers have some issues with the logging of tokenizers. Maybe @n1t0 knows more about this ?", "I'm having a similar issue but when I try to do multiprocessing with the `DataLoader`\r\n\r\nCode to reproduce:\r\n\r\n```\r\nfrom datasets import load_dataset\r\n\r\nbook_corpus = load_dataset('bookcorpus', 'plain_text', cache_dir='/home/ad/Desktop/bookcorpus', split='train[:1%]')\r\nbook_corpus = book_corpus.map(encode, batched=True, num_proc=20, load_from_cache_file=True, batch_size=5000)\r\nbook_corpus.set_format(type='torch', columns=['text', \"input_ids\", \"attention_mask\", \"token_type_ids\"])\r\n\r\nfrom transformers import DataCollatorForWholeWordMask\r\nfrom transformers import Trainer, TrainingArguments\r\n\r\ndata_collator = DataCollatorForWholeWordMask(\r\n tokenizer=tokenizer, mlm=True, mlm_probability=0.15)\r\n\r\ntraining_args = TrainingArguments(\r\n output_dir=\"./mobile_linear_att_8L_128_128_03layerdrop_shared\",\r\n overwrite_output_dir=True,\r\n num_train_epochs=1,\r\n per_device_train_batch_size=64,\r\n save_steps=50,\r\n save_total_limit=2,\r\n logging_first_step=True,\r\n warmup_steps=100,\r\n logging_steps=50,\r\n gradient_accumulation_steps=1,\r\n fp16=True,\r\n **dataloader_num_workers=10**,\r\n)\r\n\r\ntrainer = Trainer(\r\n model=model,\r\n args=training_args,\r\n data_collator=data_collator,\r\n train_dataset=book_corpus,\r\n tokenizer=tokenizer)\r\n\r\ntrainer.train()\r\n```\r\n\r\n```\r\n---------------------------------------------------------------------------\r\nAssertionError Traceback (most recent call last)\r\n<timed eval> in <module>\r\n\r\n~/anaconda3/envs/tfm/lib/python3.6/site-packages/transformers/trainer.py in train(self, model_path, trial)\r\n 869 self.control = self.callback_handler.on_epoch_begin(self.args, self.state, self.control)\r\n 870 \r\n--> 871 for step, inputs in enumerate(epoch_iterator):\r\n 872 \r\n 873 # Skip past any already trained steps if resuming training\r\n\r\n~/anaconda3/envs/tfm/lib/python3.6/site-packages/torch/utils/data/dataloader.py in __next__(self)\r\n 433 if self._sampler_iter is None:\r\n 434 self._reset()\r\n--> 435 data = self._next_data()\r\n 436 self._num_yielded += 1\r\n 437 if self._dataset_kind == _DatasetKind.Iterable and \\\r\n\r\n~/anaconda3/envs/tfm/lib/python3.6/site-packages/torch/utils/data/dataloader.py in _next_data(self)\r\n 1083 else:\r\n 1084 del self._task_info[idx]\r\n-> 1085 return self._process_data(data)\r\n 1086 \r\n 1087 def _try_put_index(self):\r\n\r\n~/anaconda3/envs/tfm/lib/python3.6/site-packages/torch/utils/data/dataloader.py in _process_data(self, data)\r\n 1109 self._try_put_index()\r\n 1110 if isinstance(data, ExceptionWrapper):\r\n-> 1111 data.reraise()\r\n 1112 return data\r\n 1113 \r\n\r\n~/anaconda3/envs/tfm/lib/python3.6/site-packages/torch/_utils.py in reraise(self)\r\n 426 # have message field\r\n 427 raise self.exc_type(message=msg)\r\n--> 428 raise self.exc_type(msg)\r\n 429 \r\n 430 \r\n\r\nAssertionError: Caught AssertionError in DataLoader worker process 0.\r\nOriginal Traceback (most recent call last):\r\n File \"/home/ad/anaconda3/envs/tfm/lib/python3.6/site-packages/torch/utils/data/_utils/worker.py\", line 198, in _worker_loop\r\n data = fetcher.fetch(index)\r\n File \"/home/ad/anaconda3/envs/tfm/lib/python3.6/site-packages/torch/utils/data/_utils/fetch.py\", line 44, in fetch\r\n data = [self.dataset[idx] for idx in possibly_batched_index]\r\n File \"/home/ad/anaconda3/envs/tfm/lib/python3.6/site-packages/torch/utils/data/_utils/fetch.py\", line 44, in <listcomp>\r\n data = [self.dataset[idx] for idx in possibly_batched_index]\r\n File \"/home/ad/anaconda3/envs/tfm/lib/python3.6/site-packages/datasets/arrow_dataset.py\", line 1087, in __getitem__\r\n format_kwargs=self._format_kwargs,\r\n File \"/home/ad/anaconda3/envs/tfm/lib/python3.6/site-packages/datasets/arrow_dataset.py\", line 1074, in _getitem\r\n format_kwargs=format_kwargs,\r\n File \"/home/ad/anaconda3/envs/tfm/lib/python3.6/site-packages/datasets/arrow_dataset.py\", line 890, in _convert_outputs\r\n v = map_nested(command, v, **map_nested_kwargs)\r\n File \"/home/ad/anaconda3/envs/tfm/lib/python3.6/site-packages/datasets/utils/py_utils.py\", line 225, in map_nested\r\n return function(data_struct)\r\n File \"/home/ad/anaconda3/envs/tfm/lib/python3.6/site-packages/datasets/arrow_dataset.py\", line 851, in command\r\n return torch.tensor(x, **format_kwargs)\r\n File \"/home/ad/anaconda3/envs/tfm/lib/python3.6/warnings.py\", line 101, in _showwarnmsg\r\n _showwarnmsg_impl(msg)\r\n File \"/home/ad/anaconda3/envs/tfm/lib/python3.6/warnings.py\", line 30, in _showwarnmsg_impl\r\n file.write(text)\r\n File \"/home/ad/anaconda3/envs/tfm/lib/python3.6/site-packages/wandb/sdk/lib/redirect.py\", line 100, in new_write\r\n cb(name, data)\r\n File \"/home/ad/anaconda3/envs/tfm/lib/python3.6/site-packages/wandb/sdk/wandb_run.py\", line 723, in _console_callback\r\n self._backend.interface.publish_output(name, data)\r\n File \"/home/ad/anaconda3/envs/tfm/lib/python3.6/site-packages/wandb/sdk/interface/interface.py\", line 153, in publish_output\r\n self._publish_output(o)\r\n File \"/home/ad/anaconda3/envs/tfm/lib/python3.6/site-packages/wandb/sdk/interface/interface.py\", line 158, in _publish_output\r\n self._publish(rec)\r\n File \"/home/ad/anaconda3/envs/tfm/lib/python3.6/site-packages/wandb/sdk/interface/interface.py\", line 456, in _publish\r\n if self._process and not self._process.is_alive():\r\n File \"/home/ad/anaconda3/envs/tfm/lib/python3.6/multiprocessing/process.py\", line 134, in is_alive\r\n assert self._parent_pid == os.getpid(), 'can only test a child process'\r\nAssertionError: can only test a child process\r\n```\r\n\r\nAs a workaround I have commented line 456 and 457 in `/home/ad/anaconda3/envs/tfm/lib/python3.6/site-packages/wandb/sdk/interface/interface.py`", "Isn't it more the pytorch warning on the use of non-writable memory for tensor that trigger this here @lhoestq? (since it seems to be a warning triggered in `torch.tensor()`", "Yep this time this is a warning from pytorch that causes wandb to not work properly.\r\nCould this by a wandb issue ?", "Hi @timothyjlaurent @gaceladri \r\nIf you're running `transformers` from `master` you can try setting the env var `WAND_DISABLE=true` (from https://github.com/huggingface/transformers/pull/9896) and try again ?\r\nThis issue might be related to https://github.com/huggingface/transformers/issues/9623 ", "I have commented the lines that cause my code break. I'm now seeing my reports on Wandb and my code does not break. I am training now, so I will check probably in 6 hours. I suppose that setting wandb disable will work as well." ]

[ { "id": 2067376369, "node_id": "MDU6TGFiZWwyMDY3Mzc2MzY5", "url": "https://api.github.com/repos/huggingface/datasets/labels/dataset%20request", "name": "dataset request", "color": "e99695", "default": false, "description": "Requesting to add a new dataset" } ]

[]

[ "Hi @yjernite I'm new but wanted to contribute. Has anyone already taken this problem and do you think it is suitable for newbies?", "Hi @tenjjin! This dataset is still up for grabs! Here's the link with the guide to add it. You should play around with the library first (download and look at a few datasets), then follow the steps here:\r\n\r\nhttps://github.com/huggingface/datasets/blob/master/ADD_NEW_DATASET.md", "Closed by #1399 " ]

[]

[]

[]

[]

[]

[]

[ { "id": 2067393914, "node_id": "MDU6TGFiZWwyMDY3MzkzOTE0", "url": "https://api.github.com/repos/huggingface/datasets/labels/metric%20bug", "name": "metric bug", "color": "25b21e", "default": false, "description": "A bug in a metric script" } ]

[]

[ "Thanks for reporting ! That's a bug indeed\r\nIf you want to contribute, feel free to fix this issue and open a PR :)", "This error is because of a mismatch between `datasets` and `bert_score`. With `datasets=1.1.2` and `bert_score>=0.3.6` it works ok. So `pip install -U bert_score` should fix the problem. ", "Thanks for the heads up @pvl and for the PR as well :)", "Hello everyone,\r\n\r\nI think the problem is not solved: \r\n\r\n```\r\nfrom datasets import load_metric\r\nmetric=load_metric('bertscore')\r\nmetric.compute(\r\n predictions=predictions,\r\n references=references,\r\n lang='fr',\r\n rescale_with_baseline=True\r\n)\r\nTypeError: get_hash() missing 2 required positional arguments: 'use_custom_baseline' and 'use_fast_tokenizer'\r\n```\r\nThis code is produced using `Python 3.6.9 datasets==1.1.2 and bert_score==0.3.10`", "Hi ! This has been fixed by https://github.com/huggingface/datasets/pull/2770, we'll do a new release soon to make the fix available :)\r\n\r\nIn the meantime please use an older version of `bert_score`" ]

[]

[]

[ "Right now multiprocessing only runs on single node.\r\n\r\nHowever it's probably possible to extend it to support multi nodes. Indeed we're using the `multiprocess` library from the `pathos` project to do multiprocessing in `datasets`, and `pathos` is made to support parallelism on several nodes. More info about pathos [on the pathos repo](https://github.com/uqfoundation/pathos).\r\n\r\nIf you're familiar with pathos or if you want to give it a try, it could be a nice addition to the library :)" ]