Datasets:

---

RajkNakka

/

github-issues-comments

like 0

[]

[]

[ "Hi! This code works for me locally or in Colab. What's the output of `python -c \"import pyarrow as pa; print(pa.__version__)\"` when you run it inside your environment?", "Thanks for looking into this!\r\nThe output of `python -c \"import pyarrow as pa; print(pa.__version__)\"` is:\r\n```\r\n11.0.0\r\n```\r\n\r\nI did the following to setup the environment:\r\n```\r\nconda create -n datasets_debug python=3.9\r\nconda activate datasets_debug\r\npip install datasets==2.9.0\r\n```\r\n\r\nI just tested this on another machine (Ubuntu 18.04.6 LTS) with the same result as mentioned in the issue description.\r\n" ]

[]

[]

[ "I agree that it would be a good idea to introduce a `combiner` argument in another PR.\r\n\r\nI did take quite a lot of inspiration from the implementation of `map`, but it did not seem obvious how to resuse `map` for the implementation. Do you have any suggestions, i could give a try?\r\n\r\nThose were exactly my thoughts, regarding the non-obvious initializer for batched and formatted datasets, so i agree! I'll introduce a `initializer` argument, and have it mandatory when `batched=True`.", "I added `initializer`. It is optional for `batched=False` and mandatory for `batched=True`. It has to be of the same length as `input_columns`, if `input_columns=None` it has to have the same length as `_data.column_names`. \r\n\r\nIf the initializer is not set for `batched=False` the first example is set as the `initializer`. \r\n\r\nThe initializer is used to initiliaze for each shard, so that means if that:\r\n```python\r\ndset = Dataset.from_dict({\"x\": [1, 2, 3]})\r\nsum_reduce = lambda x, y: x + y\r\nreduction = dset.reduce(sum_reduce, batched=True, initializer=1, input_columns='x', num_proc=2)\r\n# reduction is 8, i.e. reduction + num_proc * initializer\r\n```", "> I added initializer. It is optional for batched=False and mandatory for batched=True. It has to be of the same length as input_columns, if input_columns=None it has to have the same length as _data.column_names.\r\n> \r\n> If the initializer is not set for batched=False the first example is set as the initializer.\r\n\r\nSounds good to me !\r\n\r\n> The initializer is used to initiliaze for each shard, so that means if that:\r\n> \r\n> ```python\r\n> dset = Dataset.from_dict({\"x\": [1, 2, 3]})\r\n> sum_reduce = lambda x, y: x + y\r\n> reduction = dset.reduce(sum_reduce, batched=True, initializer=1, input_columns='x', num_proc=2)\r\n> # reduction is 8, i.e. reduction + num_proc * initializer\r\n> ```\r\n\r\nHmm this can be confusing for some users. Maybe we should consider making `combiner` mandatory for multiprocessing.\r\n\r\nIf we agree on this, maybe for this PR you can either:\r\n- remove multiprocessing (and we add combiner + multiprocessing in a subsequent PR)\r\n- OR add `combiner` directly\r\n\r\nMaybe we can get more feedback from @huggingface/datasets as well", "> > I added initializer. It is optional for batched=False and mandatory for batched=True. It has to be of the same length as input_columns, if input_columns=None it has to have the same length as _data.column_names.\r\n> > If the initializer is not set for batched=False the first example is set as the initializer.\r\n> \r\n> Sounds good to me !\r\n> \r\n> > The initializer is used to initiliaze for each shard, so that means if that:\r\n> > ```python\r\n> > dset = Dataset.from_dict({\"x\": [1, 2, 3]})\r\n> > sum_reduce = lambda x, y: x + y\r\n> > reduction = dset.reduce(sum_reduce, batched=True, initializer=1, input_columns='x', num_proc=2)\r\n> > # reduction is 8, i.e. reduction + num_proc * initializer\r\n> > ```\r\n> \r\n> Hmm this can be confusing for some users. Maybe we should consider making `combiner` mandatory for multiprocessing.\r\n> \r\n> If we agree on this, maybe for this PR you can either:\r\n> \r\n> * remove multiprocessing (and we add combiner + multiprocessing in a subsequent PR)\r\n> * OR add `combiner` directly\r\n> \r\n> Maybe we can get more feedback from @huggingface/datasets as well\r\n\r\nI think i prefer adding `combiner` in this PR. I think ill make `combiner` mandatory for `batched=True`, instead of assuming that `combiner=function`. Ill look at this one of the coming days. Also at some point i have to define `reduce` for `DatasetDict`, and not just `Dataset`.", "I added the `combiner` parameter as described. I added some examples in the docstring, as i felt it might still be a bit confusing what happens during multiprocessing / batching.\r\n\r\nStill need to look at `DatasetDict`.", "The docs for this PR live [here](https://moon-ci-docs.huggingface.co/docs/datasets/pr_5533). All of your documentation changes will be reflected on that endpoint.", "Feel free to merge `main` into your branch - we fixed some CI failures today", "The proposed API doesn't seem intuitive to me - one can already use `functools.reduce` or `Dataset.map` for this purpose ([Colab](https://colab.research.google.com/drive/1jCLv31Y4cDfqD0lhO0AnqEv3Or-LLvWe?usp=sharing) with examples), so perhaps we could have a section in the docs that uses these methods to perform reductions rather than introducing a new method (which needs to be maintained later)", "Thanks for sharing this google colab, it has nice examples !\r\n\r\nThough I still think `functools.reduce` with multiprocessing can be a pain - we offer something easier here:\r\n- no need to use a pool yourself\r\n- no need to use `map` just to iterate on the dataset (not its main purpose)\r\n- native support for lambdas (using dill)\r\n- the combiner is **mandatory** for multiprocessing to avoid ending up with an incorrect result as in your example\r\n\r\nHowever I agree that maintaining this can be challenging, especially if you think about how `map` already is, and if we also have to deal with dataset formatting.", "> native support for lambdas (using dill)\r\n\r\nReplacing `multiprocessing` with `multiprocess` in the example would allow that.\r\n\r\n> no need to use map just to iterate on the dataset (not its main purpose)\r\n\r\nNot the main purpose, but this was mentioned as a \"feature\" in the previous docs if I remember.\r\n\r\nAnd all this is related to the multi-processing case, which we can document.\r\n\r\nBesides the linked issue, I can't find requests for `Dataset.reduce`, which makes me think `functools.reduce` does the job for most users.", "> Besides the linked issue, I can't find requests for Dataset.reduce, which makes me think functools.reduce does the job for most users.\r\n\r\nI think @srush was looking for a way to do a word count but ended up using a single processed `map`. I also saw some users on the forum wanting to compute `max`\r\n\r\n> Not the main purpose, but this was mentioned as a \"feature\" in the previous docs if I remember.\r\n> \r\n> And all this is related to the multi-processing case, which we can document.\r\n\r\nYup indeed", "While counting is one example, I often find I want to compute different statistics over a dataset. This seems like a natural way to do it in a stateless manner.\n\n\nI guess you could use functools reduce, but that wouldn't allow batching, right?", "I've updated the [Colab](https://colab.research.google.com/drive/1jCLv31Y4cDfqD0lhO0AnqEv3Or-LLvWe?usp=sharing) with an example that reduces batches with `map` and then computes the final result. It would be nice to have a similar example (explained in detail) in the docs to show the full power of `map`.\r\n\r\nPlus, for simple reductions such as `max`, one can do `pc.max(ds.with_format(\"arrow\")[\"col\"])` to directly get the result (without loading the entire column in RAM).\r\n\r\n@srush \r\n\r\n> I guess you could use functools reduce, but that wouldn't allow batching, right?\r\n\r\nYou can use `.iter(batch_size)` to get batches\r\n ", "That `functools` tools example is clean. I didn't know about `iter`. That would handle my use case.\n\nThe stateful `map` with a global variable is pretty hairy. I don't think we should recommend people do that.\n\n", "Whenever I in the past wanted to calculate statistics for datasets I used `functools` similarly to how it's described in the colab, but I always felt it was a bit of a hassle to use it together with multiprocessing, which is why I picked up the issue, to do it \"once and for all\".", "Should i close this and open another PR, with descriptions of how to use `map` for reduction, or?", "Yes I think good documentation is the way to go here. @mariosasko 's examples are clear and efficient.\r\n\r\nMaybe we could have an `Aggregations` section in the `Process` page with some guides on how to:\r\n- use `.map()` to compute aggregates\r\n- use `.with_format(\"arrow\")` for max, min, etc. to save RAM and get max speed\r\n- use a multiprocessed `.map()` to get partial results in parallel and combine them (max text length example)\r\n- (advanced) use multiprocessing with an arbitrary accumulator (word count example)\r\n\r\nAnd also a new conceptual guide on `Multiprocessed mapping` to say that it helps speed up CPU intensive processing but why it may lead to incorrect results when computing aggregates.\r\n\r\ncc @stevhliu for visibility and if you have some comments", "I would create a `Reduce` - to be more exact - subsection under `Map` to demonstrate these examples since we're showing how they can be done with the `Dataset.map` function. It'd also be good to add a link to the new concept guide from this section to solidify user understanding :)", "Coolio. Ill close this PR and get going on another one adding what we've discussed during the next couple of days!" ]

[]

[]

[ "Hi! You can get this behavior by specifying `stratify_by_column=\"label\"` in `train_test_split`.\r\n\r\nThis is the full example:\r\n```python\r\nimport numpy as np\r\nfrom datasets import Dataset, ClassLabel\r\n\r\ndata = [\r\n {'label': 0, 'text': \"example1\"},\r\n {'label': 1, 'text': \"example2\"},\r\n {'label': 1, 'text': \"example3\"},\r\n {'label': 1, 'text': \"example4\"},\r\n {'label': 0, 'text': \"example5\"},\r\n {'label': 1, 'text': \"example6\"},\r\n {'label': 2, 'text': \"example7\"},\r\n {'label': 2, 'text': \"example8\"}\r\n]\r\n\r\nfor _ in range(10):\r\n data_set = Dataset.from_list(data)\r\n data_set = data_set.cast_column(\"label\", ClassLabel(num_classes=3))\r\n data_set = data_set.train_test_split(test_size=0.5, stratify_by_column=\"label\")\r\n unique_labels_train = np.unique(data_set[\"train\"][:][\"label\"])\r\n unique_labels_test = np.unique(data_set[\"test\"][:][\"label\"])\r\n assert len(unique_labels_train) >= len(unique_labels_test) \r\n```\r\n" ]

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

[]

[]

[]

[]

[ "_The documentation is not available anymore as the PR was closed or merged._", "<details>\n<summary>Show benchmarks</summary>\n\nPyArrow==6.0.0\n\n<details>\n<summary>Show updated benchmarks!</summary>\n\n### Benchmark: benchmark_array_xd.json\n\n| metric | read_batch_formatted_as_numpy after write_array2d | read_batch_formatted_as_numpy after write_flattened_sequence | read_batch_formatted_as_numpy after write_nested_sequence | read_batch_unformated after write_array2d | read_batch_unformated after write_flattened_sequence | read_batch_unformated after write_nested_sequence | read_col_formatted_as_numpy after write_array2d | read_col_formatted_as_numpy after write_flattened_sequence | read_col_formatted_as_numpy after write_nested_sequence | read_col_unformated after write_array2d | read_col_unformated after write_flattened_sequence | read_col_unformated after write_nested_sequence | read_formatted_as_numpy after write_array2d | read_formatted_as_numpy after write_flattened_sequence | read_formatted_as_numpy after write_nested_sequence | read_unformated after write_array2d | read_unformated after write_flattened_sequence | read_unformated after write_nested_sequence | write_array2d | write_flattened_sequence | write_nested_sequence |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.008837 / 0.011353 (-0.002516) | 0.004608 / 0.011008 (-0.006400) | 0.101821 / 0.038508 (0.063312) | 0.030300 / 0.023109 (0.007191) | 0.301275 / 0.275898 (0.025377) | 0.365027 / 0.323480 (0.041547) | 0.007043 / 0.007986 (-0.000943) | 0.003493 / 0.004328 (-0.000835) | 0.078444 / 0.004250 (0.074194) | 0.036963 / 0.037052 (-0.000089) | 0.310510 / 0.258489 (0.052020) | 0.343769 / 0.293841 (0.049928) | 0.033560 / 0.128546 (-0.094986) | 0.011427 / 0.075646 (-0.064220) | 0.323542 / 0.419271 (-0.095730) | 0.043063 / 0.043533 (-0.000470) | 0.308869 / 0.255139 (0.053730) | 0.326436 / 0.283200 (0.043236) | 0.091775 / 0.141683 (-0.049908) | 1.471020 / 1.452155 (0.018865) | 1.494328 / 1.492716 (0.001612) |\n\n### Benchmark: benchmark_getitem\\_100B.json\n\n| metric | get_batch_of\\_1024\\_random_rows | get_batch_of\\_1024\\_rows | get_first_row | get_last_row |\n|--------|---|---|---|---|\n| new / old (diff) | 0.009299 / 0.018006 (-0.008707) | 0.415705 / 0.000490 (0.415215) | 0.002406 / 0.000200 (0.002206) | 0.000066 / 0.000054 (0.000012) |\n\n### Benchmark: benchmark_indices_mapping.json\n\n| metric | select | shard | shuffle | sort | train_test_split |\n|--------|---|---|---|---|---|\n| new / old (diff) | 0.022959 / 0.037411 (-0.014452) | 0.097111 / 0.014526 (0.082585) | 0.103399 / 0.176557 (-0.073157) | 0.144385 / 0.737135 (-0.592750) | 0.109069 / 0.296338 (-0.187269) |\n\n### Benchmark: benchmark_iterating.json\n\n| metric | read 5000 | read 50000 | read_batch 50000 10 | read_batch 50000 100 | read_batch 50000 1000 | read_formatted numpy 5000 | read_formatted pandas 5000 | read_formatted tensorflow 5000 | read_formatted torch 5000 | read_formatted_batch numpy 5000 10 | read_formatted_batch numpy 5000 1000 | shuffled read 5000 | shuffled read 50000 | shuffled read_batch 50000 10 | shuffled read_batch 50000 100 | shuffled read_batch 50000 1000 | shuffled read_formatted numpy 5000 | shuffled read_formatted_batch numpy 5000 10 | shuffled read_formatted_batch numpy 5000 1000 |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.417796 / 0.215209 (0.202587) | 4.158198 / 2.077655 (2.080543) | 1.862036 / 1.504120 (0.357916) | 1.650130 / 1.541195 (0.108936) | 1.717150 / 1.468490 (0.248660) | 0.691704 / 4.584777 (-3.893073) | 3.328254 / 3.745712 (-0.417458) | 1.850070 / 5.269862 (-3.419792) | 1.154331 / 4.565676 (-3.411346) | 0.082199 / 0.424275 (-0.342076) | 0.012226 / 0.007607 (0.004619) | 0.522491 / 0.226044 (0.296446) | 5.244181 / 2.268929 (2.975253) | 2.286651 / 55.444624 (-53.157973) | 1.954439 / 6.876477 (-4.922038) | 1.992052 / 2.142072 (-0.150020) | 0.804779 / 4.805227 (-4.000449) | 0.147341 / 6.500664 (-6.353323) | 0.063863 / 0.075469 (-0.011606) |\n\n### Benchmark: benchmark_map_filter.json\n\n| metric | filter | map fast-tokenizer batched | map identity | map identity batched | map no-op batched | map no-op batched numpy | map no-op batched pandas | map no-op batched pytorch | map no-op batched tensorflow |\n|--------|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 1.270778 / 1.841788 (-0.571010) | 13.676378 / 8.074308 (5.602070) | 14.253498 / 10.191392 (4.062106) | 0.170748 / 0.680424 (-0.509676) | 0.028451 / 0.534201 (-0.505750) | 0.395034 / 0.579283 (-0.184249) | 0.407512 / 0.434364 (-0.026852) | 0.466740 / 0.540337 (-0.073598) | 0.564338 / 1.386936 (-0.822598) |\n\n</details>\nPyArrow==latest\n\n<details>\n<summary>Show updated benchmarks!</summary>\n\n### Benchmark: benchmark_array_xd.json\n\n| metric | read_batch_formatted_as_numpy after write_array2d | read_batch_formatted_as_numpy after write_flattened_sequence | read_batch_formatted_as_numpy after write_nested_sequence | read_batch_unformated after write_array2d | read_batch_unformated after write_flattened_sequence | read_batch_unformated after write_nested_sequence | read_col_formatted_as_numpy after write_array2d | read_col_formatted_as_numpy after write_flattened_sequence | read_col_formatted_as_numpy after write_nested_sequence | read_col_unformated after write_array2d | read_col_unformated after write_flattened_sequence | read_col_unformated after write_nested_sequence | read_formatted_as_numpy after write_array2d | read_formatted_as_numpy after write_flattened_sequence | read_formatted_as_numpy after write_nested_sequence | read_unformated after write_array2d | read_unformated after write_flattened_sequence | read_unformated after write_nested_sequence | write_array2d | write_flattened_sequence | write_nested_sequence |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.006733 / 0.011353 (-0.004620) | 0.004635 / 0.011008 (-0.006373) | 0.075464 / 0.038508 (0.036956) | 0.027732 / 0.023109 (0.004623) | 0.343622 / 0.275898 (0.067724) | 0.380388 / 0.323480 (0.056908) | 0.005177 / 0.007986 (-0.002808) | 0.003435 / 0.004328 (-0.000893) | 0.074546 / 0.004250 (0.070296) | 0.039115 / 0.037052 (0.002063) | 0.342207 / 0.258489 (0.083718) | 0.390324 / 0.293841 (0.096483) | 0.031665 / 0.128546 (-0.096882) | 0.011695 / 0.075646 (-0.063951) | 0.085788 / 0.419271 (-0.333484) | 0.042423 / 0.043533 (-0.001110) | 0.340748 / 0.255139 (0.085609) | 0.372813 / 0.283200 (0.089614) | 0.092395 / 0.141683 (-0.049288) | 1.502158 / 1.452155 (0.050004) | 1.618233 / 1.492716 (0.125516) |\n\n### Benchmark: benchmark_getitem\\_100B.json\n\n| metric | get_batch_of\\_1024\\_random_rows | get_batch_of\\_1024\\_rows | get_first_row | get_last_row |\n|--------|---|---|---|---|\n| new / old (diff) | 0.224451 / 0.018006 (0.206444) | 0.398712 / 0.000490 (0.398222) | 0.002739 / 0.000200 (0.002539) | 0.000074 / 0.000054 (0.000020) |\n\n### Benchmark: benchmark_indices_mapping.json\n\n| metric | select | shard | shuffle | sort | train_test_split |\n|--------|---|---|---|---|---|\n| new / old (diff) | 0.025393 / 0.037411 (-0.012018) | 0.100480 / 0.014526 (0.085954) | 0.106913 / 0.176557 (-0.069644) | 0.148639 / 0.737135 (-0.588496) | 0.110098 / 0.296338 (-0.186240) |\n\n### Benchmark: benchmark_iterating.json\n\n| metric | read 5000 | read 50000 | read_batch 50000 10 | read_batch 50000 100 | read_batch 50000 1000 | read_formatted numpy 5000 | read_formatted pandas 5000 | read_formatted tensorflow 5000 | read_formatted torch 5000 | read_formatted_batch numpy 5000 10 | read_formatted_batch numpy 5000 1000 | shuffled read 5000 | shuffled read 50000 | shuffled read_batch 50000 10 | shuffled read_batch 50000 100 | shuffled read_batch 50000 1000 | shuffled read_formatted numpy 5000 | shuffled read_formatted_batch numpy 5000 10 | shuffled read_formatted_batch numpy 5000 1000 |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.439359 / 0.215209 (0.224150) | 4.396801 / 2.077655 (2.319146) | 2.069809 / 1.504120 (0.565689) | 1.851014 / 1.541195 (0.309820) | 1.885003 / 1.468490 (0.416513) | 0.701387 / 4.584777 (-3.883390) | 3.404943 / 3.745712 (-0.340769) | 1.874506 / 5.269862 (-3.395355) | 1.174925 / 4.565676 (-3.390752) | 0.083282 / 0.424275 (-0.340993) | 0.012352 / 0.007607 (0.004745) | 0.543058 / 0.226044 (0.317013) | 5.458186 / 2.268929 (3.189258) | 2.562159 / 55.444624 (-52.882466) | 2.198810 / 6.876477 (-4.677667) | 2.238976 / 2.142072 (0.096903) | 0.810958 / 4.805227 (-3.994269) | 0.153341 / 6.500664 (-6.347323) | 0.067773 / 0.075469 (-0.007696) |\n\n### Benchmark: benchmark_map_filter.json\n\n| metric | filter | map fast-tokenizer batched | map identity | map identity batched | map no-op batched | map no-op batched numpy | map no-op batched pandas | map no-op batched pytorch | map no-op batched tensorflow |\n|--------|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 1.303938 / 1.841788 (-0.537850) | 14.170363 / 8.074308 (6.096055) | 13.727012 / 10.191392 (3.535620) | 0.129118 / 0.680424 (-0.551306) | 0.016746 / 0.534201 (-0.517455) | 0.382759 / 0.579283 (-0.196524) | 0.391070 / 0.434364 (-0.043294) | 0.461197 / 0.540337 (-0.079141) | 0.557641 / 1.386936 (-0.829295) |\n\n</details>\n</details>\n\n![](https://cml.dev/watermark.png#004bba88db03fb87d57252e38a4d7abdb0a5f0a9 \"CML watermark\")\n" ]

[]

[]

[ "_The documentation is not available anymore as the PR was closed or merged._", "Hmm, should this also be updated in `Dataset.load_from_disk` and `DatasetDict.load_from_disk`? https://github.com/huggingface/datasets/pull/5466 As there the paths are joined using `Path(..., ...)` and it won't work on Windows OS according to that PR, right?", "Hi, @lhoestq could you review this PR? Thank you in advance and sorry for the ping 🤗 ", "Besides that, I was also thinking of adding a `skip_validation` boolean arg in both `Dataset.load_from_disk` and `DatasetDict.load_from_disk` to avoid duplicating those calls too when those functions are called from `datasets.load_from_disk`.\r\n\r\nSo that `skip_validation` is set to `False` by default, but passed as `True` if called from `datasets.load_from_disk`, and that just affects the file checking part of the code on both functions, do you agree @lhoestq?", "I think we should always verify", "> I think we should always verify\r\n\r\nBut with the current way we're also verifying twice right? First on `datasets.load_from_disk` then on `Dataset.load_from_disk`, right?\r\n\r\nMaybe a warning before calling either `Dataset.load_from_disk` or `DatasetDict.load_from_disk` is enough?\r\n\r\ne.g. **\"Consider using `Dataset.load_from_disk` instead to avoid `fsspec` from verifying the presence of `dataset_info.json` and `state.json` in the remote filesystem twice.\"** to be showed just when `fs` is remote.", "I don't think it's worth adding a new argument just for that. Usually we keep the set of arguments to the strict minimum", "> I don't think it's worth adding a new argument just for that. Usually we keep the set of arguments to the strict minimum\r\n\r\nWhat about the warning?\r\n\r\nAnyway, if you don't think that's worth it feel free to merge 👍🏻 ", "> What about the warning?\r\n\r\nWe may show warnings for suggestions, but only if the user does a very unoptimized thing. Here we're not at that level ^^'", "Thanks for the explanation and feedback @lhoestq 🤗 ", "> Thank you :) Added my last suggestions:\r\n\r\nThanks for the feedback, I agree with everything besides one nit! 👍🏻 ", "<details>\n<summary>Show benchmarks</summary>\n\nPyArrow==6.0.0\n\n<details>\n<summary>Show updated benchmarks!</summary>\n\n### Benchmark: benchmark_array_xd.json\n\n| metric | read_batch_formatted_as_numpy after write_array2d | read_batch_formatted_as_numpy after write_flattened_sequence | read_batch_formatted_as_numpy after write_nested_sequence | read_batch_unformated after write_array2d | read_batch_unformated after write_flattened_sequence | read_batch_unformated after write_nested_sequence | read_col_formatted_as_numpy after write_array2d | read_col_formatted_as_numpy after write_flattened_sequence | read_col_formatted_as_numpy after write_nested_sequence | read_col_unformated after write_array2d | read_col_unformated after write_flattened_sequence | read_col_unformated after write_nested_sequence | read_formatted_as_numpy after write_array2d | read_formatted_as_numpy after write_flattened_sequence | read_formatted_as_numpy after write_nested_sequence | read_unformated after write_array2d | read_unformated after write_flattened_sequence | read_unformated after write_nested_sequence | write_array2d | write_flattened_sequence | write_nested_sequence |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.011556 / 0.011353 (0.000203) | 0.006213 / 0.011008 (-0.004796) | 0.132390 / 0.038508 (0.093882) | 0.034609 / 0.023109 (0.011500) | 0.361156 / 0.275898 (0.085258) | 0.402524 / 0.323480 (0.079044) | 0.009138 / 0.007986 (0.001152) | 0.005728 / 0.004328 (0.001399) | 0.115406 / 0.004250 (0.111156) | 0.041440 / 0.037052 (0.004388) | 0.370232 / 0.258489 (0.111742) | 0.409944 / 0.293841 (0.116103) | 0.053803 / 0.128546 (-0.074744) | 0.022029 / 0.075646 (-0.053617) | 0.400325 / 0.419271 (-0.018946) | 0.055324 / 0.043533 (0.011791) | 0.368699 / 0.255139 (0.113560) | 0.391836 / 0.283200 (0.108636) | 0.099356 / 0.141683 (-0.042327) | 1.687881 / 1.452155 (0.235726) | 1.752202 / 1.492716 (0.259485) |\n\n### Benchmark: benchmark_getitem\\_100B.json\n\n| metric | get_batch_of\\_1024\\_random_rows | get_batch_of\\_1024\\_rows | get_first_row | get_last_row |\n|--------|---|---|---|---|\n| new / old (diff) | 0.012992 / 0.018006 (-0.005014) | 0.518756 / 0.000490 (0.518267) | 0.004702 / 0.000200 (0.004502) | 0.000105 / 0.000054 (0.000050) |\n\n### Benchmark: benchmark_indices_mapping.json\n\n| metric | select | shard | shuffle | sort | train_test_split |\n|--------|---|---|---|---|---|\n| new / old (diff) | 0.028371 / 0.037411 (-0.009041) | 0.127058 / 0.014526 (0.112532) | 0.136908 / 0.176557 (-0.039649) | 0.210168 / 0.737135 (-0.526968) | 0.139600 / 0.296338 (-0.156738) |\n\n### Benchmark: benchmark_iterating.json\n\n| metric | read 5000 | read 50000 | read_batch 50000 10 | read_batch 50000 100 | read_batch 50000 1000 | read_formatted numpy 5000 | read_formatted pandas 5000 | read_formatted tensorflow 5000 | read_formatted torch 5000 | read_formatted_batch numpy 5000 10 | read_formatted_batch numpy 5000 1000 | shuffled read 5000 | shuffled read 50000 | shuffled read_batch 50000 10 | shuffled read_batch 50000 100 | shuffled read_batch 50000 1000 | shuffled read_formatted numpy 5000 | shuffled read_formatted_batch numpy 5000 10 | shuffled read_formatted_batch numpy 5000 1000 |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.570901 / 0.215209 (0.355692) | 5.967213 / 2.077655 (3.889558) | 2.286745 / 1.504120 (0.782626) | 1.950682 / 1.541195 (0.409487) | 2.062536 / 1.468490 (0.594046) | 1.255671 / 4.584777 (-3.329106) | 5.454951 / 3.745712 (1.709238) | 3.076429 / 5.269862 (-2.193433) | 2.082871 / 4.565676 (-2.482806) | 0.150069 / 0.424275 (-0.274206) | 0.014864 / 0.007607 (0.007257) | 0.774672 / 0.226044 (0.548627) | 7.873992 / 2.268929 (5.605064) | 3.196165 / 55.444624 (-52.248459) | 2.366854 / 6.876477 (-4.509623) | 2.407381 / 2.142072 (0.265309) | 1.419130 / 4.805227 (-3.386097) | 0.249210 / 6.500664 (-6.251454) | 0.088648 / 0.075469 (0.013179) |\n\n### Benchmark: benchmark_map_filter.json\n\n| metric | filter | map fast-tokenizer batched | map identity | map identity batched | map no-op batched | map no-op batched numpy | map no-op batched pandas | map no-op batched pytorch | map no-op batched tensorflow |\n|--------|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 1.528368 / 1.841788 (-0.313420) | 17.554000 / 8.074308 (9.479692) | 20.773300 / 10.191392 (10.581908) | 0.216903 / 0.680424 (-0.463521) | 0.046544 / 0.534201 (-0.487657) | 0.538238 / 0.579283 (-0.041045) | 0.673926 / 0.434364 (0.239562) | 0.656108 / 0.540337 (0.115770) | 0.774026 / 1.386936 (-0.612910) |\n\n</details>\nPyArrow==latest\n\n<details>\n<summary>Show updated benchmarks!</summary>\n\n### Benchmark: benchmark_array_xd.json\n\n| metric | read_batch_formatted_as_numpy after write_array2d | read_batch_formatted_as_numpy after write_flattened_sequence | read_batch_formatted_as_numpy after write_nested_sequence | read_batch_unformated after write_array2d | read_batch_unformated after write_flattened_sequence | read_batch_unformated after write_nested_sequence | read_col_formatted_as_numpy after write_array2d | read_col_formatted_as_numpy after write_flattened_sequence | read_col_formatted_as_numpy after write_nested_sequence | read_col_unformated after write_array2d | read_col_unformated after write_flattened_sequence | read_col_unformated after write_nested_sequence | read_formatted_as_numpy after write_array2d | read_formatted_as_numpy after write_flattened_sequence | read_formatted_as_numpy after write_nested_sequence | read_unformated after write_array2d | read_unformated after write_flattened_sequence | read_unformated after write_nested_sequence | write_array2d | write_flattened_sequence | write_nested_sequence |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.010177 / 0.011353 (-0.001176) | 0.006334 / 0.011008 (-0.004675) | 0.100097 / 0.038508 (0.061589) | 0.039996 / 0.023109 (0.016887) | 0.420225 / 0.275898 (0.144327) | 0.437694 / 0.323480 (0.114214) | 0.007987 / 0.007986 (0.000002) | 0.005782 / 0.004328 (0.001454) | 0.106421 / 0.004250 (0.102171) | 0.046993 / 0.037052 (0.009941) | 0.397304 / 0.258489 (0.138815) | 0.441780 / 0.293841 (0.147939) | 0.064594 / 0.128546 (-0.063952) | 0.020823 / 0.075646 (-0.054823) | 0.108854 / 0.419271 (-0.310417) | 0.076457 / 0.043533 (0.032924) | 0.401712 / 0.255139 (0.146573) | 0.459292 / 0.283200 (0.176093) | 0.125044 / 0.141683 (-0.016639) | 1.765531 / 1.452155 (0.313377) | 1.845429 / 1.492716 (0.352713) |\n\n### Benchmark: benchmark_getitem\\_100B.json\n\n| metric | get_batch_of\\_1024\\_random_rows | get_batch_of\\_1024\\_rows | get_first_row | get_last_row |\n|--------|---|---|---|---|\n| new / old (diff) | 0.225549 / 0.018006 (0.207543) | 0.524402 / 0.000490 (0.523913) | 0.006994 / 0.000200 (0.006794) | 0.000120 / 0.000054 (0.000065) |\n\n### Benchmark: benchmark_indices_mapping.json\n\n| metric | select | shard | shuffle | sort | train_test_split |\n|--------|---|---|---|---|---|\n| new / old (diff) | 0.033787 / 0.037411 (-0.003624) | 0.144895 / 0.014526 (0.130369) | 0.147185 / 0.176557 (-0.029371) | 0.228227 / 0.737135 (-0.508908) | 0.164967 / 0.296338 (-0.131371) |\n\n### Benchmark: benchmark_iterating.json\n\n| metric | read 5000 | read 50000 | read_batch 50000 10 | read_batch 50000 100 | read_batch 50000 1000 | read_formatted numpy 5000 | read_formatted pandas 5000 | read_formatted tensorflow 5000 | read_formatted torch 5000 | read_formatted_batch numpy 5000 10 | read_formatted_batch numpy 5000 1000 | shuffled read 5000 | shuffled read 50000 | shuffled read_batch 50000 10 | shuffled read_batch 50000 100 | shuffled read_batch 50000 1000 | shuffled read_formatted numpy 5000 | shuffled read_formatted_batch numpy 5000 10 | shuffled read_formatted_batch numpy 5000 1000 |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.628242 / 0.215209 (0.413033) | 6.348176 / 2.077655 (4.270522) | 2.615832 / 1.504120 (1.111712) | 2.217481 / 1.541195 (0.676286) | 2.287058 / 1.468490 (0.818568) | 1.322854 / 4.584777 (-3.261923) | 5.547831 / 3.745712 (1.802119) | 3.199467 / 5.269862 (-2.070395) | 2.135297 / 4.565676 (-2.430380) | 0.165134 / 0.424275 (-0.259141) | 0.014753 / 0.007607 (0.007146) | 0.778579 / 0.226044 (0.552535) | 7.982329 / 2.268929 (5.713401) | 3.331712 / 55.444624 (-52.112913) | 2.642606 / 6.876477 (-4.233871) | 2.699362 / 2.142072 (0.557290) | 1.572268 / 4.805227 (-3.232959) | 0.273348 / 6.500664 (-6.227316) | 0.082975 / 0.075469 (0.007506) |\n\n### Benchmark: benchmark_map_filter.json\n\n| metric | filter | map fast-tokenizer batched | map identity | map identity batched | map no-op batched | map no-op batched numpy | map no-op batched pandas | map no-op batched pytorch | map no-op batched tensorflow |\n|--------|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 1.730421 / 1.841788 (-0.111367) | 18.154495 / 8.074308 (10.080187) | 20.969885 / 10.191392 (10.778493) | 0.233652 / 0.680424 (-0.446772) | 0.026609 / 0.534201 (-0.507592) | 0.546874 / 0.579283 (-0.032410) | 0.602891 / 0.434364 (0.168527) | 0.641073 / 0.540337 (0.100736) | 0.772138 / 1.386936 (-0.614798) |\n\n</details>\n</details>\n\n![](https://cml.dev/watermark.png#20703458e3c42ee7bfc1a26e47805c0db4dda2d6 \"CML watermark\")\n" ]

[]

[]

[ "The docs for this PR live [here](https://moon-ci-docs.huggingface.co/docs/datasets/pr_5528). All of your documentation changes will be reflected on that endpoint.", "It seems that the parameter `create_pr` is available for [`0.8.0`](https://huggingface.co/docs/huggingface_hub/v0.8.1/en/package_reference/hf_api#huggingface_hub.HfApi.upload_file) (its not here: [`0.7.0`](https://huggingface.co/docs/huggingface_hub/v0.7.0.rc0/en/package_reference/hf_api#huggingface_hub.HfApi.upload_file)) and onwards. I included a warning, informing the user that no PR was created.", "@nateraw you are completely right! Actually, the dataset shards is never added to the created pr, only the metadata, as the code is now. Ill look into you suggestion asap. Thank!", "@nateraw Nothing more to add, that's a perfect usage of `huggingface_hub` as far as I can tell ! :fire: \r\n\r\nA very nit improvement would be to use the [for .. else ... python statement](https://book.pythontips.com/en/latest/for_-_else.html).\r\ni.e:\r\n\r\n```py\r\nif create_pr is True and revision is not None:\r\n for discussion in get_repo_discussions(repo_id, repo_type='dataset'):\r\n if discussion.is_pull_request and discussion.git_reference == revision:\r\n create_pr = False\r\n break\r\n else:\r\n raise ValueError(\"Provided revision not found\")\r\n```\r\nNo need for the `revision_found` temporary flag when do so. Yeah ok, it's niche :wink: ", "I added the suggestions from @nateraw and @Wauplin .", "> Thanks. Some comments/suggestions below...\r\n> \r\n> Why have you removed the test for create_pr? You could add it again and just add a pytest skipif when version of huggingface_hub is lower than 0.8.1.\r\n\r\nI have added the test again. I removed it because i kept getting errors when calling `create_pull_request` with `repo_id=ds_name` where `temporary_repo = ds_name`, and thought i might look more thoroughly at it later. I have added a test called `test_test` showing this, it gives:\r\n```\r\ntests/test_upstream_hub.py:360: \r\n_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _\r\n.venv/lib/python3.10/site-packages/huggingface_hub/utils/_validators.py:124: in _inner_fn\r\n return fn(*args, **kwargs)\r\n.venv/lib/python3.10/site-packages/huggingface_hub/hf_api.py:3451: in create_pull_request\r\n return self.create_discussion(\r\n.venv/lib/python3.10/site-packages/huggingface_hub/utils/_validators.py:124: in _inner_fn\r\n return fn(*args, **kwargs)\r\n.venv/lib/python3.10/site-packages/huggingface_hub/hf_api.py:3393: in create_discussion\r\n hf_raise_for_status(resp)\r\n(...)\r\nE huggingface_hub.utils._errors.RepositoryNotFoundError: 401 Client Error. (Request ID: Root=1-63ecd2cb-2cf2557a332c86ad27f687b3)\r\nE \r\nE Repository Not Found for url: https://huggingface.co/api/models/__DUMMY_TRANSFORMERS_USER__/test-16764648321590/discussions.\r\nE Please make sure you specified the correct `repo_id` and `repo_type`.\r\nE If you are trying to access a private or gated repo, make sure you are authenticated.\r\nE Invalid username or password.\r\n```", "> > Thanks. Some comments/suggestions below...\r\n> > Why have you removed the test for create_pr? You could add it again and just add a pytest skipif when version of huggingface_hub is lower than 0.8.1.\r\n> \r\n> I have added the test again. I removed it because i kept getting errors when calling `create_pull_request` with `repo_id=ds_name` where `temporary_repo = ds_name`, and thought i might look more thoroughly at it later. I have added a test called `test_test` showing this, it gives:\r\n> \r\n> ```\r\n> tests/test_upstream_hub.py:360: \r\n> _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _\r\n> .venv/lib/python3.10/site-packages/huggingface_hub/utils/_validators.py:124: in _inner_fn\r\n> return fn(*args, **kwargs)\r\n> .venv/lib/python3.10/site-packages/huggingface_hub/hf_api.py:3451: in create_pull_request\r\n> return self.create_discussion(\r\n> .venv/lib/python3.10/site-packages/huggingface_hub/utils/_validators.py:124: in _inner_fn\r\n> return fn(*args, **kwargs)\r\n> .venv/lib/python3.10/site-packages/huggingface_hub/hf_api.py:3393: in create_discussion\r\n> hf_raise_for_status(resp)\r\n> (...)\r\n> E huggingface_hub.utils._errors.RepositoryNotFoundError: 401 Client Error. (Request ID: Root=1-63ecd2cb-2cf2557a332c86ad27f687b3)\r\n> E \r\n> E Repository Not Found for url: https://huggingface.co/api/models/__DUMMY_TRANSFORMERS_USER__/test-16764648321590/discussions.\r\n> E Please make sure you specified the correct `repo_id` and `repo_type`.\r\n> E If you are trying to access a private or gated repo, make sure you are authenticated.\r\n> E Invalid username or password.\r\n> ```\r\n\r\n@albertvillanova, @lhoestq , FYI I have looked at this again, and i haven't figured it out, so the test`test_push_dataset_to_hub_with_pull_request` and the minimal example `test_test` are still failing locally, while the other tests succeed. Do you have any advice?", "I tried to move all of the \"create pr safely\"-logic to a seperate function in `_hf_hub_fixes`. I looked at how the exceptions were raised before `huggingface_hub.utils.RepositoryNotFoundError`existed, and make changes accordingly. ", "`create_pr` was set during `push_to_hub`, even though it was `None` from the outset, hence causing tests to fail for older versions of `huggingface_hub`. This is now fixed.\r\n\r\nWith the implementation of `_hf_hub_fixes.upload_file` the function call expected `commit_message`, `commit_description`. If these are not set we call the function without them, even though we are on a version of `huggingface_hub` where they are not available in `upload_file`.\r\n\r\nWhen `huggingface_hub < 0.5.0` we assume `repo_id` of them form `organisation/name`, so now that we are calling `create_repo` in the tests with `repo_id` not of this form, we need to handle this case, this is now done.\r\n\r\nMany tests failed for `dataset_dict` for the above reasons, so the fixes from `arrow_dataset.py` were also added to `dataset_dict.py`. \r\n\r\n**All tests are now passing locally for `huggingface_hub==0.2.0` and `huggingface_hub==0.12.1`…** Im sorry I should have downgraded and went through this a long time ago, but I didn’t realise the extend of these version fixes until recently…", "Hi ! FYI bumped the `huggingface-hub` dependency to 0.11 and removed the `_hf_hub_fixes.py` - which should make this PR much easier" ]

[]

[]

[ "_The documentation is not available anymore as the PR was closed or merged._", "<details>\n<summary>Show benchmarks</summary>\n\nPyArrow==6.0.0\n\n<details>\n<summary>Show updated benchmarks!</summary>\n\n### Benchmark: benchmark_array_xd.json\n\n| metric | read_batch_formatted_as_numpy after write_array2d | read_batch_formatted_as_numpy after write_flattened_sequence | read_batch_formatted_as_numpy after write_nested_sequence | read_batch_unformated after write_array2d | read_batch_unformated after write_flattened_sequence | read_batch_unformated after write_nested_sequence | read_col_formatted_as_numpy after write_array2d | read_col_formatted_as_numpy after write_flattened_sequence | read_col_formatted_as_numpy after write_nested_sequence | read_col_unformated after write_array2d | read_col_unformated after write_flattened_sequence | read_col_unformated after write_nested_sequence | read_formatted_as_numpy after write_array2d | read_formatted_as_numpy after write_flattened_sequence | read_formatted_as_numpy after write_nested_sequence | read_unformated after write_array2d | read_unformated after write_flattened_sequence | read_unformated after write_nested_sequence | write_array2d | write_flattened_sequence | write_nested_sequence |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.011142 / 0.011353 (-0.000211) | 0.005885 / 0.011008 (-0.005123) | 0.115374 / 0.038508 (0.076866) | 0.041704 / 0.023109 (0.018594) | 0.356996 / 0.275898 (0.081098) | 0.395076 / 0.323480 (0.071596) | 0.008726 / 0.007986 (0.000740) | 0.005528 / 0.004328 (0.001199) | 0.087817 / 0.004250 (0.083566) | 0.049273 / 0.037052 (0.012221) | 0.363778 / 0.258489 (0.105289) | 0.408801 / 0.293841 (0.114960) | 0.045232 / 0.128546 (-0.083314) | 0.013788 / 0.075646 (-0.061859) | 0.395634 / 0.419271 (-0.023637) | 0.056583 / 0.043533 (0.013051) | 0.360779 / 0.255139 (0.105640) | 0.386843 / 0.283200 (0.103643) | 0.116632 / 0.141683 (-0.025051) | 1.830020 / 1.452155 (0.377865) | 1.808720 / 1.492716 (0.316003) |\n\n### Benchmark: benchmark_getitem\\_100B.json\n\n| metric | get_batch_of\\_1024\\_random_rows | get_batch_of\\_1024\\_rows | get_first_row | get_last_row |\n|--------|---|---|---|---|\n| new / old (diff) | 0.221029 / 0.018006 (0.203023) | 0.489463 / 0.000490 (0.488973) | 0.002104 / 0.000200 (0.001904) | 0.000098 / 0.000054 (0.000043) |\n\n### Benchmark: benchmark_indices_mapping.json\n\n| metric | select | shard | shuffle | sort | train_test_split |\n|--------|---|---|---|---|---|\n| new / old (diff) | 0.032873 / 0.037411 (-0.004539) | 0.129526 / 0.014526 (0.115000) | 0.141446 / 0.176557 (-0.035111) | 0.189222 / 0.737135 (-0.547913) | 0.149329 / 0.296338 (-0.147010) |\n\n### Benchmark: benchmark_iterating.json\n\n| metric | read 5000 | read 50000 | read_batch 50000 10 | read_batch 50000 100 | read_batch 50000 1000 | read_formatted numpy 5000 | read_formatted pandas 5000 | read_formatted tensorflow 5000 | read_formatted torch 5000 | read_formatted_batch numpy 5000 10 | read_formatted_batch numpy 5000 1000 | shuffled read 5000 | shuffled read 50000 | shuffled read_batch 50000 10 | shuffled read_batch 50000 100 | shuffled read_batch 50000 1000 | shuffled read_formatted numpy 5000 | shuffled read_formatted_batch numpy 5000 10 | shuffled read_formatted_batch numpy 5000 1000 |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.471389 / 0.215209 (0.256180) | 4.710174 / 2.077655 (2.632519) | 2.239122 / 1.504120 (0.735002) | 1.977789 / 1.541195 (0.436595) | 2.107336 / 1.468490 (0.638846) | 0.816852 / 4.584777 (-3.767925) | 4.944056 / 3.745712 (1.198344) | 4.637939 / 5.269862 (-0.631922) | 2.355546 / 4.565676 (-2.210131) | 0.099324 / 0.424275 (-0.324951) | 0.014529 / 0.007607 (0.006922) | 0.596322 / 0.226044 (0.370277) | 5.972216 / 2.268929 (3.703287) | 2.697281 / 55.444624 (-52.747344) | 2.293836 / 6.876477 (-4.582641) | 2.380271 / 2.142072 (0.238199) | 1.001307 / 4.805227 (-3.803920) | 0.196981 / 6.500664 (-6.303683) | 0.074390 / 0.075469 (-0.001079) |\n\n### Benchmark: benchmark_map_filter.json\n\n| metric | filter | map fast-tokenizer batched | map identity | map identity batched | map no-op batched | map no-op batched numpy | map no-op batched pandas | map no-op batched pytorch | map no-op batched tensorflow |\n|--------|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 1.482915 / 1.841788 (-0.358872) | 18.739511 / 8.074308 (10.665202) | 16.768191 / 10.191392 (6.576799) | 0.203163 / 0.680424 (-0.477261) | 0.037514 / 0.534201 (-0.496687) | 0.529017 / 0.579283 (-0.050266) | 0.577591 / 0.434364 (0.143227) | 0.634057 / 0.540337 (0.093720) | 0.759812 / 1.386936 (-0.627124) |\n\n</details>\nPyArrow==latest\n\n<details>\n<summary>Show updated benchmarks!</summary>\n\n### Benchmark: benchmark_array_xd.json\n\n| metric | read_batch_formatted_as_numpy after write_array2d | read_batch_formatted_as_numpy after write_flattened_sequence | read_batch_formatted_as_numpy after write_nested_sequence | read_batch_unformated after write_array2d | read_batch_unformated after write_flattened_sequence | read_batch_unformated after write_nested_sequence | read_col_formatted_as_numpy after write_array2d | read_col_formatted_as_numpy after write_flattened_sequence | read_col_formatted_as_numpy after write_nested_sequence | read_col_unformated after write_array2d | read_col_unformated after write_flattened_sequence | read_col_unformated after write_nested_sequence | read_formatted_as_numpy after write_array2d | read_formatted_as_numpy after write_flattened_sequence | read_formatted_as_numpy after write_nested_sequence | read_unformated after write_array2d | read_unformated after write_flattened_sequence | read_unformated after write_nested_sequence | write_array2d | write_flattened_sequence | write_nested_sequence |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.008815 / 0.011353 (-0.002537) | 0.005956 / 0.011008 (-0.005052) | 0.087912 / 0.038508 (0.049404) | 0.040291 / 0.023109 (0.017182) | 0.404079 / 0.275898 (0.128181) | 0.447309 / 0.323480 (0.123829) | 0.006515 / 0.007986 (-0.001471) | 0.005917 / 0.004328 (0.001588) | 0.085560 / 0.004250 (0.081310) | 0.057077 / 0.037052 (0.020025) | 0.403349 / 0.258489 (0.144860) | 0.465644 / 0.293841 (0.171803) | 0.043530 / 0.128546 (-0.085016) | 0.014234 / 0.075646 (-0.061412) | 0.102203 / 0.419271 (-0.317068) | 0.058335 / 0.043533 (0.014802) | 0.398488 / 0.255139 (0.143349) | 0.424127 / 0.283200 (0.140927) | 0.119058 / 0.141683 (-0.022625) | 1.748748 / 1.452155 (0.296593) | 1.822190 / 1.492716 (0.329474) |\n\n### Benchmark: benchmark_getitem\\_100B.json\n\n| metric | get_batch_of\\_1024\\_random_rows | get_batch_of\\_1024\\_rows | get_first_row | get_last_row |\n|--------|---|---|---|---|\n| new / old (diff) | 0.255782 / 0.018006 (0.237776) | 0.496665 / 0.000490 (0.496176) | 0.000471 / 0.000200 (0.000271) | 0.000069 / 0.000054 (0.000014) |\n\n### Benchmark: benchmark_indices_mapping.json\n\n| metric | select | shard | shuffle | sort | train_test_split |\n|--------|---|---|---|---|---|\n| new / old (diff) | 0.034111 / 0.037411 (-0.003301) | 0.131442 / 0.014526 (0.116917) | 0.144660 / 0.176557 (-0.031897) | 0.188156 / 0.737135 (-0.548979) | 0.149875 / 0.296338 (-0.146463) |\n\n### Benchmark: benchmark_iterating.json\n\n| metric | read 5000 | read 50000 | read_batch 50000 10 | read_batch 50000 100 | read_batch 50000 1000 | read_formatted numpy 5000 | read_formatted pandas 5000 | read_formatted tensorflow 5000 | read_formatted torch 5000 | read_formatted_batch numpy 5000 10 | read_formatted_batch numpy 5000 1000 | shuffled read 5000 | shuffled read 50000 | shuffled read_batch 50000 10 | shuffled read_batch 50000 100 | shuffled read_batch 50000 1000 | shuffled read_formatted numpy 5000 | shuffled read_formatted_batch numpy 5000 10 | shuffled read_formatted_batch numpy 5000 1000 |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.502218 / 0.215209 (0.287009) | 5.004486 / 2.077655 (2.926832) | 2.420379 / 1.504120 (0.916259) | 2.194671 / 1.541195 (0.653476) | 2.306376 / 1.468490 (0.837886) | 0.856623 / 4.584777 (-3.728154) | 4.963211 / 3.745712 (1.217499) | 2.517965 / 5.269862 (-2.751896) | 1.743880 / 4.565676 (-2.821797) | 0.105270 / 0.424275 (-0.319005) | 0.014725 / 0.007607 (0.007118) | 0.621934 / 0.226044 (0.395890) | 6.183827 / 2.268929 (3.914898) | 2.945868 / 55.444624 (-52.498757) | 2.557676 / 6.876477 (-4.318801) | 2.622282 / 2.142072 (0.480210) | 1.011647 / 4.805227 (-3.793580) | 0.199573 / 6.500664 (-6.301091) | 0.076283 / 0.075469 (0.000814) |\n\n### Benchmark: benchmark_map_filter.json\n\n| metric | filter | map fast-tokenizer batched | map identity | map identity batched | map no-op batched | map no-op batched numpy | map no-op batched pandas | map no-op batched pytorch | map no-op batched tensorflow |\n|--------|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 1.518813 / 1.841788 (-0.322975) | 18.833017 / 8.074308 (10.758709) | 16.095249 / 10.191392 (5.903857) | 0.196667 / 0.680424 (-0.483757) | 0.022060 / 0.534201 (-0.512141) | 0.537802 / 0.579283 (-0.041481) | 0.523676 / 0.434364 (0.089312) | 0.629387 / 0.540337 (0.089049) | 0.738042 / 1.386936 (-0.648894) |\n\n</details>\n</details>\n\n![](https://cml.dev/watermark.png#813712c3cd133f72f496d279e02344d6ee743fdf \"CML watermark\")\n", "<details>\n<summary>Show benchmarks</summary>\n\nPyArrow==6.0.0\n\n<details>\n<summary>Show updated benchmarks!</summary>\n\n### Benchmark: benchmark_array_xd.json\n\n| metric | read_batch_formatted_as_numpy after write_array2d | read_batch_formatted_as_numpy after write_flattened_sequence | read_batch_formatted_as_numpy after write_nested_sequence | read_batch_unformated after write_array2d | read_batch_unformated after write_flattened_sequence | read_batch_unformated after write_nested_sequence | read_col_formatted_as_numpy after write_array2d | read_col_formatted_as_numpy after write_flattened_sequence | read_col_formatted_as_numpy after write_nested_sequence | read_col_unformated after write_array2d | read_col_unformated after write_flattened_sequence | read_col_unformated after write_nested_sequence | read_formatted_as_numpy after write_array2d | read_formatted_as_numpy after write_flattened_sequence | read_formatted_as_numpy after write_nested_sequence | read_unformated after write_array2d | read_unformated after write_flattened_sequence | read_unformated after write_nested_sequence | write_array2d | write_flattened_sequence | write_nested_sequence |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.008608 / 0.011353 (-0.002745) | 0.004553 / 0.011008 (-0.006455) | 0.100031 / 0.038508 (0.061523) | 0.029498 / 0.023109 (0.006389) | 0.306913 / 0.275898 (0.031015) | 0.367369 / 0.323480 (0.043889) | 0.006883 / 0.007986 (-0.001103) | 0.004768 / 0.004328 (0.000440) | 0.077424 / 0.004250 (0.073173) | 0.034005 / 0.037052 (-0.003047) | 0.317772 / 0.258489 (0.059283) | 0.356859 / 0.293841 (0.063018) | 0.033717 / 0.128546 (-0.094829) | 0.011386 / 0.075646 (-0.064260) | 0.322832 / 0.419271 (-0.096439) | 0.043930 / 0.043533 (0.000397) | 0.308087 / 0.255139 (0.052948) | 0.338349 / 0.283200 (0.055149) | 0.094780 / 0.141683 (-0.046903) | 1.463454 / 1.452155 (0.011300) | 1.495055 / 1.492716 (0.002338) |\n\n### Benchmark: benchmark_getitem\\_100B.json\n\n| metric | get_batch_of\\_1024\\_random_rows | get_batch_of\\_1024\\_rows | get_first_row | get_last_row |\n|--------|---|---|---|---|\n| new / old (diff) | 0.191039 / 0.018006 (0.173033) | 0.414650 / 0.000490 (0.414160) | 0.002435 / 0.000200 (0.002235) | 0.000075 / 0.000054 (0.000021) |\n\n### Benchmark: benchmark_indices_mapping.json\n\n| metric | select | shard | shuffle | sort | train_test_split |\n|--------|---|---|---|---|---|\n| new / old (diff) | 0.023871 / 0.037411 (-0.013540) | 0.097140 / 0.014526 (0.082614) | 0.105914 / 0.176557 (-0.070643) | 0.147375 / 0.737135 (-0.589760) | 0.107985 / 0.296338 (-0.188354) |\n\n### Benchmark: benchmark_iterating.json\n\n| metric | read 5000 | read 50000 | read_batch 50000 10 | read_batch 50000 100 | read_batch 50000 1000 | read_formatted numpy 5000 | read_formatted pandas 5000 | read_formatted tensorflow 5000 | read_formatted torch 5000 | read_formatted_batch numpy 5000 10 | read_formatted_batch numpy 5000 1000 | shuffled read 5000 | shuffled read 50000 | shuffled read_batch 50000 10 | shuffled read_batch 50000 100 | shuffled read_batch 50000 1000 | shuffled read_formatted numpy 5000 | shuffled read_formatted_batch numpy 5000 10 | shuffled read_formatted_batch numpy 5000 1000 |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.420174 / 0.215209 (0.204965) | 4.208354 / 2.077655 (2.130700) | 1.904568 / 1.504120 (0.400448) | 1.687406 / 1.541195 (0.146212) | 1.723901 / 1.468490 (0.255411) | 0.693554 / 4.584777 (-3.891223) | 3.445474 / 3.745712 (-0.300238) | 1.904919 / 5.269862 (-3.364943) | 1.284378 / 4.565676 (-3.281298) | 0.082539 / 0.424275 (-0.341736) | 0.012490 / 0.007607 (0.004883) | 0.527778 / 0.226044 (0.301733) | 5.300766 / 2.268929 (3.031838) | 2.324666 / 55.444624 (-53.119958) | 1.977166 / 6.876477 (-4.899311) | 2.054396 / 2.142072 (-0.087677) | 0.820966 / 4.805227 (-3.984261) | 0.148584 / 6.500664 (-6.352080) | 0.063618 / 0.075469 (-0.011851) |\n\n### Benchmark: benchmark_map_filter.json\n\n| metric | filter | map fast-tokenizer batched | map identity | map identity batched | map no-op batched | map no-op batched numpy | map no-op batched pandas | map no-op batched pytorch | map no-op batched tensorflow |\n|--------|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 1.188075 / 1.841788 (-0.653712) | 13.706950 / 8.074308 (5.632642) | 13.725122 / 10.191392 (3.533730) | 0.167379 / 0.680424 (-0.513045) | 0.028729 / 0.534201 (-0.505472) | 0.395373 / 0.579283 (-0.183910) | 0.403604 / 0.434364 (-0.030760) | 0.464290 / 0.540337 (-0.076047) | 0.553792 / 1.386936 (-0.833144) |\n\n</details>\nPyArrow==latest\n\n<details>\n<summary>Show updated benchmarks!</summary>\n\n### Benchmark: benchmark_array_xd.json\n\n| metric | read_batch_formatted_as_numpy after write_array2d | read_batch_formatted_as_numpy after write_flattened_sequence | read_batch_formatted_as_numpy after write_nested_sequence | read_batch_unformated after write_array2d | read_batch_unformated after write_flattened_sequence | read_batch_unformated after write_nested_sequence | read_col_formatted_as_numpy after write_array2d | read_col_formatted_as_numpy after write_flattened_sequence | read_col_formatted_as_numpy after write_nested_sequence | read_col_unformated after write_array2d | read_col_unformated after write_flattened_sequence | read_col_unformated after write_nested_sequence | read_formatted_as_numpy after write_array2d | read_formatted_as_numpy after write_flattened_sequence | read_formatted_as_numpy after write_nested_sequence | read_unformated after write_array2d | read_unformated after write_flattened_sequence | read_unformated after write_nested_sequence | write_array2d | write_flattened_sequence | write_nested_sequence |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.006565 / 0.011353 (-0.004787) | 0.004588 / 0.011008 (-0.006420) | 0.077312 / 0.038508 (0.038804) | 0.027348 / 0.023109 (0.004239) | 0.367753 / 0.275898 (0.091855) | 0.403250 / 0.323480 (0.079770) | 0.005201 / 0.007986 (-0.002785) | 0.004695 / 0.004328 (0.000366) | 0.076203 / 0.004250 (0.071953) | 0.039388 / 0.037052 (0.002336) | 0.374418 / 0.258489 (0.115929) | 0.413623 / 0.293841 (0.119782) | 0.031731 / 0.128546 (-0.096815) | 0.011644 / 0.075646 (-0.064002) | 0.086339 / 0.419271 (-0.332932) | 0.048902 / 0.043533 (0.005369) | 0.352064 / 0.255139 (0.096925) | 0.386637 / 0.283200 (0.103437) | 0.093662 / 0.141683 (-0.048021) | 1.479863 / 1.452155 (0.027709) | 1.562475 / 1.492716 (0.069758) |\n\n### Benchmark: benchmark_getitem\\_100B.json\n\n| metric | get_batch_of\\_1024\\_random_rows | get_batch_of\\_1024\\_rows | get_first_row | get_last_row |\n|--------|---|---|---|---|\n| new / old (diff) | 0.231874 / 0.018006 (0.213867) | 0.402185 / 0.000490 (0.401695) | 0.005252 / 0.000200 (0.005052) | 0.000086 / 0.000054 (0.000032) |\n\n### Benchmark: benchmark_indices_mapping.json\n\n| metric | select | shard | shuffle | sort | train_test_split |\n|--------|---|---|---|---|---|\n| new / old (diff) | 0.025402 / 0.037411 (-0.012010) | 0.099896 / 0.014526 (0.085370) | 0.106365 / 0.176557 (-0.070192) | 0.143309 / 0.737135 (-0.593827) | 0.112311 / 0.296338 (-0.184027) |\n\n### Benchmark: benchmark_iterating.json\n\n| metric | read 5000 | read 50000 | read_batch 50000 10 | read_batch 50000 100 | read_batch 50000 1000 | read_formatted numpy 5000 | read_formatted pandas 5000 | read_formatted tensorflow 5000 | read_formatted torch 5000 | read_formatted_batch numpy 5000 10 | read_formatted_batch numpy 5000 1000 | shuffled read 5000 | shuffled read 50000 | shuffled read_batch 50000 10 | shuffled read_batch 50000 100 | shuffled read_batch 50000 1000 | shuffled read_formatted numpy 5000 | shuffled read_formatted_batch numpy 5000 10 | shuffled read_formatted_batch numpy 5000 1000 |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.447637 / 0.215209 (0.232428) | 4.469337 / 2.077655 (2.391682) | 2.164332 / 1.504120 (0.660212) | 1.957826 / 1.541195 (0.416631) | 1.984580 / 1.468490 (0.516090) | 0.702909 / 4.584777 (-3.881868) | 3.361725 / 3.745712 (-0.383987) | 2.818102 / 5.269862 (-2.451760) | 1.589815 / 4.565676 (-2.975862) | 0.083647 / 0.424275 (-0.340628) | 0.012502 / 0.007607 (0.004895) | 0.545578 / 0.226044 (0.319534) | 5.480894 / 2.268929 (3.211966) | 2.605599 / 55.444624 (-52.839026) | 2.253444 / 6.876477 (-4.623032) | 2.289818 / 2.142072 (0.147746) | 0.803680 / 4.805227 (-4.001547) | 0.151870 / 6.500664 (-6.348794) | 0.066610 / 0.075469 (-0.008859) |\n\n### Benchmark: benchmark_map_filter.json\n\n| metric | filter | map fast-tokenizer batched | map identity | map identity batched | map no-op batched | map no-op batched numpy | map no-op batched pandas | map no-op batched pytorch | map no-op batched tensorflow |\n|--------|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 1.327390 / 1.841788 (-0.514398) | 14.046936 / 8.074308 (5.972628) | 13.643169 / 10.191392 (3.451777) | 0.128223 / 0.680424 (-0.552201) | 0.016941 / 0.534201 (-0.517260) | 0.383887 / 0.579283 (-0.195396) | 0.383891 / 0.434364 (-0.050473) | 0.440191 / 0.540337 (-0.100146) | 0.525357 / 1.386936 (-0.861579) |\n\n</details>\n</details>\n\n![](https://cml.dev/watermark.png#1575be339bc14a12229d782e2788746f27aeeb2a \"CML watermark\")\n", "Yea there must have been an update in another package that unconstrained the protobuf dependency - idk which one though", "It is `tensorboard`. I have reported the issue to `tensorflow`:\r\n- https://github.com/tensorflow/tensorflow/issues/59665" ]

[]

[]

[ "_The documentation is not available anymore as the PR was closed or merged._", "Thanks for the quick review! I updated the code with your suggestion", "Thanks for the quick review @albertvillanova! I updated the code with your suggestions", "<details>\n<summary>Show benchmarks</summary>\n\nPyArrow==8.0.0\n\n<details>\n<summary>Show updated benchmarks!</summary>\n\n### Benchmark: benchmark_array_xd.json\n\n| metric | read_batch_formatted_as_numpy after write_array2d | read_batch_formatted_as_numpy after write_flattened_sequence | read_batch_formatted_as_numpy after write_nested_sequence | read_batch_unformated after write_array2d | read_batch_unformated after write_flattened_sequence | read_batch_unformated after write_nested_sequence | read_col_formatted_as_numpy after write_array2d | read_col_formatted_as_numpy after write_flattened_sequence | read_col_formatted_as_numpy after write_nested_sequence | read_col_unformated after write_array2d | read_col_unformated after write_flattened_sequence | read_col_unformated after write_nested_sequence | read_formatted_as_numpy after write_array2d | read_formatted_as_numpy after write_flattened_sequence | read_formatted_as_numpy after write_nested_sequence | read_unformated after write_array2d | read_unformated after write_flattened_sequence | read_unformated after write_nested_sequence | write_array2d | write_flattened_sequence | write_nested_sequence |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.008577 / 0.011353 (-0.002776) | 0.005714 / 0.011008 (-0.005294) | 0.114718 / 0.038508 (0.076210) | 0.039799 / 0.023109 (0.016690) | 0.387530 / 0.275898 (0.111632) | 0.395739 / 0.323480 (0.072259) | 0.006775 / 0.007986 (-0.001211) | 0.006280 / 0.004328 (0.001952) | 0.086470 / 0.004250 (0.082220) | 0.054424 / 0.037052 (0.017371) | 0.361989 / 0.258489 (0.103500) | 0.424678 / 0.293841 (0.130837) | 0.043081 / 0.128546 (-0.085465) | 0.013903 / 0.075646 (-0.061743) | 0.397625 / 0.419271 (-0.021647) | 0.059789 / 0.043533 (0.016256) | 0.375195 / 0.255139 (0.120056) | 0.403724 / 0.283200 (0.120524) | 0.121470 / 0.141683 (-0.020213) | 1.734496 / 1.452155 (0.282341) | 1.820479 / 1.492716 (0.327763) |\n\n### Benchmark: benchmark_getitem\\_100B.json\n\n| metric | get_batch_of\\_1024\\_random_rows | get_batch_of\\_1024\\_rows | get_first_row | get_last_row |\n|--------|---|---|---|---|\n| new / old (diff) | 0.239672 / 0.018006 (0.221665) | 0.499373 / 0.000490 (0.498883) | 0.005034 / 0.000200 (0.004834) | 0.000099 / 0.000054 (0.000045) |\n\n### Benchmark: benchmark_indices_mapping.json\n\n| metric | select | shard | shuffle | sort | train_test_split |\n|--------|---|---|---|---|---|\n| new / old (diff) | 0.033000 / 0.037411 (-0.004411) | 0.130930 / 0.014526 (0.116404) | 0.151690 / 0.176557 (-0.024866) | 0.211839 / 0.737135 (-0.525296) | 0.148727 / 0.296338 (-0.147612) |\n\n### Benchmark: benchmark_iterating.json\n\n| metric | read 5000 | read 50000 | read_batch 50000 10 | read_batch 50000 100 | read_batch 50000 1000 | read_formatted numpy 5000 | read_formatted pandas 5000 | read_formatted tensorflow 5000 | read_formatted torch 5000 | read_formatted_batch numpy 5000 10 | read_formatted_batch numpy 5000 1000 | shuffled read 5000 | shuffled read 50000 | shuffled read_batch 50000 10 | shuffled read_batch 50000 100 | shuffled read_batch 50000 1000 | shuffled read_formatted numpy 5000 | shuffled read_formatted_batch numpy 5000 10 | shuffled read_formatted_batch numpy 5000 1000 |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.480592 / 0.215209 (0.265382) | 4.809700 / 2.077655 (2.732046) | 2.232414 / 1.504120 (0.728294) | 2.035432 / 1.541195 (0.494237) | 2.115991 / 1.468490 (0.647501) | 0.817841 / 4.584777 (-3.766936) | 4.718035 / 3.745712 (0.972323) | 4.107102 / 5.269862 (-1.162759) | 2.166838 / 4.565676 (-2.398839) | 0.102207 / 0.424275 (-0.322068) | 0.014686 / 0.007607 (0.007079) | 0.599922 / 0.226044 (0.373877) | 5.985840 / 2.268929 (3.716912) | 2.769199 / 55.444624 (-52.675425) | 2.427095 / 6.876477 (-4.449382) | 2.586666 / 2.142072 (0.444593) | 0.987650 / 4.805227 (-3.817578) | 0.199419 / 6.500664 (-6.301245) | 0.076710 / 0.075469 (0.001240) |\n\n### Benchmark: benchmark_map_filter.json\n\n| metric | filter | map fast-tokenizer batched | map identity | map identity batched | map no-op batched | map no-op batched numpy | map no-op batched pandas | map no-op batched pytorch | map no-op batched tensorflow |\n|--------|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 1.454509 / 1.841788 (-0.387278) | 18.267849 / 8.074308 (10.193541) | 16.701880 / 10.191392 (6.510488) | 0.204225 / 0.680424 (-0.476199) | 0.020295 / 0.534201 (-0.513906) | 0.504254 / 0.579283 (-0.075029) | 0.535071 / 0.434364 (0.100707) | 0.611825 / 0.540337 (0.071488) | 0.697289 / 1.386936 (-0.689647) |\n\n</details>\nPyArrow==latest\n\n<details>\n<summary>Show updated benchmarks!</summary>\n\n### Benchmark: benchmark_array_xd.json\n\n| metric | read_batch_formatted_as_numpy after write_array2d | read_batch_formatted_as_numpy after write_flattened_sequence | read_batch_formatted_as_numpy after write_nested_sequence | read_batch_unformated after write_array2d | read_batch_unformated after write_flattened_sequence | read_batch_unformated after write_nested_sequence | read_col_formatted_as_numpy after write_array2d | read_col_formatted_as_numpy after write_flattened_sequence | read_col_formatted_as_numpy after write_nested_sequence | read_col_unformated after write_array2d | read_col_unformated after write_flattened_sequence | read_col_unformated after write_nested_sequence | read_formatted_as_numpy after write_array2d | read_formatted_as_numpy after write_flattened_sequence | read_formatted_as_numpy after write_nested_sequence | read_unformated after write_array2d | read_unformated after write_flattened_sequence | read_unformated after write_nested_sequence | write_array2d | write_flattened_sequence | write_nested_sequence |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.009141 / 0.011353 (-0.002211) | 0.005987 / 0.011008 (-0.005021) | 0.092003 / 0.038508 (0.053495) | 0.043239 / 0.023109 (0.020130) | 0.400425 / 0.275898 (0.124527) | 0.464849 / 0.323480 (0.141369) | 0.008256 / 0.007986 (0.000270) | 0.006251 / 0.004328 (0.001923) | 0.095263 / 0.004250 (0.091013) | 0.057899 / 0.037052 (0.020847) | 0.402899 / 0.258489 (0.144410) | 0.477411 / 0.293841 (0.183570) | 0.044122 / 0.128546 (-0.084424) | 0.014158 / 0.075646 (-0.061489) | 0.116354 / 0.419271 (-0.302917) | 0.061045 / 0.043533 (0.017512) | 0.411635 / 0.255139 (0.156497) | 0.466281 / 0.283200 (0.183082) | 0.129423 / 0.141683 (-0.012260) | 1.799790 / 1.452155 (0.347635) | 2.004578 / 1.492716 (0.511862) |\n\n### Benchmark: benchmark_getitem\\_100B.json\n\n| metric | get_batch_of\\_1024\\_random_rows | get_batch_of\\_1024\\_rows | get_first_row | get_last_row |\n|--------|---|---|---|---|\n| new / old (diff) | 0.224012 / 0.018006 (0.206006) | 0.502972 / 0.000490 (0.502482) | 0.003560 / 0.000200 (0.003360) | 0.000110 / 0.000054 (0.000056) |\n\n### Benchmark: benchmark_indices_mapping.json\n\n| metric | select | shard | shuffle | sort | train_test_split |\n|--------|---|---|---|---|---|\n| new / old (diff) | 0.034794 / 0.037411 (-0.002618) | 0.139646 / 0.014526 (0.125120) | 0.144330 / 0.176557 (-0.032226) | 0.202528 / 0.737135 (-0.534607) | 0.151561 / 0.296338 (-0.144777) |\n\n### Benchmark: benchmark_iterating.json\n\n| metric | read 5000 | read 50000 | read_batch 50000 10 | read_batch 50000 100 | read_batch 50000 1000 | read_formatted numpy 5000 | read_formatted pandas 5000 | read_formatted tensorflow 5000 | read_formatted torch 5000 | read_formatted_batch numpy 5000 10 | read_formatted_batch numpy 5000 1000 | shuffled read 5000 | shuffled read 50000 | shuffled read_batch 50000 10 | shuffled read_batch 50000 100 | shuffled read_batch 50000 1000 | shuffled read_formatted numpy 5000 | shuffled read_formatted_batch numpy 5000 10 | shuffled read_formatted_batch numpy 5000 1000 |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.504343 / 0.215209 (0.289133) | 5.050690 / 2.077655 (2.973035) | 2.433107 / 1.504120 (0.928987) | 2.197443 / 1.541195 (0.656248) | 2.331225 / 1.468490 (0.862734) | 0.834066 / 4.584777 (-3.750711) | 4.837648 / 3.745712 (1.091936) | 4.105672 / 5.269862 (-1.164189) | 2.281557 / 4.565676 (-2.284120) | 0.102257 / 0.424275 (-0.322018) | 0.014425 / 0.007607 (0.006818) | 0.629290 / 0.226044 (0.403245) | 6.251513 / 2.268929 (3.982585) | 2.959012 / 55.444624 (-52.485613) | 2.570031 / 6.876477 (-4.306446) | 2.657525 / 2.142072 (0.515453) | 1.002861 / 4.805227 (-3.802367) | 0.199326 / 6.500664 (-6.301338) | 0.078428 / 0.075469 (0.002958) |\n\n### Benchmark: benchmark_map_filter.json\n\n| metric | filter | map fast-tokenizer batched | map identity | map identity batched | map no-op batched | map no-op batched numpy | map no-op batched pandas | map no-op batched pytorch | map no-op batched tensorflow |\n|--------|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 1.579587 / 1.841788 (-0.262201) | 18.567509 / 8.074308 (10.493201) | 17.162144 / 10.191392 (6.970752) | 0.193460 / 0.680424 (-0.486964) | 0.020819 / 0.534201 (-0.513382) | 0.501929 / 0.579283 (-0.077354) | 0.508039 / 0.434364 (0.073675) | 0.582656 / 0.540337 (0.042319) | 0.693624 / 1.386936 (-0.693312) |\n\n</details>\n</details>\n\n![](https://cml.dev/watermark.png#c410d321cd1289c6a630192b078f4892c2e13ff9 \"CML watermark\")\n" ]

[]

[]

[ "Thanks for reporting, @TJ-Solergibert.\r\n\r\nWe cannot access your Colab notebook: `There was an error loading this notebook. Ensure that the file is accessible and try again.`\r\nCould you please make it publicly accessible?\r\n", "I swear it's public, I've checked the settings and I've been able to open it in incognito mode.\r\n\r\nNotebook: https://colab.research.google.com/drive/1JCrS7FlGfu_kFqChMrwKZ_bpabnIMqbP?usp=sharing\r\n\r\nAnyway, this is the code to reproduce the error:\r\n\r\n```python3\r\nfrom datasets import ClassLabel\r\nfrom datasets import load_dataset\r\n\r\neuroparl_ds = load_dataset(\"tj-solergibert/Europarl-ST\")\r\n\r\nsource_lang = \"nl\"\r\nlanguages = list(europarl_ds[\"train\"][0][\"transcriptions\"].keys())\r\nClassLabels = ClassLabel(num_classes = len(languages), names = languages)\r\n\r\ndef map_label2id(example):\r\n example['dest_lang'] = ClassLabels.str2int(example['dest_lang'])\r\n return example\r\n\r\ndef unfold_transcriptions(example):\r\n for lang in languages:\r\n example[lang] = example[\"transcriptions\"][lang]\r\n return example\r\n\r\ndef unroll(batch, src_lang, dest_langs):\r\n source_t, dest_t, dest_l = [], [], []\r\n for lang in dest_langs: \r\n source_t += batch[src_lang]\r\n dest_t += batch[lang]\r\n dest_l += [lang]\r\n return_dict = {\"source_text\": source_t, \"dest_text\": dest_t, \"dest_lang\": dest_l}\r\n return return_dict\r\n\r\ndef preprocess_split(ds_split, src_lang):\r\n dest_langs = [x for x in languages if x != src_lang]\r\n\r\n ds_split = ds_split.map(unroll, fn_kwargs= {\"src_lang\": src_lang, \"dest_langs\": dest_langs}, batched = True, batch_size = 1, remove_columns= list(languages))\r\n ds_split = ds_split.filter(lambda x: x[\"source_text\"] != None and x[\"dest_text\"] != None) # Remove incomplete translations\r\n ds_split = ds_split.filter(lambda x: x[\"source_text\"] != \"None\" and x[\"dest_text\"] != \"None\")\r\n ds_split = ds_split.map(map_label2id) \r\n ds_split = ds_split.cast_column(\"dest_lang\", ClassLabels)\r\n return ds_split\r\n\r\ndef reset_cortas(example):\r\n for lang in languages:\r\n if isinstance(example[lang], str):\r\n if example[lang].isnumeric () or len(example[lang]) <= 5:\r\n example[lang] = \"None\"\r\n return example\r\n\r\ndef clean_dataset(dataset):\r\n # Remove columns\r\n dataset = dataset.remove_columns([\"original_speech\", \"original_language\", \"audio_path\", \"segment_start\", \"segment_end\"])\r\n # Unfold\r\n dataset = dataset.map(unfold_transcriptions, remove_columns = [\"transcriptions\"])\r\n dataset = dataset.map(reset_cortas)\r\n return dataset\r\n\r\nprocessed_europarl = clean_dataset(europarl_ds[\"test\"])\r\nnew_train_ds = preprocess_split(processed_europarl, 'nl')\r\n```", "Thanks, @TJ-Solergibert. I can access your notebook now. Maybe it was just a temporary issue.\r\n\r\nAt first sight, it seems something related to your data: maybe some of the examples do not have all the transcriptions for all the languages. Then, some of them are null when unrolled. And when trying to concatenate with the other rows containing strings, the cast issue is raised (the arrays to be concatenated have different types).\r\n\r\nDo you think this could be the case?", "See, in this example, \"nl\" and \"ro\" transcripts are null:\r\n```python\r\n>>> europarl_ds[\"test\"][:1]\r\n{'original_speech': ['− Señor Presidente, en primer lugar, quisiera felicitar al señor Seeber por el trabajo realizado, porque en su informe se recogen muchas de las preocupaciones manifestadas en esta'],\r\n 'original_language': ['es'],\r\n 'audio_path': ['es/audios/en.20081008.24.3-238.m4a'],\r\n 'segment_start': [0.6200000047683716],\r\n 'segment_end': [11.319999694824219],\r\n 'transcriptions': [{'de': '− Herr Präsident! Zunächst möchte ich Richard Seeber zu der von ihm geleisteten Arbeit gratulieren, denn sein Bericht greift viele der in diesem Haus zum Ausdruck gebrachten Anliegen',\r\n 'en': '− Mr President, firstly I would like to congratulate Mr Seeber on the work he has done, because his report picks up many of the concerns expressed in this',\r\n 'es': '− Señor Presidente, en primer lugar, quisiera felicitar al señor Seeber por el trabajo realizado, porque en su informe se recogen muchas de las preocupaciones manifestadas en esta',\r\n 'fr': '− Monsieur le Président, je voudrais tout d ’ abord féliciter M. Seeber pour le travail qu ’ il a effectué, parce que son rapport reprend beaucoup des inquiétudes exprimées au sein de cette',\r\n 'it': \"− Signor Presidente, mi congratulo innanzi tutto con l'onorevole Seeber per il lavoro svolto, perché la sua relazione accoglie molti dei timori espressi da quest'Aula\",\r\n 'nl': None,\r\n 'pl': '− Panie przewodniczący! Po pierwsze chciałabym pogratulować panu posłowi Seeberowi wykonanej pracy, ponieważ jego sprawozdanie podejmuje szereg podnoszonych w tej Izbie',\r\n 'pt': '− Senhor Presidente, começo por felicitar o senhor deputado Seeber pelo trabalho que desenvolveu em torno deste relatório, que retoma muitas das preocupações expressas nesta',\r\n 'ro': None}]}\r\n```\r\n```python\r\n>>> processed_europarl[0]\r\n{'de': '− Herr Präsident! Zunächst möchte ich Richard Seeber zu der von ihm geleisteten Arbeit gratulieren, denn sein Bericht greift viele der in diesem Haus zum Ausdruck gebrachten Anliegen',\r\n 'en': '− Mr President, firstly I would like to congratulate Mr Seeber on the work he has done, because his report picks up many of the concerns expressed in this',\r\n 'es': '− Señor Presidente, en primer lugar, quisiera felicitar al señor Seeber por el trabajo realizado, porque en su informe se recogen muchas de las preocupaciones manifestadas en esta',\r\n 'fr': '− Monsieur le Président, je voudrais tout d ’ abord féliciter M. Seeber pour le travail qu ’ il a effectué, parce que son rapport reprend beaucoup des inquiétudes exprimées au sein de cette',\r\n 'it': \"− Signor Presidente, mi congratulo innanzi tutto con l'onorevole Seeber per il lavoro svolto, perché la sua relazione accoglie molti dei timori espressi da quest'Aula\",\r\n 'nl': None,\r\n 'pl': '− Panie przewodniczący! Po pierwsze chciałabym pogratulować panu posłowi Seeberowi wykonanej pracy, ponieważ jego sprawozdanie podejmuje szereg podnoszonych w tej Izbie',\r\n 'pt': '− Senhor Presidente, começo por felicitar o senhor deputado Seeber pelo trabalho que desenvolveu em torno deste relatório, que retoma muitas das preocupações expressas nesta',\r\n 'ro': None}\r\n```", "You can fix this issue by forcing the cast of None to str by hand:\r\n- If you replace this line:\r\n```python\r\nsource_t += batch[src_lang]\r\n```\r\n- With this line (because the batch size is 1):\r\n```python\r\nsource_t += [str(batch[src_lang][0])]\r\n```\r\n- Or with this line (if the batch size were larger than 1):\r\n```python\r\nsource_t += [str(text) for text in batch[src_lang]]\r\n```", "Problem solved! Thanks @albertvillanova, now I have even increased the batch size and it's crazy fast :rocket: !" ]

[]

[]

[ "_The documentation is not available anymore as the PR was closed or merged._" ]

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

[ { "avatar_url": "https://avatars.githubusercontent.com/u/16348744?v=4", "events_url": "https://api.github.com/users/polinaeterna/events{/privacy}", "followers_url": "https://api.github.com/users/polinaeterna/followers", "following_url": "https://api.github.com/users/polinaeterna/following{/other_user}", "gists_url": "https://api.github.com/users/polinaeterna/gists{/gist_id}", "gravatar_id": "", "html_url": "https://github.com/polinaeterna", "id": 16348744, "login": "polinaeterna", "node_id": "MDQ6VXNlcjE2MzQ4NzQ0", "organizations_url": "https://api.github.com/users/polinaeterna/orgs", "received_events_url": "https://api.github.com/users/polinaeterna/received_events", "repos_url": "https://api.github.com/users/polinaeterna/repos", "site_admin": false, "starred_url": "https://api.github.com/users/polinaeterna/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/polinaeterna/subscriptions", "type": "User", "url": "https://api.github.com/users/polinaeterna" } ]

[]

[]

[]

[ "P.S. For more context, I'm currently exploring the integration of 🤗`datasets` with JAX, so in case you need any help or want me to try something specific just let me know! (`jnp.asarray`/`jnp.array(..., copy=False)` still no zero-copy 😭)", "_The documentation is not available anymore as the PR was closed or merged._", "> Hi ! Thanks for improving this :)\r\n\r\nGlad to help, @lhoestq! Also, regarding the questions in the `## What's missing?` can I have your input? Thanks 🤗 ", "Whoops forgot to reply to these matters - sorry x)\r\n\r\nYea a JAX guide would be welcome in the documentation ! This can be done in a separate PR if you want :)\r\n\r\nPyarrow is always imported with `datasets`, so it doesn't really matter if it's under TYPE_CHECKING or not.\r\n\r\nRegarding the license : yes indeed it should be in every file, thanks for reporting.\r\n\r\nNo big preference between jnp.array and jnp.asarray, unless one offers better performance", "> Whoops forgot to reply to these matters - sorry x)\r\n> \r\n> Yea a JAX guide would be welcome in the documentation ! This can be done in a separate PR if you want :)\r\n> \r\n> Pyarrow is always imported with `datasets`, so it doesn't really matter if it's under TYPE_CHECKING or not.\r\n> \r\n> Regarding the license : yes indeed it should be in every file, thanks for reporting.\r\n> \r\n> No big preference between jnp.array and jnp.asarray, unless one offers better performance\r\n\r\nCool @lhoestq thanks for the input there!\r\n\r\n1. I can create a separate PR for JAX-format usage\r\n2. Regarding that, makes sense, we can just not put it there, unless it's more clear that in that file `pyarrow` is just required for typing?\r\n3. Do you want me to add the License? In this PR? In a separate one?\r\n4. Ideally `jnp.asarray` is similar to `np.asarray` which in the case of `numpy` tends to be more efficient as it does zero-copy when possible, while `np.array` has `copy=True` by default, anyway as I mentioned before (and as you already know) the copy from `numpy` to `jax` is not zero-copy, while the other way around (`jax` to `numpy`) it is", "Thanks, feel free to create separate PRs for the docs and the license.\r\n\r\nI guess you can move the `pyarrow` import back to where it was for consistency with the other files and we can merge this one ;)", "> Thanks, feel free to create separate PRs for the docs and the license.\r\n> \r\n> I guess you can move the `pyarrow` import back to where it was for consistency with the other files and we can merge this one ;)\r\n\r\nCool thanks I'll do that! 👍🏻 ", "Actually I just checked and there are still tens of thousands of users with jax 0.3.25 - so we need to support older versions as well. I guess it comes from `transformers` which doesn't support jax 0.4 (and doesn't want to until the jax team stops breaking the lib all the time).\r\n\r\nCould you make sure your changes work with older versions as well ? Sorry for not spotting this earlier.\r\nIf we have `\"jax>=0.2.8,!=0.3.2,<=0.4.3\"` that'b be nice, and we can update the latest supported release from time to time.\r\n\r\nIn the CI you can add `jax==0.2.8` for the `deps-minimum` job, and use `jax~=0.4.1` for the `deps-latest`.", "> Actually I just checked and there are still tens of thousands of users with jax 0.3.25 - so we need to support older versions as well. I guess it comes from `transformers` which doesn't support jax 0.4 (and doesn't want to until the jax team stops breaking the lib all the time).\r\n> \r\n> Could you make sure your changes work with older versions as well ? Sorry for not spotting this earlier. If we have `\"jax>=0.2.8,!=0.3.2,<=0.4.3\"` that'b be nice, and we can update the latest supported release from time to time.\r\n> \r\n> In the CI you can add `jax==0.2.8` for the `deps-minimum` job, and use `jax~=0.4.1` for the `deps-latest`.\r\n\r\nOk, didn't know that @lhoestq thanks for the detailed context! Sure, I'll update it and make sure it's also compatible with older versions.", "Oops forgot to add you as co-author of the last commit @lhoestq my bad 😞 ", "So it should be fixed right now @lhoestq! The thing is that `jax` doesn't provide support for Python 3.7 due to its EOL next June (more information at https://endoflife.date/python)...\r\n\r\nAnyway, I can confirm that `jax.Array` type works with 0.3.25 and that the following code works fine:\r\n\r\n```python\r\nimport jax\r\nimport jax.numpy as jnp\r\n\r\nx = jnp.ones((1, 10), dtype=jnp.float32) # Is a `jnp.DeviceArray`\r\nassert isinstance(x, jax.Array) # Is `True`\r\n```\r\n\r\nSo we can still use 0.3.25 as the maximum supported version, as well as 0.3.6 for `jaxlib` so as to be consistent with 🤗`transformers`.\r\n\r\nThanks for your comments @lhoestq those were really useful!", "Sorry for the spam, pinning versions leads to failure runs (not related to the type-hinting); I'll check that locally instead of here to avoid spam... Not pinning the dependencies work but I'll check the minimum required versions for both `jax` and `jaxlib` in Python 3.7", "> Cool ! Thanks for trying to make the CI support it, but it's maybe not worth spending more time on this for now ^^\r\n> \r\n> merging :)\r\n\r\nDo you want me to work on the CI in a separate branch? Thanks for merging and for your help as always :)", "> Do you want me to work on the CI in a separate branch? Thanks for merging and for your help as always :)\r\n\r\nIn the end I think we can keep it as is since we didn't modify the core code for jax. Maybe later if we do further changes and need to make sure we don't break anything ;) For example when we decide to add support for more recent versions", "<details>\n<summary>Show benchmarks</summary>\n\nPyArrow==6.0.0\n\n<details>\n<summary>Show updated benchmarks!</summary>\n\n### Benchmark: benchmark_array_xd.json\n\n| metric | read_batch_formatted_as_numpy after write_array2d | read_batch_formatted_as_numpy after write_flattened_sequence | read_batch_formatted_as_numpy after write_nested_sequence | read_batch_unformated after write_array2d | read_batch_unformated after write_flattened_sequence | read_batch_unformated after write_nested_sequence | read_col_formatted_as_numpy after write_array2d | read_col_formatted_as_numpy after write_flattened_sequence | read_col_formatted_as_numpy after write_nested_sequence | read_col_unformated after write_array2d | read_col_unformated after write_flattened_sequence | read_col_unformated after write_nested_sequence | read_formatted_as_numpy after write_array2d | read_formatted_as_numpy after write_flattened_sequence | read_formatted_as_numpy after write_nested_sequence | read_unformated after write_array2d | read_unformated after write_flattened_sequence | read_unformated after write_nested_sequence | write_array2d | write_flattened_sequence | write_nested_sequence |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.010798 / 0.011353 (-0.000555) | 0.005690 / 0.011008 (-0.005318) | 0.116840 / 0.038508 (0.078332) | 0.041376 / 0.023109 (0.018266) | 0.345616 / 0.275898 (0.069718) | 0.413914 / 0.323480 (0.090434) | 0.009237 / 0.007986 (0.001252) | 0.004490 / 0.004328 (0.000162) | 0.085833 / 0.004250 (0.081582) | 0.050231 / 0.037052 (0.013179) | 0.367276 / 0.258489 (0.108787) | 0.393735 / 0.293841 (0.099894) | 0.043775 / 0.128546 (-0.084772) | 0.013215 / 0.075646 (-0.062432) | 0.391020 / 0.419271 (-0.028252) | 0.055102 / 0.043533 (0.011569) | 0.360333 / 0.255139 (0.105194) | 0.370531 / 0.283200 (0.087331) | 0.115484 / 0.141683 (-0.026199) | 1.694779 / 1.452155 (0.242625) | 1.756249 / 1.492716 (0.263532) |\n\n### Benchmark: benchmark_getitem\\_100B.json\n\n| metric | get_batch_of\\_1024\\_random_rows | get_batch_of\\_1024\\_rows | get_first_row | get_last_row |\n|--------|---|---|---|---|\n| new / old (diff) | 0.230508 / 0.018006 (0.212501) | 0.478681 / 0.000490 (0.478191) | 0.010305 / 0.000200 (0.010105) | 0.000147 / 0.000054 (0.000093) |\n\n### Benchmark: benchmark_indices_mapping.json\n\n| metric | select | shard | shuffle | sort | train_test_split |\n|--------|---|---|---|---|---|\n| new / old (diff) | 0.030953 / 0.037411 (-0.006459) | 0.124320 / 0.014526 (0.109794) | 0.140417 / 0.176557 (-0.036140) | 0.189522 / 0.737135 (-0.547613) | 0.143635 / 0.296338 (-0.152704) |\n\n### Benchmark: benchmark_iterating.json\n\n| metric | read 5000 | read 50000 | read_batch 50000 10 | read_batch 50000 100 | read_batch 50000 1000 | read_formatted numpy 5000 | read_formatted pandas 5000 | read_formatted tensorflow 5000 | read_formatted torch 5000 | read_formatted_batch numpy 5000 10 | read_formatted_batch numpy 5000 1000 | shuffled read 5000 | shuffled read 50000 | shuffled read_batch 50000 10 | shuffled read_batch 50000 100 | shuffled read_batch 50000 1000 | shuffled read_formatted numpy 5000 | shuffled read_formatted_batch numpy 5000 10 | shuffled read_formatted_batch numpy 5000 1000 |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.485995 / 0.215209 (0.270786) | 4.799668 / 2.077655 (2.722014) | 2.195655 / 1.504120 (0.691535) | 1.940073 / 1.541195 (0.398879) | 2.053853 / 1.468490 (0.585363) | 0.825399 / 4.584777 (-3.759378) | 4.522180 / 3.745712 (0.776468) | 2.484626 / 5.269862 (-2.785236) | 1.727617 / 4.565676 (-2.838059) | 0.098808 / 0.424275 (-0.325467) | 0.014753 / 0.007607 (0.007146) | 0.606798 / 0.226044 (0.380754) | 5.918090 / 2.268929 (3.649162) | 2.668124 / 55.444624 (-52.776500) | 2.300447 / 6.876477 (-4.576030) | 2.411203 / 2.142072 (0.269130) | 0.999826 / 4.805227 (-3.805401) | 0.193683 / 6.500664 (-6.306981) | 0.069341 / 0.075469 (-0.006129) |\n\n### Benchmark: benchmark_map_filter.json\n\n| metric | filter | map fast-tokenizer batched | map identity | map identity batched | map no-op batched | map no-op batched numpy | map no-op batched pandas | map no-op batched pytorch | map no-op batched tensorflow |\n|--------|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 1.455816 / 1.841788 (-0.385972) | 17.176476 / 8.074308 (9.102168) | 16.359100 / 10.191392 (6.167708) | 0.199669 / 0.680424 (-0.480755) | 0.033456 / 0.534201 (-0.500745) | 0.512478 / 0.579283 (-0.066805) | 0.526350 / 0.434364 (0.091986) | 0.637669 / 0.540337 (0.097332) | 0.753821 / 1.386936 (-0.633115) |\n\n</details>\nPyArrow==latest\n\n<details>\n<summary>Show updated benchmarks!</summary>\n\n### Benchmark: benchmark_array_xd.json\n\n| metric | read_batch_formatted_as_numpy after write_array2d | read_batch_formatted_as_numpy after write_flattened_sequence | read_batch_formatted_as_numpy after write_nested_sequence | read_batch_unformated after write_array2d | read_batch_unformated after write_flattened_sequence | read_batch_unformated after write_nested_sequence | read_col_formatted_as_numpy after write_array2d | read_col_formatted_as_numpy after write_flattened_sequence | read_col_formatted_as_numpy after write_nested_sequence | read_col_unformated after write_array2d | read_col_unformated after write_flattened_sequence | read_col_unformated after write_nested_sequence | read_formatted_as_numpy after write_array2d | read_formatted_as_numpy after write_flattened_sequence | read_formatted_as_numpy after write_nested_sequence | read_unformated after write_array2d | read_unformated after write_flattened_sequence | read_unformated after write_nested_sequence | write_array2d | write_flattened_sequence | write_nested_sequence |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.008176 / 0.011353 (-0.003177) | 0.005862 / 0.011008 (-0.005147) | 0.086123 / 0.038508 (0.047615) | 0.037144 / 0.023109 (0.014035) | 0.398328 / 0.275898 (0.122430) | 0.439126 / 0.323480 (0.115647) | 0.006455 / 0.007986 (-0.001531) | 0.004575 / 0.004328 (0.000246) | 0.083396 / 0.004250 (0.079146) | 0.052827 / 0.037052 (0.015775) | 0.401039 / 0.258489 (0.142550) | 0.441374 / 0.293841 (0.147533) | 0.041671 / 0.128546 (-0.086875) | 0.014098 / 0.075646 (-0.061548) | 0.100873 / 0.419271 (-0.318398) | 0.058690 / 0.043533 (0.015157) | 0.395817 / 0.255139 (0.140678) | 0.409226 / 0.283200 (0.126026) | 0.119804 / 0.141683 (-0.021879) | 1.704583 / 1.452155 (0.252428) | 1.782527 / 1.492716 (0.289811) |\n\n### Benchmark: benchmark_getitem\\_100B.json\n\n| metric | get_batch_of\\_1024\\_random_rows | get_batch_of\\_1024\\_rows | get_first_row | get_last_row |\n|--------|---|---|---|---|\n| new / old (diff) | 0.255166 / 0.018006 (0.237160) | 0.485091 / 0.000490 (0.484601) | 0.007458 / 0.000200 (0.007258) | 0.000116 / 0.000054 (0.000061) |\n\n### Benchmark: benchmark_indices_mapping.json\n\n| metric | select | shard | shuffle | sort | train_test_split |\n|--------|---|---|---|---|---|\n| new / old (diff) | 0.034531 / 0.037411 (-0.002880) | 0.134332 / 0.014526 (0.119806) | 0.144944 / 0.176557 (-0.031613) | 0.199352 / 0.737135 (-0.537783) | 0.152243 / 0.296338 (-0.144095) |\n\n### Benchmark: benchmark_iterating.json\n\n| metric | read 5000 | read 50000 | read_batch 50000 10 | read_batch 50000 100 | read_batch 50000 1000 | read_formatted numpy 5000 | read_formatted pandas 5000 | read_formatted tensorflow 5000 | read_formatted torch 5000 | read_formatted_batch numpy 5000 10 | read_formatted_batch numpy 5000 1000 | shuffled read 5000 | shuffled read 50000 | shuffled read_batch 50000 10 | shuffled read_batch 50000 100 | shuffled read_batch 50000 1000 | shuffled read_formatted numpy 5000 | shuffled read_formatted_batch numpy 5000 10 | shuffled read_formatted_batch numpy 5000 1000 |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.495361 / 0.215209 (0.280152) | 4.895144 / 2.077655 (2.817489) | 2.350419 / 1.504120 (0.846299) | 2.112131 / 1.541195 (0.570937) | 2.234469 / 1.468490 (0.765978) | 0.815862 / 4.584777 (-3.768915) | 4.531638 / 3.745712 (0.785926) | 2.405186 / 5.269862 (-2.864676) | 1.559020 / 4.565676 (-3.006656) | 0.100432 / 0.424275 (-0.323843) | 0.014217 / 0.007607 (0.006610) | 0.614622 / 0.226044 (0.388577) | 5.984541 / 2.268929 (3.715613) | 2.929897 / 55.444624 (-52.514727) | 2.484010 / 6.876477 (-4.392467) | 2.533538 / 2.142072 (0.391466) | 0.972119 / 4.805227 (-3.833108) | 0.193630 / 6.500664 (-6.307034) | 0.073694 / 0.075469 (-0.001775) |\n\n### Benchmark: benchmark_map_filter.json\n\n| metric | filter | map fast-tokenizer batched | map identity | map identity batched | map no-op batched | map no-op batched numpy | map no-op batched pandas | map no-op batched pytorch | map no-op batched tensorflow |\n|--------|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 1.503725 / 1.841788 (-0.338063) | 17.421529 / 8.074308 (9.347221) | 15.686433 / 10.191392 (5.495041) | 0.216688 / 0.680424 (-0.463736) | 0.020929 / 0.534201 (-0.513272) | 0.512523 / 0.579283 (-0.066760) | 0.499878 / 0.434364 (0.065514) | 0.639238 / 0.540337 (0.098900) | 0.769598 / 1.386936 (-0.617338) |\n\n</details>\n</details>\n\n![](https://cml.dev/watermark.png#99200127ade6d7b7d2cfb7b88365e5844b5c9c2e \"CML watermark\")\n", "> > Do you want me to work on the CI in a separate branch? Thanks for merging and for your help as always :)\r\n> \r\n> In the end I think we can keep it as is since we didn't modify the core code for jax. Maybe later if we do further changes and need to make sure we don't break anything ;) For example when we decide to add support for more recent versions\r\n\r\nMakes sense, thank you @lhoestq!" ]

[]

[]

[ "_The documentation is not available anymore as the PR was closed or merged._" ]

[]

[]

[]

[]

[]

[ "_The documentation is not available anymore as the PR was closed or merged._", "<details>\n<summary>Show benchmarks</summary>\n\nPyArrow==6.0.0\n\n<details>\n<summary>Show updated benchmarks!</summary>\n\n### Benchmark: benchmark_array_xd.json\n\n| metric | read_batch_formatted_as_numpy after write_array2d | read_batch_formatted_as_numpy after write_flattened_sequence | read_batch_formatted_as_numpy after write_nested_sequence | read_batch_unformated after write_array2d | read_batch_unformated after write_flattened_sequence | read_batch_unformated after write_nested_sequence | read_col_formatted_as_numpy after write_array2d | read_col_formatted_as_numpy after write_flattened_sequence | read_col_formatted_as_numpy after write_nested_sequence | read_col_unformated after write_array2d | read_col_unformated after write_flattened_sequence | read_col_unformated after write_nested_sequence | read_formatted_as_numpy after write_array2d | read_formatted_as_numpy after write_flattened_sequence | read_formatted_as_numpy after write_nested_sequence | read_unformated after write_array2d | read_unformated after write_flattened_sequence | read_unformated after write_nested_sequence | write_array2d | write_flattened_sequence | write_nested_sequence |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.009729 / 0.011353 (-0.001624) | 0.005342 / 0.011008 (-0.005666) | 0.100194 / 0.038508 (0.061686) | 0.036391 / 0.023109 (0.013282) | 0.294163 / 0.275898 (0.018264) | 0.364117 / 0.323480 (0.040637) | 0.008231 / 0.007986 (0.000246) | 0.005954 / 0.004328 (0.001626) | 0.076484 / 0.004250 (0.072234) | 0.045028 / 0.037052 (0.007976) | 0.308163 / 0.258489 (0.049674) | 0.339473 / 0.293841 (0.045632) | 0.039268 / 0.128546 (-0.089279) | 0.012357 / 0.075646 (-0.063289) | 0.334176 / 0.419271 (-0.085096) | 0.049502 / 0.043533 (0.005969) | 0.294134 / 0.255139 (0.038995) | 0.319370 / 0.283200 (0.036170) | 0.113040 / 0.141683 (-0.028643) | 1.450750 / 1.452155 (-0.001405) | 1.490265 / 1.492716 (-0.002452) |\n\n### Benchmark: benchmark_getitem\\_100B.json\n\n| metric | get_batch_of\\_1024\\_random_rows | get_batch_of\\_1024\\_rows | get_first_row | get_last_row |\n|--------|---|---|---|---|\n| new / old (diff) | 0.252860 / 0.018006 (0.234854) | 0.554299 / 0.000490 (0.553810) | 0.002105 / 0.000200 (0.001905) | 0.000091 / 0.000054 (0.000037) |\n\n### Benchmark: benchmark_indices_mapping.json\n\n| metric | select | shard | shuffle | sort | train_test_split |\n|--------|---|---|---|---|---|\n| new / old (diff) | 0.026557 / 0.037411 (-0.010854) | 0.104464 / 0.014526 (0.089938) | 0.116724 / 0.176557 (-0.059833) | 0.154736 / 0.737135 (-0.582399) | 0.122017 / 0.296338 (-0.174322) |\n\n### Benchmark: benchmark_iterating.json\n\n| metric | read 5000 | read 50000 | read_batch 50000 10 | read_batch 50000 100 | read_batch 50000 1000 | read_formatted numpy 5000 | read_formatted pandas 5000 | read_formatted tensorflow 5000 | read_formatted torch 5000 | read_formatted_batch numpy 5000 10 | read_formatted_batch numpy 5000 1000 | shuffled read 5000 | shuffled read 50000 | shuffled read_batch 50000 10 | shuffled read_batch 50000 100 | shuffled read_batch 50000 1000 | shuffled read_formatted numpy 5000 | shuffled read_formatted_batch numpy 5000 10 | shuffled read_formatted_batch numpy 5000 1000 |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.398170 / 0.215209 (0.182961) | 3.979309 / 2.077655 (1.901654) | 1.773051 / 1.504120 (0.268931) | 1.587247 / 1.541195 (0.046053) | 1.620446 / 1.468490 (0.151956) | 0.692152 / 4.584777 (-3.892625) | 3.724821 / 3.745712 (-0.020891) | 2.133122 / 5.269862 (-3.136739) | 1.455612 / 4.565676 (-3.110065) | 0.084721 / 0.424275 (-0.339554) | 0.012461 / 0.007607 (0.004854) | 0.498909 / 0.226044 (0.272865) | 4.983837 / 2.268929 (2.714908) | 2.258489 / 55.444624 (-53.186135) | 1.891690 / 6.876477 (-4.984786) | 1.976944 / 2.142072 (-0.165128) | 0.836950 / 4.805227 (-3.968277) | 0.165401 / 6.500664 (-6.335263) | 0.061623 / 0.075469 (-0.013846) |\n\n### Benchmark: benchmark_map_filter.json\n\n| metric | filter | map fast-tokenizer batched | map identity | map identity batched | map no-op batched | map no-op batched numpy | map no-op batched pandas | map no-op batched pytorch | map no-op batched tensorflow |\n|--------|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 1.205945 / 1.841788 (-0.635842) | 15.101603 / 8.074308 (7.027295) | 14.393739 / 10.191392 (4.202347) | 0.176313 / 0.680424 (-0.504110) | 0.029102 / 0.534201 (-0.505099) | 0.439785 / 0.579283 (-0.139498) | 0.437360 / 0.434364 (0.002996) | 0.539668 / 0.540337 (-0.000669) | 0.641452 / 1.386936 (-0.745484) |\n\n</details>\nPyArrow==latest\n\n<details>\n<summary>Show updated benchmarks!</summary>\n\n### Benchmark: benchmark_array_xd.json\n\n| metric | read_batch_formatted_as_numpy after write_array2d | read_batch_formatted_as_numpy after write_flattened_sequence | read_batch_formatted_as_numpy after write_nested_sequence | read_batch_unformated after write_array2d | read_batch_unformated after write_flattened_sequence | read_batch_unformated after write_nested_sequence | read_col_formatted_as_numpy after write_array2d | read_col_formatted_as_numpy after write_flattened_sequence | read_col_formatted_as_numpy after write_nested_sequence | read_col_unformated after write_array2d | read_col_unformated after write_flattened_sequence | read_col_unformated after write_nested_sequence | read_formatted_as_numpy after write_array2d | read_formatted_as_numpy after write_flattened_sequence | read_formatted_as_numpy after write_nested_sequence | read_unformated after write_array2d | read_unformated after write_flattened_sequence | read_unformated after write_nested_sequence | write_array2d | write_flattened_sequence | write_nested_sequence |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.007184 / 0.011353 (-0.004169) | 0.005215 / 0.011008 (-0.005793) | 0.074617 / 0.038508 (0.036109) | 0.033209 / 0.023109 (0.010100) | 0.334304 / 0.275898 (0.058406) | 0.370270 / 0.323480 (0.046790) | 0.005851 / 0.007986 (-0.002135) | 0.004106 / 0.004328 (-0.000222) | 0.075487 / 0.004250 (0.071237) | 0.051133 / 0.037052 (0.014080) | 0.335401 / 0.258489 (0.076912) | 0.391457 / 0.293841 (0.097616) | 0.036525 / 0.128546 (-0.092021) | 0.012423 / 0.075646 (-0.063223) | 0.086446 / 0.419271 (-0.332825) | 0.050707 / 0.043533 (0.007174) | 0.336186 / 0.255139 (0.081047) | 0.353273 / 0.283200 (0.070074) | 0.105625 / 0.141683 (-0.036057) | 1.486118 / 1.452155 (0.033963) | 1.584931 / 1.492716 (0.092214) |\n\n### Benchmark: benchmark_getitem\\_100B.json\n\n| metric | get_batch_of\\_1024\\_random_rows | get_batch_of\\_1024\\_rows | get_first_row | get_last_row |\n|--------|---|---|---|---|\n| new / old (diff) | 0.237589 / 0.018006 (0.219583) | 0.552030 / 0.000490 (0.551540) | 0.002863 / 0.000200 (0.002663) | 0.000097 / 0.000054 (0.000042) |\n\n### Benchmark: benchmark_indices_mapping.json\n\n| metric | select | shard | shuffle | sort | train_test_split |\n|--------|---|---|---|---|---|\n| new / old (diff) | 0.028078 / 0.037411 (-0.009333) | 0.112516 / 0.014526 (0.097990) | 0.121119 / 0.176557 (-0.055438) | 0.158874 / 0.737135 (-0.578262) | 0.129501 / 0.296338 (-0.166837) |\n\n### Benchmark: benchmark_iterating.json\n\n| metric | read 5000 | read 50000 | read_batch 50000 10 | read_batch 50000 100 | read_batch 50000 1000 | read_formatted numpy 5000 | read_formatted pandas 5000 | read_formatted tensorflow 5000 | read_formatted torch 5000 | read_formatted_batch numpy 5000 10 | read_formatted_batch numpy 5000 1000 | shuffled read 5000 | shuffled read 50000 | shuffled read_batch 50000 10 | shuffled read_batch 50000 100 | shuffled read_batch 50000 1000 | shuffled read_formatted numpy 5000 | shuffled read_formatted_batch numpy 5000 10 | shuffled read_formatted_batch numpy 5000 1000 |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.419479 / 0.215209 (0.204270) | 4.192216 / 2.077655 (2.114561) | 1.990513 / 1.504120 (0.486393) | 1.792892 / 1.541195 (0.251697) | 1.853904 / 1.468490 (0.385413) | 0.712702 / 4.584777 (-3.872074) | 3.820682 / 3.745712 (0.074970) | 2.143695 / 5.269862 (-3.126166) | 1.369621 / 4.565676 (-3.196055) | 0.087451 / 0.424275 (-0.336824) | 0.012622 / 0.007607 (0.005014) | 0.521056 / 0.226044 (0.295011) | 5.204873 / 2.268929 (2.935944) | 2.481169 / 55.444624 (-52.963455) | 2.112134 / 6.876477 (-4.764342) | 2.200681 / 2.142072 (0.058609) | 0.860323 / 4.805227 (-3.944904) | 0.171452 / 6.500664 (-6.329212) | 0.065235 / 0.075469 (-0.010234) |\n\n### Benchmark: benchmark_map_filter.json\n\n| metric | filter | map fast-tokenizer batched | map identity | map identity batched | map no-op batched | map no-op batched numpy | map no-op batched pandas | map no-op batched pytorch | map no-op batched tensorflow |\n|--------|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 1.241047 / 1.841788 (-0.600741) | 14.977890 / 8.074308 (6.903582) | 13.584265 / 10.191392 (3.392873) | 0.180050 / 0.680424 (-0.500374) | 0.018247 / 0.534201 (-0.515954) | 0.429585 / 0.579283 (-0.149698) | 0.429448 / 0.434364 (-0.004916) | 0.542663 / 0.540337 (0.002326) | 0.649525 / 1.386936 (-0.737411) |\n\n</details>\n</details>\n\n![](https://cml.dev/watermark.png#26cf1d2548eb313a06565d36bd400436e350bc86 \"CML watermark\")\n", "<details>\n<summary>Show benchmarks</summary>\n\nPyArrow==6.0.0\n\n<details>\n<summary>Show updated benchmarks!</summary>\n\n### Benchmark: benchmark_array_xd.json\n\n| metric | read_batch_formatted_as_numpy after write_array2d | read_batch_formatted_as_numpy after write_flattened_sequence | read_batch_formatted_as_numpy after write_nested_sequence | read_batch_unformated after write_array2d | read_batch_unformated after write_flattened_sequence | read_batch_unformated after write_nested_sequence | read_col_formatted_as_numpy after write_array2d | read_col_formatted_as_numpy after write_flattened_sequence | read_col_formatted_as_numpy after write_nested_sequence | read_col_unformated after write_array2d | read_col_unformated after write_flattened_sequence | read_col_unformated after write_nested_sequence | read_formatted_as_numpy after write_array2d | read_formatted_as_numpy after write_flattened_sequence | read_formatted_as_numpy after write_nested_sequence | read_unformated after write_array2d | read_unformated after write_flattened_sequence | read_unformated after write_nested_sequence | write_array2d | write_flattened_sequence | write_nested_sequence |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.011289 / 0.011353 (-0.000064) | 0.005841 / 0.011008 (-0.005167) | 0.120994 / 0.038508 (0.082486) | 0.043627 / 0.023109 (0.020517) | 0.353254 / 0.275898 (0.077356) | 0.394685 / 0.323480 (0.071205) | 0.009520 / 0.007986 (0.001535) | 0.004770 / 0.004328 (0.000442) | 0.088857 / 0.004250 (0.084607) | 0.048426 / 0.037052 (0.011373) | 0.353815 / 0.258489 (0.095326) | 0.404109 / 0.293841 (0.110268) | 0.060079 / 0.128546 (-0.068467) | 0.013840 / 0.075646 (-0.061806) | 0.403133 / 0.419271 (-0.016139) | 0.072227 / 0.043533 (0.028694) | 0.354585 / 0.255139 (0.099446) | 0.377937 / 0.283200 (0.094737) | 0.139080 / 0.141683 (-0.002602) | 1.733266 / 1.452155 (0.281112) | 1.828402 / 1.492716 (0.335686) |\n\n### Benchmark: benchmark_getitem\\_100B.json\n\n| metric | get_batch_of\\_1024\\_random_rows | get_batch_of\\_1024\\_rows | get_first_row | get_last_row |\n|--------|---|---|---|---|\n| new / old (diff) | 0.215095 / 0.018006 (0.197088) | 0.486669 / 0.000490 (0.486179) | 0.001425 / 0.000200 (0.001225) | 0.000097 / 0.000054 (0.000043) |\n\n### Benchmark: benchmark_indices_mapping.json\n\n| metric | select | shard | shuffle | sort | train_test_split |\n|--------|---|---|---|---|---|\n| new / old (diff) | 0.032832 / 0.037411 (-0.004579) | 0.136335 / 0.014526 (0.121809) | 0.141827 / 0.176557 (-0.034730) | 0.185917 / 0.737135 (-0.551218) | 0.149046 / 0.296338 (-0.147293) |\n\n### Benchmark: benchmark_iterating.json\n\n| metric | read 5000 | read 50000 | read_batch 50000 10 | read_batch 50000 100 | read_batch 50000 1000 | read_formatted numpy 5000 | read_formatted pandas 5000 | read_formatted tensorflow 5000 | read_formatted torch 5000 | read_formatted_batch numpy 5000 10 | read_formatted_batch numpy 5000 1000 | shuffled read 5000 | shuffled read 50000 | shuffled read_batch 50000 10 | shuffled read_batch 50000 100 | shuffled read_batch 50000 1000 | shuffled read_formatted numpy 5000 | shuffled read_formatted_batch numpy 5000 10 | shuffled read_formatted_batch numpy 5000 1000 |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.474587 / 0.215209 (0.259378) | 4.753686 / 2.077655 (2.676031) | 2.152147 / 1.504120 (0.648027) | 1.941762 / 1.541195 (0.400567) | 2.077493 / 1.468490 (0.609003) | 0.822432 / 4.584777 (-3.762345) | 4.860151 / 3.745712 (1.114439) | 2.527292 / 5.269862 (-2.742569) | 1.580442 / 4.565676 (-2.985234) | 0.102104 / 0.424275 (-0.322171) | 0.015060 / 0.007607 (0.007453) | 0.598780 / 0.226044 (0.372736) | 5.998318 / 2.268929 (3.729390) | 2.754115 / 55.444624 (-52.690509) | 2.317509 / 6.876477 (-4.558967) | 2.409942 / 2.142072 (0.267870) | 1.008830 / 4.805227 (-3.796397) | 0.196203 / 6.500664 (-6.304461) | 0.075378 / 0.075469 (-0.000091) |\n\n### Benchmark: benchmark_map_filter.json\n\n| metric | filter | map fast-tokenizer batched | map identity | map identity batched | map no-op batched | map no-op batched numpy | map no-op batched pandas | map no-op batched pytorch | map no-op batched tensorflow |\n|--------|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 1.430676 / 1.841788 (-0.411112) | 19.597628 / 8.074308 (11.523320) | 17.364673 / 10.191392 (7.173281) | 0.216621 / 0.680424 (-0.463803) | 0.039505 / 0.534201 (-0.494696) | 0.529027 / 0.579283 (-0.050256) | 0.572014 / 0.434364 (0.137650) | 0.702898 / 0.540337 (0.162560) | 0.785748 / 1.386936 (-0.601188) |\n\n</details>\nPyArrow==latest\n\n<details>\n<summary>Show updated benchmarks!</summary>\n\n### Benchmark: benchmark_array_xd.json\n\n| metric | read_batch_formatted_as_numpy after write_array2d | read_batch_formatted_as_numpy after write_flattened_sequence | read_batch_formatted_as_numpy after write_nested_sequence | read_batch_unformated after write_array2d | read_batch_unformated after write_flattened_sequence | read_batch_unformated after write_nested_sequence | read_col_formatted_as_numpy after write_array2d | read_col_formatted_as_numpy after write_flattened_sequence | read_col_formatted_as_numpy after write_nested_sequence | read_col_unformated after write_array2d | read_col_unformated after write_flattened_sequence | read_col_unformated after write_nested_sequence | read_formatted_as_numpy after write_array2d | read_formatted_as_numpy after write_flattened_sequence | read_formatted_as_numpy after write_nested_sequence | read_unformated after write_array2d | read_unformated after write_flattened_sequence | read_unformated after write_nested_sequence | write_array2d | write_flattened_sequence | write_nested_sequence |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.009150 / 0.011353 (-0.002203) | 0.006088 / 0.011008 (-0.004920) | 0.090629 / 0.038508 (0.052121) | 0.044284 / 0.023109 (0.021174) | 0.411363 / 0.275898 (0.135465) | 0.445499 / 0.323480 (0.122020) | 0.007129 / 0.007986 (-0.000856) | 0.004843 / 0.004328 (0.000515) | 0.087919 / 0.004250 (0.083668) | 0.060329 / 0.037052 (0.023277) | 0.405802 / 0.258489 (0.147313) | 0.468301 / 0.293841 (0.174460) | 0.044271 / 0.128546 (-0.084275) | 0.014895 / 0.075646 (-0.060751) | 0.103728 / 0.419271 (-0.315544) | 0.084190 / 0.043533 (0.040657) | 0.407210 / 0.255139 (0.152071) | 0.432585 / 0.283200 (0.149386) | 0.137132 / 0.141683 (-0.004550) | 1.720261 / 1.452155 (0.268107) | 1.858575 / 1.492716 (0.365858) |\n\n### Benchmark: benchmark_getitem\\_100B.json\n\n| metric | get_batch_of\\_1024\\_random_rows | get_batch_of\\_1024\\_rows | get_first_row | get_last_row |\n|--------|---|---|---|---|\n| new / old (diff) | 0.331395 / 0.018006 (0.313389) | 0.494757 / 0.000490 (0.494267) | 0.043426 / 0.000200 (0.043226) | 0.000470 / 0.000054 (0.000415) |\n\n### Benchmark: benchmark_indices_mapping.json\n\n| metric | select | shard | shuffle | sort | train_test_split |\n|--------|---|---|---|---|---|\n| new / old (diff) | 0.035288 / 0.037411 (-0.002123) | 0.140856 / 0.014526 (0.126330) | 0.146597 / 0.176557 (-0.029959) | 0.192775 / 0.737135 (-0.544360) | 0.155307 / 0.296338 (-0.141032) |\n\n### Benchmark: benchmark_iterating.json\n\n| metric | read 5000 | read 50000 | read_batch 50000 10 | read_batch 50000 100 | read_batch 50000 1000 | read_formatted numpy 5000 | read_formatted pandas 5000 | read_formatted tensorflow 5000 | read_formatted torch 5000 | read_formatted_batch numpy 5000 10 | read_formatted_batch numpy 5000 1000 | shuffled read 5000 | shuffled read 50000 | shuffled read_batch 50000 10 | shuffled read_batch 50000 100 | shuffled read_batch 50000 1000 | shuffled read_formatted numpy 5000 | shuffled read_formatted_batch numpy 5000 10 | shuffled read_formatted_batch numpy 5000 1000 |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.504000 / 0.215209 (0.288791) | 5.011081 / 2.077655 (2.933427) | 2.380420 / 1.504120 (0.876300) | 2.154819 / 1.541195 (0.613624) | 2.293883 / 1.468490 (0.825393) | 0.864429 / 4.584777 (-3.720348) | 5.134475 / 3.745712 (1.388763) | 4.984024 / 5.269862 (-0.285837) | 2.333754 / 4.565676 (-2.231923) | 0.105854 / 0.424275 (-0.318422) | 0.015833 / 0.007607 (0.008226) | 0.633614 / 0.226044 (0.407569) | 6.330974 / 2.268929 (4.062046) | 3.020498 / 55.444624 (-52.424126) | 2.578234 / 6.876477 (-4.298243) | 2.654429 / 2.142072 (0.512357) | 1.022041 / 4.805227 (-3.783186) | 0.205085 / 6.500664 (-6.295579) | 0.081122 / 0.075469 (0.005653) |\n\n### Benchmark: benchmark_map_filter.json\n\n| metric | filter | map fast-tokenizer batched | map identity | map identity batched | map no-op batched | map no-op batched numpy | map no-op batched pandas | map no-op batched pytorch | map no-op batched tensorflow |\n|--------|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 1.538929 / 1.841788 (-0.302859) | 19.907799 / 8.074308 (11.833490) | 17.174568 / 10.191392 (6.983176) | 0.228165 / 0.680424 (-0.452258) | 0.024688 / 0.534201 (-0.509513) | 0.508958 / 0.579283 (-0.070326) | 0.544469 / 0.434364 (0.110105) | 0.590805 / 0.540337 (0.050468) | 0.705947 / 1.386936 (-0.680989) |\n\n</details>\n</details>\n\n![](https://cml.dev/watermark.png#2573861afb170fd575dbe67270294a4e88ab4be6 \"CML watermark\")\n", "<details>\n<summary>Show benchmarks</summary>\n\nPyArrow==6.0.0\n\n<details>\n<summary>Show updated benchmarks!</summary>\n\n### Benchmark: benchmark_array_xd.json\n\n| metric | read_batch_formatted_as_numpy after write_array2d | read_batch_formatted_as_numpy after write_flattened_sequence | read_batch_formatted_as_numpy after write_nested_sequence | read_batch_unformated after write_array2d | read_batch_unformated after write_flattened_sequence | read_batch_unformated after write_nested_sequence | read_col_formatted_as_numpy after write_array2d | read_col_formatted_as_numpy after write_flattened_sequence | read_col_formatted_as_numpy after write_nested_sequence | read_col_unformated after write_array2d | read_col_unformated after write_flattened_sequence | read_col_unformated after write_nested_sequence | read_formatted_as_numpy after write_array2d | read_formatted_as_numpy after write_flattened_sequence | read_formatted_as_numpy after write_nested_sequence | read_unformated after write_array2d | read_unformated after write_flattened_sequence | read_unformated after write_nested_sequence | write_array2d | write_flattened_sequence | write_nested_sequence |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.008377 / 0.011353 (-0.002975) | 0.004445 / 0.011008 (-0.006563) | 0.100671 / 0.038508 (0.062163) | 0.029216 / 0.023109 (0.006107) | 0.300311 / 0.275898 (0.024413) | 0.356907 / 0.323480 (0.033427) | 0.006921 / 0.007986 (-0.001065) | 0.003384 / 0.004328 (-0.000944) | 0.078529 / 0.004250 (0.074278) | 0.034689 / 0.037052 (-0.002364) | 0.304647 / 0.258489 (0.046158) | 0.343584 / 0.293841 (0.049743) | 0.032700 / 0.128546 (-0.095846) | 0.011403 / 0.075646 (-0.064244) | 0.321540 / 0.419271 (-0.097732) | 0.040770 / 0.043533 (-0.002762) | 0.306900 / 0.255139 (0.051761) | 0.322482 / 0.283200 (0.039282) | 0.085396 / 0.141683 (-0.056287) | 1.450735 / 1.452155 (-0.001419) | 1.491829 / 1.492716 (-0.000888) |\n\n### Benchmark: benchmark_getitem\\_100B.json\n\n| metric | get_batch_of\\_1024\\_random_rows | get_batch_of\\_1024\\_rows | get_first_row | get_last_row |\n|--------|---|---|---|---|\n| new / old (diff) | 0.009439 / 0.018006 (-0.008567) | 0.406805 / 0.000490 (0.406315) | 0.002993 / 0.000200 (0.002793) | 0.000102 / 0.000054 (0.000048) |\n\n### Benchmark: benchmark_indices_mapping.json\n\n| metric | select | shard | shuffle | sort | train_test_split |\n|--------|---|---|---|---|---|\n| new / old (diff) | 0.025034 / 0.037411 (-0.012378) | 0.100567 / 0.014526 (0.086042) | 0.107267 / 0.176557 (-0.069290) | 0.149945 / 0.737135 (-0.587190) | 0.111150 / 0.296338 (-0.185189) |\n\n### Benchmark: benchmark_iterating.json\n\n| metric | read 5000 | read 50000 | read_batch 50000 10 | read_batch 50000 100 | read_batch 50000 1000 | read_formatted numpy 5000 | read_formatted pandas 5000 | read_formatted tensorflow 5000 | read_formatted torch 5000 | read_formatted_batch numpy 5000 10 | read_formatted_batch numpy 5000 1000 | shuffled read 5000 | shuffled read 50000 | shuffled read_batch 50000 10 | shuffled read_batch 50000 100 | shuffled read_batch 50000 1000 | shuffled read_formatted numpy 5000 | shuffled read_formatted_batch numpy 5000 10 | shuffled read_formatted_batch numpy 5000 1000 |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.418387 / 0.215209 (0.203178) | 4.177979 / 2.077655 (2.100324) | 1.886650 / 1.504120 (0.382530) | 1.685692 / 1.541195 (0.144497) | 1.728270 / 1.468490 (0.259780) | 0.700904 / 4.584777 (-3.883873) | 3.379998 / 3.745712 (-0.365714) | 1.874779 / 5.269862 (-3.395083) | 1.170366 / 4.565676 (-3.395310) | 0.083190 / 0.424275 (-0.341085) | 0.012506 / 0.007607 (0.004899) | 0.528633 / 0.226044 (0.302589) | 5.301793 / 2.268929 (3.032865) | 2.334050 / 55.444624 (-53.110574) | 1.986988 / 6.876477 (-4.889488) | 2.020508 / 2.142072 (-0.121565) | 0.817227 / 4.805227 (-3.988000) | 0.150284 / 6.500664 (-6.350380) | 0.065489 / 0.075469 (-0.009980) |\n\n### Benchmark: benchmark_map_filter.json\n\n| metric | filter | map fast-tokenizer batched | map identity | map identity batched | map no-op batched | map no-op batched numpy | map no-op batched pandas | map no-op batched pytorch | map no-op batched tensorflow |\n|--------|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 1.224216 / 1.841788 (-0.617572) | 13.729808 / 8.074308 (5.655500) | 14.283402 / 10.191392 (4.092010) | 0.159434 / 0.680424 (-0.520990) | 0.028471 / 0.534201 (-0.505730) | 0.395102 / 0.579283 (-0.184181) | 0.402733 / 0.434364 (-0.031631) | 0.470852 / 0.540337 (-0.069485) | 0.568530 / 1.386936 (-0.818406) |\n\n</details>\nPyArrow==latest\n\n<details>\n<summary>Show updated benchmarks!</summary>\n\n### Benchmark: benchmark_array_xd.json\n\n| metric | read_batch_formatted_as_numpy after write_array2d | read_batch_formatted_as_numpy after write_flattened_sequence | read_batch_formatted_as_numpy after write_nested_sequence | read_batch_unformated after write_array2d | read_batch_unformated after write_flattened_sequence | read_batch_unformated after write_nested_sequence | read_col_formatted_as_numpy after write_array2d | read_col_formatted_as_numpy after write_flattened_sequence | read_col_formatted_as_numpy after write_nested_sequence | read_col_unformated after write_array2d | read_col_unformated after write_flattened_sequence | read_col_unformated after write_nested_sequence | read_formatted_as_numpy after write_array2d | read_formatted_as_numpy after write_flattened_sequence | read_formatted_as_numpy after write_nested_sequence | read_unformated after write_array2d | read_unformated after write_flattened_sequence | read_unformated after write_nested_sequence | write_array2d | write_flattened_sequence | write_nested_sequence |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.006750 / 0.011353 (-0.004603) | 0.004479 / 0.011008 (-0.006529) | 0.074926 / 0.038508 (0.036418) | 0.027619 / 0.023109 (0.004510) | 0.342070 / 0.275898 (0.066172) | 0.372452 / 0.323480 (0.048972) | 0.005094 / 0.007986 (-0.002892) | 0.003494 / 0.004328 (-0.000834) | 0.074963 / 0.004250 (0.070713) | 0.038457 / 0.037052 (0.001405) | 0.340587 / 0.258489 (0.082098) | 0.381212 / 0.293841 (0.087371) | 0.031597 / 0.128546 (-0.096950) | 0.011631 / 0.075646 (-0.064015) | 0.084646 / 0.419271 (-0.334626) | 0.042072 / 0.043533 (-0.001461) | 0.340977 / 0.255139 (0.085838) | 0.366502 / 0.283200 (0.083302) | 0.091181 / 0.141683 (-0.050502) | 1.435119 / 1.452155 (-0.017035) | 1.520426 / 1.492716 (0.027710) |\n\n### Benchmark: benchmark_getitem\\_100B.json\n\n| metric | get_batch_of\\_1024\\_random_rows | get_batch_of\\_1024\\_rows | get_first_row | get_last_row |\n|--------|---|---|---|---|\n| new / old (diff) | 0.211320 / 0.018006 (0.193313) | 0.466154 / 0.000490 (0.465664) | 0.002901 / 0.000200 (0.002701) | 0.000080 / 0.000054 (0.000026) |\n\n### Benchmark: benchmark_indices_mapping.json\n\n| metric | select | shard | shuffle | sort | train_test_split |\n|--------|---|---|---|---|---|\n| new / old (diff) | 0.025122 / 0.037411 (-0.012289) | 0.098929 / 0.014526 (0.084403) | 0.106551 / 0.176557 (-0.070005) | 0.142820 / 0.737135 (-0.594316) | 0.110701 / 0.296338 (-0.185637) |\n\n### Benchmark: benchmark_iterating.json\n\n| metric | read 5000 | read 50000 | read_batch 50000 10 | read_batch 50000 100 | read_batch 50000 1000 | read_formatted numpy 5000 | read_formatted pandas 5000 | read_formatted tensorflow 5000 | read_formatted torch 5000 | read_formatted_batch numpy 5000 10 | read_formatted_batch numpy 5000 1000 | shuffled read 5000 | shuffled read 50000 | shuffled read_batch 50000 10 | shuffled read_batch 50000 100 | shuffled read_batch 50000 1000 | shuffled read_formatted numpy 5000 | shuffled read_formatted_batch numpy 5000 10 | shuffled read_formatted_batch numpy 5000 1000 |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.445187 / 0.215209 (0.229978) | 4.457524 / 2.077655 (2.379870) | 2.088323 / 1.504120 (0.584203) | 1.888076 / 1.541195 (0.346881) | 1.923340 / 1.468490 (0.454850) | 0.723354 / 4.584777 (-3.861423) | 3.428479 / 3.745712 (-0.317233) | 1.914580 / 5.269862 (-3.355281) | 1.191810 / 4.565676 (-3.373866) | 0.087008 / 0.424275 (-0.337267) | 0.013431 / 0.007607 (0.005824) | 0.545089 / 0.226044 (0.319044) | 5.465887 / 2.268929 (3.196958) | 2.527431 / 55.444624 (-52.917194) | 2.240622 / 6.876477 (-4.635854) | 2.232472 / 2.142072 (0.090399) | 0.815968 / 4.805227 (-3.989259) | 0.152842 / 6.500664 (-6.347822) | 0.067152 / 0.075469 (-0.008317) |\n\n### Benchmark: benchmark_map_filter.json\n\n| metric | filter | map fast-tokenizer batched | map identity | map identity batched | map no-op batched | map no-op batched numpy | map no-op batched pandas | map no-op batched pytorch | map no-op batched tensorflow |\n|--------|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 1.328360 / 1.841788 (-0.513427) | 14.163349 / 8.074308 (6.089040) | 13.814255 / 10.191392 (3.622863) | 0.131684 / 0.680424 (-0.548740) | 0.016980 / 0.534201 (-0.517221) | 0.396045 / 0.579283 (-0.183238) | 0.395078 / 0.434364 (-0.039286) | 0.471728 / 0.540337 (-0.068609) | 0.567830 / 1.386936 (-0.819106) |\n\n</details>\n</details>\n\n![](https://cml.dev/watermark.png#82331b032891671c334afe30c5f3cc21245b2d72 \"CML watermark\")\n", "<details>\n<summary>Show benchmarks</summary>\n\nPyArrow==6.0.0\n\n<details>\n<summary>Show updated benchmarks!</summary>\n\n### Benchmark: benchmark_array_xd.json\n\n| metric | read_batch_formatted_as_numpy after write_array2d | read_batch_formatted_as_numpy after write_flattened_sequence | read_batch_formatted_as_numpy after write_nested_sequence | read_batch_unformated after write_array2d | read_batch_unformated after write_flattened_sequence | read_batch_unformated after write_nested_sequence | read_col_formatted_as_numpy after write_array2d | read_col_formatted_as_numpy after write_flattened_sequence | read_col_formatted_as_numpy after write_nested_sequence | read_col_unformated after write_array2d | read_col_unformated after write_flattened_sequence | read_col_unformated after write_nested_sequence | read_formatted_as_numpy after write_array2d | read_formatted_as_numpy after write_flattened_sequence | read_formatted_as_numpy after write_nested_sequence | read_unformated after write_array2d | read_unformated after write_flattened_sequence | read_unformated after write_nested_sequence | write_array2d | write_flattened_sequence | write_nested_sequence |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.012630 / 0.011353 (0.001277) | 0.007038 / 0.011008 (-0.003970) | 0.158816 / 0.038508 (0.120308) | 0.044142 / 0.023109 (0.021032) | 0.389393 / 0.275898 (0.113495) | 0.479745 / 0.323480 (0.156265) | 0.009335 / 0.007986 (0.001349) | 0.005434 / 0.004328 (0.001105) | 0.107747 / 0.004250 (0.103497) | 0.048382 / 0.037052 (0.011330) | 0.398144 / 0.258489 (0.139655) | 0.446373 / 0.293841 (0.152532) | 0.066285 / 0.128546 (-0.062261) | 0.021174 / 0.075646 (-0.054472) | 0.449176 / 0.419271 (0.029905) | 0.063044 / 0.043533 (0.019511) | 0.390523 / 0.255139 (0.135384) | 0.451435 / 0.283200 (0.168236) | 0.116369 / 0.141683 (-0.025314) | 1.881269 / 1.452155 (0.429114) | 1.944527 / 1.492716 (0.451811) |\n\n### Benchmark: benchmark_getitem\\_100B.json\n\n| metric | get_batch_of\\_1024\\_random_rows | get_batch_of\\_1024\\_rows | get_first_row | get_last_row |\n|--------|---|---|---|---|\n| new / old (diff) | 0.227989 / 0.018006 (0.209983) | 0.538514 / 0.000490 (0.538024) | 0.009404 / 0.000200 (0.009204) | 0.000510 / 0.000054 (0.000455) |\n\n### Benchmark: benchmark_indices_mapping.json\n\n| metric | select | shard | shuffle | sort | train_test_split |\n|--------|---|---|---|---|---|\n| new / old (diff) | 0.029826 / 0.037411 (-0.007585) | 0.129623 / 0.014526 (0.115098) | 0.142067 / 0.176557 (-0.034489) | 0.218586 / 0.737135 (-0.518549) | 0.160524 / 0.296338 (-0.135814) |\n\n### Benchmark: benchmark_iterating.json\n\n| metric | read 5000 | read 50000 | read_batch 50000 10 | read_batch 50000 100 | read_batch 50000 1000 | read_formatted numpy 5000 | read_formatted pandas 5000 | read_formatted tensorflow 5000 | read_formatted torch 5000 | read_formatted_batch numpy 5000 10 | read_formatted_batch numpy 5000 1000 | shuffled read 5000 | shuffled read 50000 | shuffled read_batch 50000 10 | shuffled read_batch 50000 100 | shuffled read_batch 50000 1000 | shuffled read_formatted numpy 5000 | shuffled read_formatted_batch numpy 5000 10 | shuffled read_formatted_batch numpy 5000 1000 |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.667195 / 0.215209 (0.451986) | 6.694192 / 2.077655 (4.616537) | 2.542493 / 1.504120 (1.038373) | 2.124042 / 1.541195 (0.582847) | 2.024854 / 1.468490 (0.556364) | 1.306222 / 4.584777 (-3.278555) | 5.631557 / 3.745712 (1.885845) | 3.405978 / 5.269862 (-1.863884) | 2.471399 / 4.565676 (-2.094278) | 0.165187 / 0.424275 (-0.259088) | 0.014880 / 0.007607 (0.007273) | 0.842718 / 0.226044 (0.616673) | 8.584358 / 2.268929 (6.315430) | 3.377228 / 55.444624 (-52.067396) | 2.667265 / 6.876477 (-4.209212) | 2.699462 / 2.142072 (0.557389) | 1.623115 / 4.805227 (-3.182112) | 0.253929 / 6.500664 (-6.246735) | 0.077189 / 0.075469 (0.001720) |\n\n### Benchmark: benchmark_map_filter.json\n\n| metric | filter | map fast-tokenizer batched | map identity | map identity batched | map no-op batched | map no-op batched numpy | map no-op batched pandas | map no-op batched pytorch | map no-op batched tensorflow |\n|--------|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 1.778962 / 1.841788 (-0.062825) | 18.997636 / 8.074308 (10.923328) | 24.255222 / 10.191392 (14.063830) | 0.304754 / 0.680424 (-0.375670) | 0.049656 / 0.534201 (-0.484545) | 0.590871 / 0.579283 (0.011588) | 0.649292 / 0.434364 (0.214928) | 0.751281 / 0.540337 (0.210943) | 0.872193 / 1.386936 (-0.514743) |\n\n</details>\nPyArrow==latest\n\n<details>\n<summary>Show updated benchmarks!</summary>\n\n### Benchmark: benchmark_array_xd.json\n\n| metric | read_batch_formatted_as_numpy after write_array2d | read_batch_formatted_as_numpy after write_flattened_sequence | read_batch_formatted_as_numpy after write_nested_sequence | read_batch_unformated after write_array2d | read_batch_unformated after write_flattened_sequence | read_batch_unformated after write_nested_sequence | read_col_formatted_as_numpy after write_array2d | read_col_formatted_as_numpy after write_flattened_sequence | read_col_formatted_as_numpy after write_nested_sequence | read_col_unformated after write_array2d | read_col_unformated after write_flattened_sequence | read_col_unformated after write_nested_sequence | read_formatted_as_numpy after write_array2d | read_formatted_as_numpy after write_flattened_sequence | read_formatted_as_numpy after write_nested_sequence | read_unformated after write_array2d | read_unformated after write_flattened_sequence | read_unformated after write_nested_sequence | write_array2d | write_flattened_sequence | write_nested_sequence |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.010660 / 0.011353 (-0.000693) | 0.006492 / 0.011008 (-0.004516) | 0.112190 / 0.038508 (0.073682) | 0.045391 / 0.023109 (0.022281) | 0.439852 / 0.275898 (0.163954) | 0.486489 / 0.323480 (0.163009) | 0.007155 / 0.007986 (-0.000830) | 0.006323 / 0.004328 (0.001995) | 0.099775 / 0.004250 (0.095525) | 0.055762 / 0.037052 (0.018709) | 0.439457 / 0.258489 (0.180968) | 0.505322 / 0.293841 (0.211481) | 0.057019 / 0.128546 (-0.071527) | 0.031382 / 0.075646 (-0.044264) | 0.121211 / 0.419271 (-0.298061) | 0.066091 / 0.043533 (0.022558) | 0.499760 / 0.255139 (0.244622) | 0.508312 / 0.283200 (0.225113) | 0.146975 / 0.141683 (0.005292) | 1.916347 / 1.452155 (0.464193) | 2.065860 / 1.492716 (0.573144) |\n\n### Benchmark: benchmark_getitem\\_100B.json\n\n| metric | get_batch_of\\_1024\\_random_rows | get_batch_of\\_1024\\_rows | get_first_row | get_last_row |\n|--------|---|---|---|---|\n| new / old (diff) | 0.247176 / 0.018006 (0.229170) | 0.565141 / 0.000490 (0.564652) | 0.004841 / 0.000200 (0.004641) | 0.000141 / 0.000054 (0.000087) |\n\n### Benchmark: benchmark_indices_mapping.json\n\n| metric | select | shard | shuffle | sort | train_test_split |\n|--------|---|---|---|---|---|\n| new / old (diff) | 0.036378 / 0.037411 (-0.001033) | 0.143470 / 0.014526 (0.128944) | 0.148096 / 0.176557 (-0.028461) | 0.225877 / 0.737135 (-0.511258) | 0.147072 / 0.296338 (-0.149266) |\n\n### Benchmark: benchmark_iterating.json\n\n| metric | read 5000 | read 50000 | read_batch 50000 10 | read_batch 50000 100 | read_batch 50000 1000 | read_formatted numpy 5000 | read_formatted pandas 5000 | read_formatted tensorflow 5000 | read_formatted torch 5000 | read_formatted_batch numpy 5000 10 | read_formatted_batch numpy 5000 1000 | shuffled read 5000 | shuffled read 50000 | shuffled read_batch 50000 10 | shuffled read_batch 50000 100 | shuffled read_batch 50000 1000 | shuffled read_formatted numpy 5000 | shuffled read_formatted_batch numpy 5000 10 | shuffled read_formatted_batch numpy 5000 1000 |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.723119 / 0.215209 (0.507910) | 6.824981 / 2.077655 (4.747326) | 2.883840 / 1.504120 (1.379720) | 2.468707 / 1.541195 (0.927513) | 2.525549 / 1.468490 (1.057059) | 1.426640 / 4.584777 (-3.158137) | 5.816045 / 3.745712 (2.070333) | 5.727037 / 5.269862 (0.457175) | 2.650307 / 4.565676 (-1.915369) | 0.160306 / 0.424275 (-0.263970) | 0.015371 / 0.007607 (0.007764) | 0.835778 / 0.226044 (0.609733) | 8.622836 / 2.268929 (6.353907) | 3.616338 / 55.444624 (-51.828287) | 2.974243 / 6.876477 (-3.902234) | 2.884557 / 2.142072 (0.742485) | 1.734874 / 4.805227 (-3.070353) | 0.277474 / 6.500664 (-6.223190) | 0.094189 / 0.075469 (0.018720) |\n\n### Benchmark: benchmark_map_filter.json\n\n| metric | filter | map fast-tokenizer batched | map identity | map identity batched | map no-op batched | map no-op batched numpy | map no-op batched pandas | map no-op batched pytorch | map no-op batched tensorflow |\n|--------|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 1.785728 / 1.841788 (-0.056059) | 19.376490 / 8.074308 (11.302182) | 24.560403 / 10.191392 (14.369011) | 0.250686 / 0.680424 (-0.429738) | 0.034333 / 0.534201 (-0.499868) | 0.557331 / 0.579283 (-0.021952) | 0.641007 / 0.434364 (0.206643) | 0.657138 / 0.540337 (0.116800) | 0.759023 / 1.386936 (-0.627913) |\n\n</details>\n</details>\n\n![](https://cml.dev/watermark.png#06ae3f678651bfbb3ca7dd3274ee2f38e0e0237e \"CML watermark\")\n" ]

[]

[]

[ "_The documentation is not available anymore as the PR was closed or merged._", "<details>\n<summary>Show benchmarks</summary>\n\nPyArrow==6.0.0\n\n<details>\n<summary>Show updated benchmarks!</summary>\n\n### Benchmark: benchmark_array_xd.json\n\n| metric | read_batch_formatted_as_numpy after write_array2d | read_batch_formatted_as_numpy after write_flattened_sequence | read_batch_formatted_as_numpy after write_nested_sequence | read_batch_unformated after write_array2d | read_batch_unformated after write_flattened_sequence | read_batch_unformated after write_nested_sequence | read_col_formatted_as_numpy after write_array2d | read_col_formatted_as_numpy after write_flattened_sequence | read_col_formatted_as_numpy after write_nested_sequence | read_col_unformated after write_array2d | read_col_unformated after write_flattened_sequence | read_col_unformated after write_nested_sequence | read_formatted_as_numpy after write_array2d | read_formatted_as_numpy after write_flattened_sequence | read_formatted_as_numpy after write_nested_sequence | read_unformated after write_array2d | read_unformated after write_flattened_sequence | read_unformated after write_nested_sequence | write_array2d | write_flattened_sequence | write_nested_sequence |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.008283 / 0.011353 (-0.003070) | 0.004450 / 0.011008 (-0.006558) | 0.099773 / 0.038508 (0.061265) | 0.029068 / 0.023109 (0.005959) | 0.296799 / 0.275898 (0.020901) | 0.350946 / 0.323480 (0.027466) | 0.007331 / 0.007986 (-0.000655) | 0.004550 / 0.004328 (0.000222) | 0.077603 / 0.004250 (0.073352) | 0.034307 / 0.037052 (-0.002746) | 0.313174 / 0.258489 (0.054685) | 0.342270 / 0.293841 (0.048429) | 0.033463 / 0.128546 (-0.095083) | 0.011421 / 0.075646 (-0.064225) | 0.317188 / 0.419271 (-0.102083) | 0.040985 / 0.043533 (-0.002548) | 0.300800 / 0.255139 (0.045661) | 0.360171 / 0.283200 (0.076972) | 0.086702 / 0.141683 (-0.054981) | 1.474679 / 1.452155 (0.022525) | 1.518319 / 1.492716 (0.025603) |\n\n### Benchmark: benchmark_getitem\\_100B.json\n\n| metric | get_batch_of\\_1024\\_random_rows | get_batch_of\\_1024\\_rows | get_first_row | get_last_row |\n|--------|---|---|---|---|\n| new / old (diff) | 0.198059 / 0.018006 (0.180052) | 0.403502 / 0.000490 (0.403012) | 0.002663 / 0.000200 (0.002463) | 0.000218 / 0.000054 (0.000164) |\n\n### Benchmark: benchmark_indices_mapping.json\n\n| metric | select | shard | shuffle | sort | train_test_split |\n|--------|---|---|---|---|---|\n| new / old (diff) | 0.022946 / 0.037411 (-0.014465) | 0.096466 / 0.014526 (0.081940) | 0.104092 / 0.176557 (-0.072465) | 0.138499 / 0.737135 (-0.598636) | 0.106941 / 0.296338 (-0.189397) |\n\n### Benchmark: benchmark_iterating.json\n\n| metric | read 5000 | read 50000 | read_batch 50000 10 | read_batch 50000 100 | read_batch 50000 1000 | read_formatted numpy 5000 | read_formatted pandas 5000 | read_formatted tensorflow 5000 | read_formatted torch 5000 | read_formatted_batch numpy 5000 10 | read_formatted_batch numpy 5000 1000 | shuffled read 5000 | shuffled read 50000 | shuffled read_batch 50000 10 | shuffled read_batch 50000 100 | shuffled read_batch 50000 1000 | shuffled read_formatted numpy 5000 | shuffled read_formatted_batch numpy 5000 10 | shuffled read_formatted_batch numpy 5000 1000 |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.416000 / 0.215209 (0.200791) | 4.153120 / 2.077655 (2.075465) | 1.843957 / 1.504120 (0.339837) | 1.650391 / 1.541195 (0.109197) | 1.684765 / 1.468490 (0.216275) | 0.688917 / 4.584777 (-3.895860) | 3.442797 / 3.745712 (-0.302916) | 1.834685 / 5.269862 (-3.435176) | 1.148046 / 4.565676 (-3.417631) | 0.082299 / 0.424275 (-0.341976) | 0.012399 / 0.007607 (0.004792) | 0.521099 / 0.226044 (0.295054) | 5.223695 / 2.268929 (2.954767) | 2.270970 / 55.444624 (-53.173654) | 1.921321 / 6.876477 (-4.955156) | 1.954675 / 2.142072 (-0.187398) | 0.809383 / 4.805227 (-3.995845) | 0.148562 / 6.500664 (-6.352102) | 0.064764 / 0.075469 (-0.010705) |\n\n### Benchmark: benchmark_map_filter.json\n\n| metric | filter | map fast-tokenizer batched | map identity | map identity batched | map no-op batched | map no-op batched numpy | map no-op batched pandas | map no-op batched pytorch | map no-op batched tensorflow |\n|--------|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 1.212687 / 1.841788 (-0.629101) | 13.491641 / 8.074308 (5.417333) | 12.972926 / 10.191392 (2.781534) | 0.137036 / 0.680424 (-0.543388) | 0.028591 / 0.534201 (-0.505610) | 0.391980 / 0.579283 (-0.187303) | 0.394474 / 0.434364 (-0.039889) | 0.456582 / 0.540337 (-0.083755) | 0.535984 / 1.386936 (-0.850952) |\n\n</details>\nPyArrow==latest\n\n<details>\n<summary>Show updated benchmarks!</summary>\n\n### Benchmark: benchmark_array_xd.json\n\n| metric | read_batch_formatted_as_numpy after write_array2d | read_batch_formatted_as_numpy after write_flattened_sequence | read_batch_formatted_as_numpy after write_nested_sequence | read_batch_unformated after write_array2d | read_batch_unformated after write_flattened_sequence | read_batch_unformated after write_nested_sequence | read_col_formatted_as_numpy after write_array2d | read_col_formatted_as_numpy after write_flattened_sequence | read_col_formatted_as_numpy after write_nested_sequence | read_col_unformated after write_array2d | read_col_unformated after write_flattened_sequence | read_col_unformated after write_nested_sequence | read_formatted_as_numpy after write_array2d | read_formatted_as_numpy after write_flattened_sequence | read_formatted_as_numpy after write_nested_sequence | read_unformated after write_array2d | read_unformated after write_flattened_sequence | read_unformated after write_nested_sequence | write_array2d | write_flattened_sequence | write_nested_sequence |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.006419 / 0.011353 (-0.004934) | 0.004295 / 0.011008 (-0.006713) | 0.077702 / 0.038508 (0.039194) | 0.027368 / 0.023109 (0.004259) | 0.336713 / 0.275898 (0.060815) | 0.370074 / 0.323480 (0.046594) | 0.004657 / 0.007986 (-0.003328) | 0.003308 / 0.004328 (-0.001021) | 0.075747 / 0.004250 (0.071496) | 0.037323 / 0.037052 (0.000271) | 0.342382 / 0.258489 (0.083893) | 0.381109 / 0.293841 (0.087269) | 0.031804 / 0.128546 (-0.096742) | 0.011761 / 0.075646 (-0.063885) | 0.086818 / 0.419271 (-0.332454) | 0.042058 / 0.043533 (-0.001475) | 0.346295 / 0.255139 (0.091156) | 0.366857 / 0.283200 (0.083658) | 0.088666 / 0.141683 (-0.053016) | 1.533711 / 1.452155 (0.081556) | 1.537422 / 1.492716 (0.044705) |\n\n### Benchmark: benchmark_getitem\\_100B.json\n\n| metric | get_batch_of\\_1024\\_random_rows | get_batch_of\\_1024\\_rows | get_first_row | get_last_row |\n|--------|---|---|---|---|\n| new / old (diff) | 0.220416 / 0.018006 (0.202410) | 0.387393 / 0.000490 (0.386903) | 0.003739 / 0.000200 (0.003539) | 0.000076 / 0.000054 (0.000021) |\n\n### Benchmark: benchmark_indices_mapping.json\n\n| metric | select | shard | shuffle | sort | train_test_split |\n|--------|---|---|---|---|---|\n| new / old (diff) | 0.024083 / 0.037411 (-0.013329) | 0.098036 / 0.014526 (0.083510) | 0.102908 / 0.176557 (-0.073648) | 0.139512 / 0.737135 (-0.597623) | 0.107703 / 0.296338 (-0.188635) |\n\n### Benchmark: benchmark_iterating.json\n\n| metric | read 5000 | read 50000 | read_batch 50000 10 | read_batch 50000 100 | read_batch 50000 1000 | read_formatted numpy 5000 | read_formatted pandas 5000 | read_formatted tensorflow 5000 | read_formatted torch 5000 | read_formatted_batch numpy 5000 10 | read_formatted_batch numpy 5000 1000 | shuffled read 5000 | shuffled read 50000 | shuffled read_batch 50000 10 | shuffled read_batch 50000 100 | shuffled read_batch 50000 1000 | shuffled read_formatted numpy 5000 | shuffled read_formatted_batch numpy 5000 10 | shuffled read_formatted_batch numpy 5000 1000 |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.437615 / 0.215209 (0.222406) | 4.373140 / 2.077655 (2.295486) | 2.065063 / 1.504120 (0.560943) | 1.863938 / 1.541195 (0.322743) | 1.907955 / 1.468490 (0.439465) | 0.695830 / 4.584777 (-3.888947) | 3.394248 / 3.745712 (-0.351464) | 1.842794 / 5.269862 (-3.427068) | 1.156928 / 4.565676 (-3.408748) | 0.082505 / 0.424275 (-0.341771) | 0.012405 / 0.007607 (0.004798) | 0.538041 / 0.226044 (0.311997) | 5.363508 / 2.268929 (3.094579) | 2.509383 / 55.444624 (-52.935241) | 2.160416 / 6.876477 (-4.716061) | 2.162054 / 2.142072 (0.019982) | 0.802419 / 4.805227 (-4.002809) | 0.150529 / 6.500664 (-6.350135) | 0.066418 / 0.075469 (-0.009051) |\n\n### Benchmark: benchmark_map_filter.json\n\n| metric | filter | map fast-tokenizer batched | map identity | map identity batched | map no-op batched | map no-op batched numpy | map no-op batched pandas | map no-op batched pytorch | map no-op batched tensorflow |\n|--------|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 1.257221 / 1.841788 (-0.584567) | 13.748839 / 8.074308 (5.674531) | 13.310555 / 10.191392 (3.119163) | 0.152997 / 0.680424 (-0.527427) | 0.016618 / 0.534201 (-0.517583) | 0.375443 / 0.579283 (-0.203840) | 0.374942 / 0.434364 (-0.059422) | 0.466704 / 0.540337 (-0.073633) | 0.553563 / 1.386936 (-0.833373) |\n\n</details>\n</details>\n\n![](https://cml.dev/watermark.png#1ac8343af4e2dc6fe0771d0be70eaf8a6e5a8fbc \"CML watermark\")\n", "<details>\n<summary>Show benchmarks</summary>\n\nPyArrow==6.0.0\n\n<details>\n<summary>Show updated benchmarks!</summary>\n\n### Benchmark: benchmark_array_xd.json\n\n| metric | read_batch_formatted_as_numpy after write_array2d | read_batch_formatted_as_numpy after write_flattened_sequence | read_batch_formatted_as_numpy after write_nested_sequence | read_batch_unformated after write_array2d | read_batch_unformated after write_flattened_sequence | read_batch_unformated after write_nested_sequence | read_col_formatted_as_numpy after write_array2d | read_col_formatted_as_numpy after write_flattened_sequence | read_col_formatted_as_numpy after write_nested_sequence | read_col_unformated after write_array2d | read_col_unformated after write_flattened_sequence | read_col_unformated after write_nested_sequence | read_formatted_as_numpy after write_array2d | read_formatted_as_numpy after write_flattened_sequence | read_formatted_as_numpy after write_nested_sequence | read_unformated after write_array2d | read_unformated after write_flattened_sequence | read_unformated after write_nested_sequence | write_array2d | write_flattened_sequence | write_nested_sequence |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.009260 / 0.011353 (-0.002092) | 0.005213 / 0.011008 (-0.005795) | 0.102151 / 0.038508 (0.063643) | 0.035619 / 0.023109 (0.012510) | 0.296266 / 0.275898 (0.020368) | 0.359884 / 0.323480 (0.036404) | 0.008176 / 0.007986 (0.000190) | 0.005031 / 0.004328 (0.000703) | 0.077178 / 0.004250 (0.072927) | 0.041898 / 0.037052 (0.004846) | 0.305640 / 0.258489 (0.047151) | 0.346275 / 0.293841 (0.052434) | 0.037684 / 0.128546 (-0.090863) | 0.011816 / 0.075646 (-0.063831) | 0.334853 / 0.419271 (-0.084419) | 0.046535 / 0.043533 (0.003002) | 0.291544 / 0.255139 (0.036405) | 0.317194 / 0.283200 (0.033994) | 0.103212 / 0.141683 (-0.038471) | 1.424994 / 1.452155 (-0.027161) | 1.486216 / 1.492716 (-0.006501) |\n\n### Benchmark: benchmark_getitem\\_100B.json\n\n| metric | get_batch_of\\_1024\\_random_rows | get_batch_of\\_1024\\_rows | get_first_row | get_last_row |\n|--------|---|---|---|---|\n| new / old (diff) | 0.011816 / 0.018006 (-0.006190) | 0.442092 / 0.000490 (0.441602) | 0.001297 / 0.000200 (0.001097) | 0.000078 / 0.000054 (0.000024) |\n\n### Benchmark: benchmark_indices_mapping.json\n\n| metric | select | shard | shuffle | sort | train_test_split |\n|--------|---|---|---|---|---|\n| new / old (diff) | 0.028277 / 0.037411 (-0.009134) | 0.110431 / 0.014526 (0.095905) | 0.118456 / 0.176557 (-0.058100) | 0.156778 / 0.737135 (-0.580357) | 0.123036 / 0.296338 (-0.173302) |\n\n### Benchmark: benchmark_iterating.json\n\n| metric | read 5000 | read 50000 | read_batch 50000 10 | read_batch 50000 100 | read_batch 50000 1000 | read_formatted numpy 5000 | read_formatted pandas 5000 | read_formatted tensorflow 5000 | read_formatted torch 5000 | read_formatted_batch numpy 5000 10 | read_formatted_batch numpy 5000 1000 | shuffled read 5000 | shuffled read 50000 | shuffled read_batch 50000 10 | shuffled read_batch 50000 100 | shuffled read_batch 50000 1000 | shuffled read_formatted numpy 5000 | shuffled read_formatted_batch numpy 5000 10 | shuffled read_formatted_batch numpy 5000 1000 |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.399006 / 0.215209 (0.183797) | 3.990367 / 2.077655 (1.912712) | 1.798739 / 1.504120 (0.294620) | 1.607133 / 1.541195 (0.065938) | 1.748897 / 1.468490 (0.280407) | 0.690666 / 4.584777 (-3.894111) | 3.795892 / 3.745712 (0.050180) | 3.479317 / 5.269862 (-1.790545) | 1.861268 / 4.565676 (-2.704409) | 0.085235 / 0.424275 (-0.339040) | 0.012997 / 0.007607 (0.005390) | 0.512489 / 0.226044 (0.286445) | 5.039515 / 2.268929 (2.770587) | 2.258079 / 55.444624 (-53.186545) | 1.907178 / 6.876477 (-4.969299) | 1.985953 / 2.142072 (-0.156119) | 0.843595 / 4.805227 (-3.961633) | 0.165286 / 6.500664 (-6.335378) | 0.063026 / 0.075469 (-0.012443) |\n\n### Benchmark: benchmark_map_filter.json\n\n| metric | filter | map fast-tokenizer batched | map identity | map identity batched | map no-op batched | map no-op batched numpy | map no-op batched pandas | map no-op batched pytorch | map no-op batched tensorflow |\n|--------|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 1.186680 / 1.841788 (-0.655108) | 14.976016 / 8.074308 (6.901708) | 14.436941 / 10.191392 (4.245549) | 0.172620 / 0.680424 (-0.507804) | 0.028760 / 0.534201 (-0.505441) | 0.443505 / 0.579283 (-0.135778) | 0.435665 / 0.434364 (0.001301) | 0.520164 / 0.540337 (-0.020174) | 0.608348 / 1.386936 (-0.778588) |\n\n</details>\nPyArrow==latest\n\n<details>\n<summary>Show updated benchmarks!</summary>\n\n### Benchmark: benchmark_array_xd.json\n\n| metric | read_batch_formatted_as_numpy after write_array2d | read_batch_formatted_as_numpy after write_flattened_sequence | read_batch_formatted_as_numpy after write_nested_sequence | read_batch_unformated after write_array2d | read_batch_unformated after write_flattened_sequence | read_batch_unformated after write_nested_sequence | read_col_formatted_as_numpy after write_array2d | read_col_formatted_as_numpy after write_flattened_sequence | read_col_formatted_as_numpy after write_nested_sequence | read_col_unformated after write_array2d | read_col_unformated after write_flattened_sequence | read_col_unformated after write_nested_sequence | read_formatted_as_numpy after write_array2d | read_formatted_as_numpy after write_flattened_sequence | read_formatted_as_numpy after write_nested_sequence | read_unformated after write_array2d | read_unformated after write_flattened_sequence | read_unformated after write_nested_sequence | write_array2d | write_flattened_sequence | write_nested_sequence |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.007510 / 0.011353 (-0.003842) | 0.005012 / 0.011008 (-0.005996) | 0.077865 / 0.038508 (0.039357) | 0.033610 / 0.023109 (0.010500) | 0.365996 / 0.275898 (0.090098) | 0.416393 / 0.323480 (0.092913) | 0.005672 / 0.007986 (-0.002314) | 0.005334 / 0.004328 (0.001006) | 0.074948 / 0.004250 (0.070698) | 0.045962 / 0.037052 (0.008909) | 0.362209 / 0.258489 (0.103719) | 0.410522 / 0.293841 (0.116681) | 0.036247 / 0.128546 (-0.092299) | 0.012432 / 0.075646 (-0.063214) | 0.088754 / 0.419271 (-0.330517) | 0.048848 / 0.043533 (0.005315) | 0.370994 / 0.255139 (0.115855) | 0.382476 / 0.283200 (0.099277) | 0.103443 / 0.141683 (-0.038240) | 1.483127 / 1.452155 (0.030972) | 1.573366 / 1.492716 (0.080650) |\n\n### Benchmark: benchmark_getitem\\_100B.json\n\n| metric | get_batch_of\\_1024\\_random_rows | get_batch_of\\_1024\\_rows | get_first_row | get_last_row |\n|--------|---|---|---|---|\n| new / old (diff) | 0.224163 / 0.018006 (0.206157) | 0.475136 / 0.000490 (0.474646) | 0.000394 / 0.000200 (0.000194) | 0.000057 / 0.000054 (0.000003) |\n\n### Benchmark: benchmark_indices_mapping.json\n\n| metric | select | shard | shuffle | sort | train_test_split |\n|--------|---|---|---|---|---|\n| new / old (diff) | 0.030612 / 0.037411 (-0.006799) | 0.113983 / 0.014526 (0.099457) | 0.121835 / 0.176557 (-0.054722) | 0.160092 / 0.737135 (-0.577043) | 0.127431 / 0.296338 (-0.168908) |\n\n### Benchmark: benchmark_iterating.json\n\n| metric | read 5000 | read 50000 | read_batch 50000 10 | read_batch 50000 100 | read_batch 50000 1000 | read_formatted numpy 5000 | read_formatted pandas 5000 | read_formatted tensorflow 5000 | read_formatted torch 5000 | read_formatted_batch numpy 5000 10 | read_formatted_batch numpy 5000 1000 | shuffled read 5000 | shuffled read 50000 | shuffled read_batch 50000 10 | shuffled read_batch 50000 100 | shuffled read_batch 50000 1000 | shuffled read_formatted numpy 5000 | shuffled read_formatted_batch numpy 5000 10 | shuffled read_formatted_batch numpy 5000 1000 |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.421389 / 0.215209 (0.206179) | 4.207638 / 2.077655 (2.129984) | 2.040265 / 1.504120 (0.536145) | 1.868617 / 1.541195 (0.327422) | 1.979016 / 1.468490 (0.510526) | 0.712499 / 4.584777 (-3.872278) | 3.783091 / 3.745712 (0.037379) | 2.124293 / 5.269862 (-3.145569) | 1.382028 / 4.565676 (-3.183649) | 0.087133 / 0.424275 (-0.337142) | 0.012634 / 0.007607 (0.005027) | 0.518965 / 0.226044 (0.292920) | 5.188330 / 2.268929 (2.919401) | 2.556593 / 55.444624 (-52.888031) | 2.243081 / 6.876477 (-4.633396) | 2.340420 / 2.142072 (0.198347) | 0.858010 / 4.805227 (-3.947218) | 0.169165 / 6.500664 (-6.331499) | 0.065177 / 0.075469 (-0.010292) |\n\n### Benchmark: benchmark_map_filter.json\n\n| metric | filter | map fast-tokenizer batched | map identity | map identity batched | map no-op batched | map no-op batched numpy | map no-op batched pandas | map no-op batched pytorch | map no-op batched tensorflow |\n|--------|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 1.297350 / 1.841788 (-0.544438) | 15.404241 / 8.074308 (7.329933) | 13.806039 / 10.191392 (3.614647) | 0.182055 / 0.680424 (-0.498369) | 0.017789 / 0.534201 (-0.516412) | 0.422828 / 0.579283 (-0.156455) | 0.418269 / 0.434364 (-0.016095) | 0.521561 / 0.540337 (-0.018777) | 0.642526 / 1.386936 (-0.744410) |\n\n</details>\n</details>\n\n![](https://cml.dev/watermark.png#0009eea6819c32a888f65b0fdb5889b6d311c436 \"CML watermark\")\n" ]

[]

[]

[ "Hi! This behavior stems from these lines:\r\n\r\nhttps://github.com/huggingface/datasets/blob/b065547654efa0ec633cf373ac1512884c68b2e1/src/datasets/formatting/np_formatter.py#L45-L46\r\n\r\nI agree we should preserve the original type whenever possible and downcast explicitly with a warning.\r\n\r\n@lhoestq Do you remember why we need this \"default dtype\" logic in our formatters?", "I was also wondering why the default type logic is needed. Me just deleting it is probably too naive of a solution.", "Hmm I think the idea was to end up with the usual default precision for deep learning models - no matter how the data was stored or where it comes from.\r\n\r\nFor example in NLP we store tokens using an optimized low precision to save disk space, but when we set the format to `torch` we actually need to get `int64`. Although the need for a default for integers also comes from numpy not returning the same integer precision depending on your machine. Finally I guess we added a default for floats as well for consistency.\r\n\r\nI'm a bit embarrassed by this though, as a user I'd have expected to get the same precision indeed as well and get a zero copy view.", "Will you fix this or should I open a PR?", "Unfortunately removing it for integers is a breaking change for most `transformers` + `datasets` users for NLP (which is a common case). Removing it for floats is a breaking change for `transformers` + `datasets` for ASR as well. And it also is a breaking change for the other users relying on this behavior.\r\n\r\nTherefore I think that the only short term solution is for the user to provide `dtype=` manually and document better this behavior. We could also extend `dtype` to accept a value that means \"return the same dtype as the underlying storage\" and make it easier to do zero copy.", "@lhoestq It should be fine to remove this conversion in Datasets 3.0, no? For now, we can warn the user (with a log message) about the future change when the default type is changed.", "Let's see with the transformers team if it sounds reasonable ? We'd have to fix multiple example scripts though.\r\n\r\nIf it's not ok we can also explore keeping this behavior only for tokens and audio data.", "IMO being coupled with Transformers can lead to unexpected behavior when one tries to use our lib without pairing it with Transformers, so I think it's still important to \"fix\" this, even if it means we will need to update Transformers' example scripts afterward.\r\n", "Ideally let's update the `transformers` example scripts before the change :P", "For others that run into the same issue: A temporary workaround for me is this:\r\n```python\r\ndef numpy_transform(batch):\r\n return {key: np.asarray(val) for key, val in batch.items()}\r\n\r\ndataset = dataset.with_transform(numpy_transform)\r\n```" ]

[]

[]

[ "Thanks a lot for your help @lhoestq. I've simplified what turned out to be a simple fix and added the unit test.\r\n\r\nDoes this look ready to be merged or is there anything I'm still missing?", "Cool ! I think you just need to remove the unused import in `io/parquet.py`\r\n```\r\nsrc/datasets/io/parquet.py:4:1: F401 'pyarrow as pa' imported but unused\r\n```\r\nand we're good to merge :)", "_The documentation is not available anymore as the PR was closed or merged._", "> Cool ! I think you just need to remove the unused import in `io/parquet.py`\r\n> \r\n> ```\r\n> src/datasets/io/parquet.py:4:1: F401 'pyarrow as pa' imported but unused\r\n> ```\r\n> \r\n> and we're good to merge :)\r\n\r\nDone! Thanks a lot, this was fun :)" ]

[]

[]

[ "_The documentation is not available anymore as the PR was closed or merged._", "The commit also includes the changes to the `DatasetDict` methods or am I missing something?", "Oh, indeed. Feel free to mark the PR as \"Ready for review\" then.", "<details>\n<summary>Show benchmarks</summary>\n\nPyArrow==6.0.0\n\n<details>\n<summary>Show updated benchmarks!</summary>\n\n### Benchmark: benchmark_array_xd.json\n\n| metric | read_batch_formatted_as_numpy after write_array2d | read_batch_formatted_as_numpy after write_flattened_sequence | read_batch_formatted_as_numpy after write_nested_sequence | read_batch_unformated after write_array2d | read_batch_unformated after write_flattened_sequence | read_batch_unformated after write_nested_sequence | read_col_formatted_as_numpy after write_array2d | read_col_formatted_as_numpy after write_flattened_sequence | read_col_formatted_as_numpy after write_nested_sequence | read_col_unformated after write_array2d | read_col_unformated after write_flattened_sequence | read_col_unformated after write_nested_sequence | read_formatted_as_numpy after write_array2d | read_formatted_as_numpy after write_flattened_sequence | read_formatted_as_numpy after write_nested_sequence | read_unformated after write_array2d | read_unformated after write_flattened_sequence | read_unformated after write_nested_sequence | write_array2d | write_flattened_sequence | write_nested_sequence |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.010149 / 0.011353 (-0.001204) | 0.005606 / 0.011008 (-0.005402) | 0.103455 / 0.038508 (0.064947) | 0.042934 / 0.023109 (0.019825) | 0.308365 / 0.275898 (0.032467) | 0.394188 / 0.323480 (0.070708) | 0.008760 / 0.007986 (0.000774) | 0.004567 / 0.004328 (0.000239) | 0.077959 / 0.004250 (0.073708) | 0.050115 / 0.037052 (0.013063) | 0.318009 / 0.258489 (0.059520) | 0.358578 / 0.293841 (0.064737) | 0.039231 / 0.128546 (-0.089315) | 0.012381 / 0.075646 (-0.063265) | 0.340046 / 0.419271 (-0.079226) | 0.048366 / 0.043533 (0.004834) | 0.307643 / 0.255139 (0.052504) | 0.342886 / 0.283200 (0.059687) | 0.109628 / 0.141683 (-0.032055) | 1.457297 / 1.452155 (0.005142) | 1.518067 / 1.492716 (0.025351) |\n\n### Benchmark: benchmark_getitem\\_100B.json\n\n| metric | get_batch_of\\_1024\\_random_rows | get_batch_of\\_1024\\_rows | get_first_row | get_last_row |\n|--------|---|---|---|---|\n| new / old (diff) | 0.295590 / 0.018006 (0.277584) | 0.531515 / 0.000490 (0.531026) | 0.005677 / 0.000200 (0.005477) | 0.000095 / 0.000054 (0.000041) |\n\n### Benchmark: benchmark_indices_mapping.json\n\n| metric | select | shard | shuffle | sort | train_test_split |\n|--------|---|---|---|---|---|\n| new / old (diff) | 0.030901 / 0.037411 (-0.006511) | 0.118312 / 0.014526 (0.103786) | 0.123146 / 0.176557 (-0.053410) | 0.163608 / 0.737135 (-0.573527) | 0.128604 / 0.296338 (-0.167734) |\n\n### Benchmark: benchmark_iterating.json\n\n| metric | read 5000 | read 50000 | read_batch 50000 10 | read_batch 50000 100 | read_batch 50000 1000 | read_formatted numpy 5000 | read_formatted pandas 5000 | read_formatted tensorflow 5000 | read_formatted torch 5000 | read_formatted_batch numpy 5000 10 | read_formatted_batch numpy 5000 1000 | shuffled read 5000 | shuffled read 50000 | shuffled read_batch 50000 10 | shuffled read_batch 50000 100 | shuffled read_batch 50000 1000 | shuffled read_formatted numpy 5000 | shuffled read_formatted_batch numpy 5000 10 | shuffled read_formatted_batch numpy 5000 1000 |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.404143 / 0.215209 (0.188934) | 4.000118 / 2.077655 (1.922464) | 1.804502 / 1.504120 (0.300382) | 1.597287 / 1.541195 (0.056093) | 1.738512 / 1.468490 (0.270022) | 0.704658 / 4.584777 (-3.880119) | 3.830101 / 3.745712 (0.084389) | 2.186598 / 5.269862 (-3.083263) | 1.367873 / 4.565676 (-3.197804) | 0.085550 / 0.424275 (-0.338725) | 0.012226 / 0.007607 (0.004619) | 0.505760 / 0.226044 (0.279716) | 5.054583 / 2.268929 (2.785655) | 2.284942 / 55.444624 (-53.159682) | 1.961413 / 6.876477 (-4.915064) | 2.059449 / 2.142072 (-0.082623) | 0.845009 / 4.805227 (-3.960218) | 0.167204 / 6.500664 (-6.333460) | 0.065998 / 0.075469 (-0.009471) |\n\n### Benchmark: benchmark_map_filter.json\n\n| metric | filter | map fast-tokenizer batched | map identity | map identity batched | map no-op batched | map no-op batched numpy | map no-op batched pandas | map no-op batched pytorch | map no-op batched tensorflow |\n|--------|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 1.221861 / 1.841788 (-0.619927) | 15.925213 / 8.074308 (7.850905) | 15.359308 / 10.191392 (5.167916) | 0.171776 / 0.680424 (-0.508648) | 0.029234 / 0.534201 (-0.504967) | 0.446349 / 0.579283 (-0.132934) | 0.447873 / 0.434364 (0.013509) | 0.527400 / 0.540337 (-0.012937) | 0.610208 / 1.386936 (-0.776728) |\n\n</details>\nPyArrow==latest\n\n<details>\n<summary>Show updated benchmarks!</summary>\n\n### Benchmark: benchmark_array_xd.json\n\n| metric | read_batch_formatted_as_numpy after write_array2d | read_batch_formatted_as_numpy after write_flattened_sequence | read_batch_formatted_as_numpy after write_nested_sequence | read_batch_unformated after write_array2d | read_batch_unformated after write_flattened_sequence | read_batch_unformated after write_nested_sequence | read_col_formatted_as_numpy after write_array2d | read_col_formatted_as_numpy after write_flattened_sequence | read_col_formatted_as_numpy after write_nested_sequence | read_col_unformated after write_array2d | read_col_unformated after write_flattened_sequence | read_col_unformated after write_nested_sequence | read_formatted_as_numpy after write_array2d | read_formatted_as_numpy after write_flattened_sequence | read_formatted_as_numpy after write_nested_sequence | read_unformated after write_array2d | read_unformated after write_flattened_sequence | read_unformated after write_nested_sequence | write_array2d | write_flattened_sequence | write_nested_sequence |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.008030 / 0.011353 (-0.003323) | 0.005686 / 0.011008 (-0.005322) | 0.076204 / 0.038508 (0.037696) | 0.037131 / 0.023109 (0.014022) | 0.341461 / 0.275898 (0.065563) | 0.378734 / 0.323480 (0.055255) | 0.006580 / 0.007986 (-0.001406) | 0.004379 / 0.004328 (0.000050) | 0.073983 / 0.004250 (0.069732) | 0.055895 / 0.037052 (0.018842) | 0.342667 / 0.258489 (0.084178) | 0.401464 / 0.293841 (0.107623) | 0.037710 / 0.128546 (-0.090837) | 0.012604 / 0.075646 (-0.063042) | 0.087563 / 0.419271 (-0.331709) | 0.050887 / 0.043533 (0.007354) | 0.333491 / 0.255139 (0.078352) | 0.357437 / 0.283200 (0.074237) | 0.109566 / 0.141683 (-0.032117) | 1.423372 / 1.452155 (-0.028783) | 1.569423 / 1.492716 (0.076706) |\n\n### Benchmark: benchmark_getitem\\_100B.json\n\n| metric | get_batch_of\\_1024\\_random_rows | get_batch_of\\_1024\\_rows | get_first_row | get_last_row |\n|--------|---|---|---|---|\n| new / old (diff) | 0.340986 / 0.018006 (0.322980) | 0.530885 / 0.000490 (0.530395) | 0.004172 / 0.000200 (0.003972) | 0.000115 / 0.000054 (0.000060) |\n\n### Benchmark: benchmark_indices_mapping.json\n\n| metric | select | shard | shuffle | sort | train_test_split |\n|--------|---|---|---|---|---|\n| new / old (diff) | 0.030424 / 0.037411 (-0.006987) | 0.121191 / 0.014526 (0.106666) | 0.129066 / 0.176557 (-0.047491) | 0.166938 / 0.737135 (-0.570198) | 0.132000 / 0.296338 (-0.164338) |\n\n### Benchmark: benchmark_iterating.json\n\n| metric | read 5000 | read 50000 | read_batch 50000 10 | read_batch 50000 100 | read_batch 50000 1000 | read_formatted numpy 5000 | read_formatted pandas 5000 | read_formatted tensorflow 5000 | read_formatted torch 5000 | read_formatted_batch numpy 5000 10 | read_formatted_batch numpy 5000 1000 | shuffled read 5000 | shuffled read 50000 | shuffled read_batch 50000 10 | shuffled read_batch 50000 100 | shuffled read_batch 50000 1000 | shuffled read_formatted numpy 5000 | shuffled read_formatted_batch numpy 5000 10 | shuffled read_formatted_batch numpy 5000 1000 |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.418718 / 0.215209 (0.203509) | 4.163973 / 2.077655 (2.086318) | 1.982665 / 1.504120 (0.478545) | 1.798866 / 1.541195 (0.257671) | 1.918867 / 1.468490 (0.450377) | 0.724634 / 4.584777 (-3.860143) | 3.864549 / 3.745712 (0.118837) | 3.697768 / 5.269862 (-1.572093) | 1.983942 / 4.565676 (-2.581735) | 0.086818 / 0.424275 (-0.337457) | 0.012336 / 0.007607 (0.004728) | 0.522314 / 0.226044 (0.296269) | 5.216813 / 2.268929 (2.947884) | 2.516187 / 55.444624 (-52.928437) | 2.172057 / 6.876477 (-4.704420) | 2.342773 / 2.142072 (0.200701) | 0.851805 / 4.805227 (-3.953422) | 0.170139 / 6.500664 (-6.330525) | 0.068494 / 0.075469 (-0.006975) |\n\n### Benchmark: benchmark_map_filter.json\n\n| metric | filter | map fast-tokenizer batched | map identity | map identity batched | map no-op batched | map no-op batched numpy | map no-op batched pandas | map no-op batched pytorch | map no-op batched tensorflow |\n|--------|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 1.307370 / 1.841788 (-0.534418) | 16.737937 / 8.074308 (8.663629) | 14.483384 / 10.191392 (4.291992) | 0.172418 / 0.680424 (-0.508006) | 0.018241 / 0.534201 (-0.515960) | 0.432049 / 0.579283 (-0.147234) | 0.447590 / 0.434364 (0.013227) | 0.550332 / 0.540337 (0.009994) | 0.646756 / 1.386936 (-0.740180) |\n\n</details>\n</details>\n\n![](https://cml.dev/watermark.png#819bc6e9f88459f363e6fb6948e9cbe5c231500d \"CML watermark\")\n" ]

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

[]

[ "Hi, thanks for noticing this! We can't just remove the cache control as this allows us to control where the arrow files generated by the ops are written (cached on disk if enabled or a temporary directory if disabled). The right way to address this inconsistency would be by having `load_from_cache_file=None` by default everywhere.", "Hi! Yes, this seems more plausible. I can implement that. One last thing is the type annotation `load_from_cache_file: bool = None`. Which I then would change to `load_from_cache_file: Optional[bool] = None`.", "PR #5515 ", "Yes, `Optional[bool]` is the correct type annotation and thanks for the PR." ]

[]

[]

[ "Hi! Let's not do this - renaming it would be a breaking change, and going through the deprecation cycle is only worth it if it improves user experience.", "Hi @mariosasko, ok it makes sense. Anyway, don't you think it's worth it at some point to start a deprecation cycle e.g. `fs` in `load_from_disk`? It doesn't affect user experience but it's for sure a bad practice IMO, but's up to you 😄 Feel free to close this issue otherwise!" ]

[]

[]

[ "_The documentation is not available anymore as the PR was closed or merged._", "<details>\n<summary>Show benchmarks</summary>\n\nPyArrow==6.0.0\n\n<details>\n<summary>Show updated benchmarks!</summary>\n\n### Benchmark: benchmark_array_xd.json\n\n| metric | read_batch_formatted_as_numpy after write_array2d | read_batch_formatted_as_numpy after write_flattened_sequence | read_batch_formatted_as_numpy after write_nested_sequence | read_batch_unformated after write_array2d | read_batch_unformated after write_flattened_sequence | read_batch_unformated after write_nested_sequence | read_col_formatted_as_numpy after write_array2d | read_col_formatted_as_numpy after write_flattened_sequence | read_col_formatted_as_numpy after write_nested_sequence | read_col_unformated after write_array2d | read_col_unformated after write_flattened_sequence | read_col_unformated after write_nested_sequence | read_formatted_as_numpy after write_array2d | read_formatted_as_numpy after write_flattened_sequence | read_formatted_as_numpy after write_nested_sequence | read_unformated after write_array2d | read_unformated after write_flattened_sequence | read_unformated after write_nested_sequence | write_array2d | write_flattened_sequence | write_nested_sequence |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.008882 / 0.011353 (-0.002471) | 0.004562 / 0.011008 (-0.006446) | 0.100035 / 0.038508 (0.061527) | 0.030654 / 0.023109 (0.007545) | 0.298745 / 0.275898 (0.022847) | 0.356869 / 0.323480 (0.033389) | 0.007170 / 0.007986 (-0.000815) | 0.003471 / 0.004328 (-0.000858) | 0.077975 / 0.004250 (0.073725) | 0.037861 / 0.037052 (0.000809) | 0.311643 / 0.258489 (0.053154) | 0.343504 / 0.293841 (0.049663) | 0.033768 / 0.128546 (-0.094778) | 0.011342 / 0.075646 (-0.064304) | 0.323953 / 0.419271 (-0.095319) | 0.040818 / 0.043533 (-0.002715) | 0.298492 / 0.255139 (0.043353) | 0.327292 / 0.283200 (0.044092) | 0.088423 / 0.141683 (-0.053260) | 1.489520 / 1.452155 (0.037366) | 1.532962 / 1.492716 (0.040245) |\n\n### Benchmark: benchmark_getitem\\_100B.json\n\n| metric | get_batch_of\\_1024\\_random_rows | get_batch_of\\_1024\\_rows | get_first_row | get_last_row |\n|--------|---|---|---|---|\n| new / old (diff) | 0.223654 / 0.018006 (0.205647) | 0.415134 / 0.000490 (0.414644) | 0.007394 / 0.000200 (0.007194) | 0.000080 / 0.000054 (0.000026) |\n\n### Benchmark: benchmark_indices_mapping.json\n\n| metric | select | shard | shuffle | sort | train_test_split |\n|--------|---|---|---|---|---|\n| new / old (diff) | 0.023616 / 0.037411 (-0.013795) | 0.096652 / 0.014526 (0.082126) | 0.105239 / 0.176557 (-0.071318) | 0.148637 / 0.737135 (-0.588498) | 0.107937 / 0.296338 (-0.188402) |\n\n### Benchmark: benchmark_iterating.json\n\n| metric | read 5000 | read 50000 | read_batch 50000 10 | read_batch 50000 100 | read_batch 50000 1000 | read_formatted numpy 5000 | read_formatted pandas 5000 | read_formatted tensorflow 5000 | read_formatted torch 5000 | read_formatted_batch numpy 5000 10 | read_formatted_batch numpy 5000 1000 | shuffled read 5000 | shuffled read 50000 | shuffled read_batch 50000 10 | shuffled read_batch 50000 100 | shuffled read_batch 50000 1000 | shuffled read_formatted numpy 5000 | shuffled read_formatted_batch numpy 5000 10 | shuffled read_formatted_batch numpy 5000 1000 |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.426816 / 0.215209 (0.211607) | 4.241533 / 2.077655 (2.163878) | 1.946493 / 1.504120 (0.442373) | 1.735765 / 1.541195 (0.194570) | 1.781424 / 1.468490 (0.312934) | 0.688082 / 4.584777 (-3.896694) | 3.396444 / 3.745712 (-0.349268) | 1.920333 / 5.269862 (-3.349528) | 1.293833 / 4.565676 (-3.271843) | 0.081967 / 0.424275 (-0.342308) | 0.012911 / 0.007607 (0.005304) | 0.536928 / 0.226044 (0.310884) | 5.452327 / 2.268929 (3.183399) | 2.505785 / 55.444624 (-52.938840) | 2.173627 / 6.876477 (-4.702850) | 2.119978 / 2.142072 (-0.022095) | 0.809012 / 4.805227 (-3.996215) | 0.149124 / 6.500664 (-6.351540) | 0.066008 / 0.075469 (-0.009461) |\n\n### Benchmark: benchmark_map_filter.json\n\n| metric | filter | map fast-tokenizer batched | map identity | map identity batched | map no-op batched | map no-op batched numpy | map no-op batched pandas | map no-op batched pytorch | map no-op batched tensorflow |\n|--------|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 1.215702 / 1.841788 (-0.626085) | 13.757525 / 8.074308 (5.683217) | 13.999208 / 10.191392 (3.807816) | 0.164875 / 0.680424 (-0.515549) | 0.028517 / 0.534201 (-0.505684) | 0.394829 / 0.579283 (-0.184454) | 0.404962 / 0.434364 (-0.029401) | 0.484455 / 0.540337 (-0.055882) | 0.575008 / 1.386936 (-0.811928) |\n\n</details>\nPyArrow==latest\n\n<details>\n<summary>Show updated benchmarks!</summary>\n\n### Benchmark: benchmark_array_xd.json\n\n| metric | read_batch_formatted_as_numpy after write_array2d | read_batch_formatted_as_numpy after write_flattened_sequence | read_batch_formatted_as_numpy after write_nested_sequence | read_batch_unformated after write_array2d | read_batch_unformated after write_flattened_sequence | read_batch_unformated after write_nested_sequence | read_col_formatted_as_numpy after write_array2d | read_col_formatted_as_numpy after write_flattened_sequence | read_col_formatted_as_numpy after write_nested_sequence | read_col_unformated after write_array2d | read_col_unformated after write_flattened_sequence | read_col_unformated after write_nested_sequence | read_formatted_as_numpy after write_array2d | read_formatted_as_numpy after write_flattened_sequence | read_formatted_as_numpy after write_nested_sequence | read_unformated after write_array2d | read_unformated after write_flattened_sequence | read_unformated after write_nested_sequence | write_array2d | write_flattened_sequence | write_nested_sequence |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.006754 / 0.011353 (-0.004598) | 0.004579 / 0.011008 (-0.006430) | 0.076617 / 0.038508 (0.038109) | 0.027902 / 0.023109 (0.004793) | 0.346278 / 0.275898 (0.070380) | 0.398060 / 0.323480 (0.074580) | 0.004938 / 0.007986 (-0.003047) | 0.004681 / 0.004328 (0.000353) | 0.076336 / 0.004250 (0.072086) | 0.038018 / 0.037052 (0.000966) | 0.358701 / 0.258489 (0.100212) | 0.408413 / 0.293841 (0.114572) | 0.031772 / 0.128546 (-0.096774) | 0.011604 / 0.075646 (-0.064042) | 0.085964 / 0.419271 (-0.333308) | 0.042030 / 0.043533 (-0.001502) | 0.343568 / 0.255139 (0.088429) | 0.381805 / 0.283200 (0.098605) | 0.090759 / 0.141683 (-0.050924) | 1.504553 / 1.452155 (0.052398) | 1.594006 / 1.492716 (0.101289) |\n\n### Benchmark: benchmark_getitem\\_100B.json\n\n| metric | get_batch_of\\_1024\\_random_rows | get_batch_of\\_1024\\_rows | get_first_row | get_last_row |\n|--------|---|---|---|---|\n| new / old (diff) | 0.227395 / 0.018006 (0.209389) | 0.403097 / 0.000490 (0.402608) | 0.000413 / 0.000200 (0.000213) | 0.000060 / 0.000054 (0.000006) |\n\n### Benchmark: benchmark_indices_mapping.json\n\n| metric | select | shard | shuffle | sort | train_test_split |\n|--------|---|---|---|---|---|\n| new / old (diff) | 0.024693 / 0.037411 (-0.012718) | 0.100470 / 0.014526 (0.085944) | 0.108481 / 0.176557 (-0.068076) | 0.142791 / 0.737135 (-0.594345) | 0.109949 / 0.296338 (-0.186389) |\n\n### Benchmark: benchmark_iterating.json\n\n| metric | read 5000 | read 50000 | read_batch 50000 10 | read_batch 50000 100 | read_batch 50000 1000 | read_formatted numpy 5000 | read_formatted pandas 5000 | read_formatted tensorflow 5000 | read_formatted torch 5000 | read_formatted_batch numpy 5000 10 | read_formatted_batch numpy 5000 1000 | shuffled read 5000 | shuffled read 50000 | shuffled read_batch 50000 10 | shuffled read_batch 50000 100 | shuffled read_batch 50000 1000 | shuffled read_formatted numpy 5000 | shuffled read_formatted_batch numpy 5000 10 | shuffled read_formatted_batch numpy 5000 1000 |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.443674 / 0.215209 (0.228465) | 4.412207 / 2.077655 (2.334553) | 2.073752 / 1.504120 (0.569632) | 1.863153 / 1.541195 (0.321958) | 1.940063 / 1.468490 (0.471573) | 0.696456 / 4.584777 (-3.888321) | 3.422120 / 3.745712 (-0.323592) | 1.902579 / 5.269862 (-3.367282) | 1.184948 / 4.565676 (-3.380729) | 0.083079 / 0.424275 (-0.341196) | 0.012649 / 0.007607 (0.005042) | 0.542035 / 0.226044 (0.315991) | 5.421826 / 2.268929 (3.152897) | 2.525092 / 55.444624 (-52.919532) | 2.177144 / 6.876477 (-4.699332) | 2.225224 / 2.142072 (0.083151) | 0.804739 / 4.805227 (-4.000488) | 0.151000 / 6.500664 (-6.349664) | 0.066987 / 0.075469 (-0.008482) |\n\n### Benchmark: benchmark_map_filter.json\n\n| metric | filter | map fast-tokenizer batched | map identity | map identity batched | map no-op batched | map no-op batched numpy | map no-op batched pandas | map no-op batched pytorch | map no-op batched tensorflow |\n|--------|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 1.277199 / 1.841788 (-0.564589) | 14.184146 / 8.074308 (6.109838) | 13.413348 / 10.191392 (3.221956) | 0.128551 / 0.680424 (-0.551872) | 0.016461 / 0.534201 (-0.517740) | 0.379963 / 0.579283 (-0.199320) | 0.381350 / 0.434364 (-0.053014) | 0.439044 / 0.540337 (-0.101293) | 0.521559 / 1.386936 (-0.865377) |\n\n</details>\n</details>\n\n![](https://cml.dev/watermark.png#4f3c152c1c35df250d2fbeb25d5823a65714f2d8 \"CML watermark\")\n", "<details>\n<summary>Show benchmarks</summary>\n\nPyArrow==6.0.0\n\n<details>\n<summary>Show updated benchmarks!</summary>\n\n### Benchmark: benchmark_array_xd.json\n\n| metric | read_batch_formatted_as_numpy after write_array2d | read_batch_formatted_as_numpy after write_flattened_sequence | read_batch_formatted_as_numpy after write_nested_sequence | read_batch_unformated after write_array2d | read_batch_unformated after write_flattened_sequence | read_batch_unformated after write_nested_sequence | read_col_formatted_as_numpy after write_array2d | read_col_formatted_as_numpy after write_flattened_sequence | read_col_formatted_as_numpy after write_nested_sequence | read_col_unformated after write_array2d | read_col_unformated after write_flattened_sequence | read_col_unformated after write_nested_sequence | read_formatted_as_numpy after write_array2d | read_formatted_as_numpy after write_flattened_sequence | read_formatted_as_numpy after write_nested_sequence | read_unformated after write_array2d | read_unformated after write_flattened_sequence | read_unformated after write_nested_sequence | write_array2d | write_flattened_sequence | write_nested_sequence |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.008876 / 0.011353 (-0.002477) | 0.004629 / 0.011008 (-0.006379) | 0.101697 / 0.038508 (0.063189) | 0.030373 / 0.023109 (0.007264) | 0.302206 / 0.275898 (0.026308) | 0.365835 / 0.323480 (0.042355) | 0.007877 / 0.007986 (-0.000109) | 0.004473 / 0.004328 (0.000144) | 0.077334 / 0.004250 (0.073084) | 0.038066 / 0.037052 (0.001014) | 0.308064 / 0.258489 (0.049575) | 0.347329 / 0.293841 (0.053488) | 0.034478 / 0.128546 (-0.094068) | 0.011651 / 0.075646 (-0.063995) | 0.323481 / 0.419271 (-0.095791) | 0.043515 / 0.043533 (-0.000018) | 0.299885 / 0.255139 (0.044746) | 0.328959 / 0.283200 (0.045760) | 0.095308 / 0.141683 (-0.046375) | 1.474058 / 1.452155 (0.021903) | 1.535335 / 1.492716 (0.042619) |\n\n### Benchmark: benchmark_getitem\\_100B.json\n\n| metric | get_batch_of\\_1024\\_random_rows | get_batch_of\\_1024\\_rows | get_first_row | get_last_row |\n|--------|---|---|---|---|\n| new / old (diff) | 0.197416 / 0.018006 (0.179410) | 0.421935 / 0.000490 (0.421446) | 0.003490 / 0.000200 (0.003290) | 0.000074 / 0.000054 (0.000020) |\n\n### Benchmark: benchmark_indices_mapping.json\n\n| metric | select | shard | shuffle | sort | train_test_split |\n|--------|---|---|---|---|---|\n| new / old (diff) | 0.024519 / 0.037411 (-0.012892) | 0.100710 / 0.014526 (0.086185) | 0.104520 / 0.176557 (-0.072036) | 0.142048 / 0.737135 (-0.595087) | 0.109274 / 0.296338 (-0.187064) |\n\n### Benchmark: benchmark_iterating.json\n\n| metric | read 5000 | read 50000 | read_batch 50000 10 | read_batch 50000 100 | read_batch 50000 1000 | read_formatted numpy 5000 | read_formatted pandas 5000 | read_formatted tensorflow 5000 | read_formatted torch 5000 | read_formatted_batch numpy 5000 10 | read_formatted_batch numpy 5000 1000 | shuffled read 5000 | shuffled read 50000 | shuffled read_batch 50000 10 | shuffled read_batch 50000 100 | shuffled read_batch 50000 1000 | shuffled read_formatted numpy 5000 | shuffled read_formatted_batch numpy 5000 10 | shuffled read_formatted_batch numpy 5000 1000 |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.408766 / 0.215209 (0.193557) | 4.101720 / 2.077655 (2.024065) | 1.812375 / 1.504120 (0.308256) | 1.605819 / 1.541195 (0.064624) | 1.688923 / 1.468490 (0.220433) | 0.691198 / 4.584777 (-3.893579) | 3.422137 / 3.745712 (-0.323575) | 1.921318 / 5.269862 (-3.348544) | 1.168770 / 4.565676 (-3.396906) | 0.082840 / 0.424275 (-0.341435) | 0.012740 / 0.007607 (0.005133) | 0.524333 / 0.226044 (0.298289) | 5.258077 / 2.268929 (2.989149) | 2.273177 / 55.444624 (-53.171447) | 1.931919 / 6.876477 (-4.944558) | 1.988415 / 2.142072 (-0.153658) | 0.812227 / 4.805227 (-3.993000) | 0.150043 / 6.500664 (-6.350622) | 0.066422 / 0.075469 (-0.009047) |\n\n### Benchmark: benchmark_map_filter.json\n\n| metric | filter | map fast-tokenizer batched | map identity | map identity batched | map no-op batched | map no-op batched numpy | map no-op batched pandas | map no-op batched pytorch | map no-op batched tensorflow |\n|--------|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 1.188069 / 1.841788 (-0.653718) | 13.942681 / 8.074308 (5.868373) | 14.104658 / 10.191392 (3.913266) | 0.151966 / 0.680424 (-0.528458) | 0.028833 / 0.534201 (-0.505368) | 0.395125 / 0.579283 (-0.184158) | 0.408512 / 0.434364 (-0.025852) | 0.487587 / 0.540337 (-0.052751) | 0.570023 / 1.386936 (-0.816913) |\n\n</details>\nPyArrow==latest\n\n<details>\n<summary>Show updated benchmarks!</summary>\n\n### Benchmark: benchmark_array_xd.json\n\n| metric | read_batch_formatted_as_numpy after write_array2d | read_batch_formatted_as_numpy after write_flattened_sequence | read_batch_formatted_as_numpy after write_nested_sequence | read_batch_unformated after write_array2d | read_batch_unformated after write_flattened_sequence | read_batch_unformated after write_nested_sequence | read_col_formatted_as_numpy after write_array2d | read_col_formatted_as_numpy after write_flattened_sequence | read_col_formatted_as_numpy after write_nested_sequence | read_col_unformated after write_array2d | read_col_unformated after write_flattened_sequence | read_col_unformated after write_nested_sequence | read_formatted_as_numpy after write_array2d | read_formatted_as_numpy after write_flattened_sequence | read_formatted_as_numpy after write_nested_sequence | read_unformated after write_array2d | read_unformated after write_flattened_sequence | read_unformated after write_nested_sequence | write_array2d | write_flattened_sequence | write_nested_sequence |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.006860 / 0.011353 (-0.004493) | 0.004582 / 0.011008 (-0.006426) | 0.079902 / 0.038508 (0.041394) | 0.027565 / 0.023109 (0.004456) | 0.341393 / 0.275898 (0.065495) | 0.378911 / 0.323480 (0.055431) | 0.005847 / 0.007986 (-0.002138) | 0.004681 / 0.004328 (0.000353) | 0.079422 / 0.004250 (0.075171) | 0.039135 / 0.037052 (0.002083) | 0.342026 / 0.258489 (0.083537) | 0.387510 / 0.293841 (0.093669) | 0.031999 / 0.128546 (-0.096547) | 0.011782 / 0.075646 (-0.063865) | 0.088563 / 0.419271 (-0.330709) | 0.042435 / 0.043533 (-0.001098) | 0.343055 / 0.255139 (0.087916) | 0.367437 / 0.283200 (0.084237) | 0.091578 / 0.141683 (-0.050104) | 1.506828 / 1.452155 (0.054673) | 1.599590 / 1.492716 (0.106874) |\n\n### Benchmark: benchmark_getitem\\_100B.json\n\n| metric | get_batch_of\\_1024\\_random_rows | get_batch_of\\_1024\\_rows | get_first_row | get_last_row |\n|--------|---|---|---|---|\n| new / old (diff) | 0.217939 / 0.018006 (0.199932) | 0.408352 / 0.000490 (0.407863) | 0.000394 / 0.000200 (0.000194) | 0.000063 / 0.000054 (0.000009) |\n\n### Benchmark: benchmark_indices_mapping.json\n\n| metric | select | shard | shuffle | sort | train_test_split |\n|--------|---|---|---|---|---|\n| new / old (diff) | 0.026344 / 0.037411 (-0.011067) | 0.102968 / 0.014526 (0.088442) | 0.110340 / 0.176557 (-0.066217) | 0.145696 / 0.737135 (-0.591439) | 0.111632 / 0.296338 (-0.184707) |\n\n### Benchmark: benchmark_iterating.json\n\n| metric | read 5000 | read 50000 | read_batch 50000 10 | read_batch 50000 100 | read_batch 50000 1000 | read_formatted numpy 5000 | read_formatted pandas 5000 | read_formatted tensorflow 5000 | read_formatted torch 5000 | read_formatted_batch numpy 5000 10 | read_formatted_batch numpy 5000 1000 | shuffled read 5000 | shuffled read 50000 | shuffled read_batch 50000 10 | shuffled read_batch 50000 100 | shuffled read_batch 50000 1000 | shuffled read_formatted numpy 5000 | shuffled read_formatted_batch numpy 5000 10 | shuffled read_formatted_batch numpy 5000 1000 |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.440764 / 0.215209 (0.225555) | 4.423179 / 2.077655 (2.345524) | 2.057016 / 1.504120 (0.552896) | 1.848741 / 1.541195 (0.307546) | 1.939827 / 1.468490 (0.471337) | 0.699370 / 4.584777 (-3.885407) | 3.472521 / 3.745712 (-0.273191) | 3.232557 / 5.269862 (-2.037305) | 1.755534 / 4.565676 (-2.810143) | 0.083469 / 0.424275 (-0.340807) | 0.012980 / 0.007607 (0.005373) | 0.557662 / 0.226044 (0.331618) | 5.435657 / 2.268929 (3.166729) | 2.545106 / 55.444624 (-52.899519) | 2.168047 / 6.876477 (-4.708430) | 2.234070 / 2.142072 (0.091997) | 0.804662 / 4.805227 (-4.000565) | 0.152832 / 6.500664 (-6.347833) | 0.069372 / 0.075469 (-0.006097) |\n\n### Benchmark: benchmark_map_filter.json\n\n| metric | filter | map fast-tokenizer batched | map identity | map identity batched | map no-op batched | map no-op batched numpy | map no-op batched pandas | map no-op batched pytorch | map no-op batched tensorflow |\n|--------|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 1.299189 / 1.841788 (-0.542598) | 14.752880 / 8.074308 (6.678572) | 13.607676 / 10.191392 (3.416284) | 0.150773 / 0.680424 (-0.529650) | 0.016701 / 0.534201 (-0.517500) | 0.379507 / 0.579283 (-0.199776) | 0.389401 / 0.434364 (-0.044963) | 0.444199 / 0.540337 (-0.096139) | 0.524264 / 1.386936 (-0.862672) |\n\n</details>\n</details>\n\n![](https://cml.dev/watermark.png#12be850b36c0b9d4841af86c75e08c0a726ffb5c \"CML watermark\")\n", "<details>\n<summary>Show benchmarks</summary>\n\nPyArrow==6.0.0\n\n<details>\n<summary>Show updated benchmarks!</summary>\n\n### Benchmark: benchmark_array_xd.json\n\n| metric | read_batch_formatted_as_numpy after write_array2d | read_batch_formatted_as_numpy after write_flattened_sequence | read_batch_formatted_as_numpy after write_nested_sequence | read_batch_unformated after write_array2d | read_batch_unformated after write_flattened_sequence | read_batch_unformated after write_nested_sequence | read_col_formatted_as_numpy after write_array2d | read_col_formatted_as_numpy after write_flattened_sequence | read_col_formatted_as_numpy after write_nested_sequence | read_col_unformated after write_array2d | read_col_unformated after write_flattened_sequence | read_col_unformated after write_nested_sequence | read_formatted_as_numpy after write_array2d | read_formatted_as_numpy after write_flattened_sequence | read_formatted_as_numpy after write_nested_sequence | read_unformated after write_array2d | read_unformated after write_flattened_sequence | read_unformated after write_nested_sequence | write_array2d | write_flattened_sequence | write_nested_sequence |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.008694 / 0.011353 (-0.002659) | 0.004549 / 0.011008 (-0.006459) | 0.101164 / 0.038508 (0.062656) | 0.029644 / 0.023109 (0.006535) | 0.294849 / 0.275898 (0.018950) | 0.366755 / 0.323480 (0.043275) | 0.007205 / 0.007986 (-0.000780) | 0.004255 / 0.004328 (-0.000074) | 0.077433 / 0.004250 (0.073183) | 0.038024 / 0.037052 (0.000972) | 0.310380 / 0.258489 (0.051891) | 0.347093 / 0.293841 (0.053252) | 0.033232 / 0.128546 (-0.095314) | 0.011404 / 0.075646 (-0.064242) | 0.323341 / 0.419271 (-0.095930) | 0.040586 / 0.043533 (-0.002946) | 0.296083 / 0.255139 (0.040944) | 0.321870 / 0.283200 (0.038671) | 0.087377 / 0.141683 (-0.054306) | 1.466869 / 1.452155 (0.014715) | 1.514763 / 1.492716 (0.022046) |\n\n### Benchmark: benchmark_getitem\\_100B.json\n\n| metric | get_batch_of\\_1024\\_random_rows | get_batch_of\\_1024\\_rows | get_first_row | get_last_row |\n|--------|---|---|---|---|\n| new / old (diff) | 0.010272 / 0.018006 (-0.007734) | 0.414645 / 0.000490 (0.414155) | 0.003730 / 0.000200 (0.003530) | 0.000076 / 0.000054 (0.000021) |\n\n### Benchmark: benchmark_indices_mapping.json\n\n| metric | select | shard | shuffle | sort | train_test_split |\n|--------|---|---|---|---|---|\n| new / old (diff) | 0.024093 / 0.037411 (-0.013318) | 0.098718 / 0.014526 (0.084192) | 0.105526 / 0.176557 (-0.071030) | 0.141578 / 0.737135 (-0.595557) | 0.109679 / 0.296338 (-0.186660) |\n\n### Benchmark: benchmark_iterating.json\n\n| metric | read 5000 | read 50000 | read_batch 50000 10 | read_batch 50000 100 | read_batch 50000 1000 | read_formatted numpy 5000 | read_formatted pandas 5000 | read_formatted tensorflow 5000 | read_formatted torch 5000 | read_formatted_batch numpy 5000 10 | read_formatted_batch numpy 5000 1000 | shuffled read 5000 | shuffled read 50000 | shuffled read_batch 50000 10 | shuffled read_batch 50000 100 | shuffled read_batch 50000 1000 | shuffled read_formatted numpy 5000 | shuffled read_formatted_batch numpy 5000 10 | shuffled read_formatted_batch numpy 5000 1000 |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.412907 / 0.215209 (0.197698) | 4.134934 / 2.077655 (2.057280) | 1.881180 / 1.504120 (0.377060) | 1.693207 / 1.541195 (0.152012) | 1.753725 / 1.468490 (0.285235) | 0.693077 / 4.584777 (-3.891700) | 3.367409 / 3.745712 (-0.378303) | 2.749035 / 5.269862 (-2.520827) | 1.565015 / 4.565676 (-3.000662) | 0.082609 / 0.424275 (-0.341666) | 0.012500 / 0.007607 (0.004892) | 0.523619 / 0.226044 (0.297575) | 5.250188 / 2.268929 (2.981259) | 2.314255 / 55.444624 (-53.130369) | 1.962357 / 6.876477 (-4.914120) | 2.020632 / 2.142072 (-0.121441) | 0.812504 / 4.805227 (-3.992724) | 0.149921 / 6.500664 (-6.350743) | 0.065816 / 0.075469 (-0.009653) |\n\n### Benchmark: benchmark_map_filter.json\n\n| metric | filter | map fast-tokenizer batched | map identity | map identity batched | map no-op batched | map no-op batched numpy | map no-op batched pandas | map no-op batched pytorch | map no-op batched tensorflow |\n|--------|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 1.230811 / 1.841788 (-0.610977) | 14.008566 / 8.074308 (5.934258) | 14.371285 / 10.191392 (4.179893) | 0.166323 / 0.680424 (-0.514101) | 0.029702 / 0.534201 (-0.504499) | 0.408629 / 0.579283 (-0.170654) | 0.410529 / 0.434364 (-0.023835) | 0.484482 / 0.540337 (-0.055855) | 0.572360 / 1.386936 (-0.814576) |\n\n</details>\nPyArrow==latest\n\n<details>\n<summary>Show updated benchmarks!</summary>\n\n### Benchmark: benchmark_array_xd.json\n\n| metric | read_batch_formatted_as_numpy after write_array2d | read_batch_formatted_as_numpy after write_flattened_sequence | read_batch_formatted_as_numpy after write_nested_sequence | read_batch_unformated after write_array2d | read_batch_unformated after write_flattened_sequence | read_batch_unformated after write_nested_sequence | read_col_formatted_as_numpy after write_array2d | read_col_formatted_as_numpy after write_flattened_sequence | read_col_formatted_as_numpy after write_nested_sequence | read_col_unformated after write_array2d | read_col_unformated after write_flattened_sequence | read_col_unformated after write_nested_sequence | read_formatted_as_numpy after write_array2d | read_formatted_as_numpy after write_flattened_sequence | read_formatted_as_numpy after write_nested_sequence | read_unformated after write_array2d | read_unformated after write_flattened_sequence | read_unformated after write_nested_sequence | write_array2d | write_flattened_sequence | write_nested_sequence |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.006873 / 0.011353 (-0.004480) | 0.004609 / 0.011008 (-0.006400) | 0.075492 / 0.038508 (0.036984) | 0.028560 / 0.023109 (0.005450) | 0.340321 / 0.275898 (0.064423) | 0.376758 / 0.323480 (0.053278) | 0.005271 / 0.007986 (-0.002715) | 0.004786 / 0.004328 (0.000457) | 0.074843 / 0.004250 (0.070592) | 0.041072 / 0.037052 (0.004019) | 0.339952 / 0.258489 (0.081463) | 0.384375 / 0.293841 (0.090534) | 0.031771 / 0.128546 (-0.096775) | 0.011607 / 0.075646 (-0.064039) | 0.084338 / 0.419271 (-0.334933) | 0.042251 / 0.043533 (-0.001282) | 0.338904 / 0.255139 (0.083765) | 0.365360 / 0.283200 (0.082160) | 0.093151 / 0.141683 (-0.048532) | 1.449833 / 1.452155 (-0.002322) | 1.601946 / 1.492716 (0.109229) |\n\n### Benchmark: benchmark_getitem\\_100B.json\n\n| metric | get_batch_of\\_1024\\_random_rows | get_batch_of\\_1024\\_rows | get_first_row | get_last_row |\n|--------|---|---|---|---|\n| new / old (diff) | 0.225149 / 0.018006 (0.207142) | 0.409855 / 0.000490 (0.409365) | 0.000384 / 0.000200 (0.000184) | 0.000060 / 0.000054 (0.000006) |\n\n### Benchmark: benchmark_indices_mapping.json\n\n| metric | select | shard | shuffle | sort | train_test_split |\n|--------|---|---|---|---|---|\n| new / old (diff) | 0.025914 / 0.037411 (-0.011497) | 0.100443 / 0.014526 (0.085917) | 0.108557 / 0.176557 (-0.067999) | 0.150338 / 0.737135 (-0.586798) | 0.111472 / 0.296338 (-0.184866) |\n\n### Benchmark: benchmark_iterating.json\n\n| metric | read 5000 | read 50000 | read_batch 50000 10 | read_batch 50000 100 | read_batch 50000 1000 | read_formatted numpy 5000 | read_formatted pandas 5000 | read_formatted tensorflow 5000 | read_formatted torch 5000 | read_formatted_batch numpy 5000 10 | read_formatted_batch numpy 5000 1000 | shuffled read 5000 | shuffled read 50000 | shuffled read_batch 50000 10 | shuffled read_batch 50000 100 | shuffled read_batch 50000 1000 | shuffled read_formatted numpy 5000 | shuffled read_formatted_batch numpy 5000 10 | shuffled read_formatted_batch numpy 5000 1000 |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.440221 / 0.215209 (0.225012) | 4.409268 / 2.077655 (2.331613) | 2.096008 / 1.504120 (0.591888) | 1.849443 / 1.541195 (0.308248) | 1.934901 / 1.468490 (0.466410) | 0.704072 / 4.584777 (-3.880705) | 3.371370 / 3.745712 (-0.374343) | 3.185478 / 5.269862 (-2.084384) | 1.514541 / 4.565676 (-3.051135) | 0.083724 / 0.424275 (-0.340551) | 0.012674 / 0.007607 (0.005067) | 0.542155 / 0.226044 (0.316111) | 5.413456 / 2.268929 (3.144528) | 2.508567 / 55.444624 (-52.936057) | 2.163235 / 6.876477 (-4.713242) | 2.193914 / 2.142072 (0.051842) | 0.810955 / 4.805227 (-3.994272) | 0.152769 / 6.500664 (-6.347895) | 0.068009 / 0.075469 (-0.007460) |\n\n### Benchmark: benchmark_map_filter.json\n\n| metric | filter | map fast-tokenizer batched | map identity | map identity batched | map no-op batched | map no-op batched numpy | map no-op batched pandas | map no-op batched pytorch | map no-op batched tensorflow |\n|--------|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 1.272511 / 1.841788 (-0.569276) | 14.334861 / 8.074308 (6.260553) | 13.555445 / 10.191392 (3.364053) | 0.160520 / 0.680424 (-0.519904) | 0.018363 / 0.534201 (-0.515838) | 0.384937 / 0.579283 (-0.194346) | 0.409138 / 0.434364 (-0.025225) | 0.484037 / 0.540337 (-0.056300) | 0.565595 / 1.386936 (-0.821341) |\n\n</details>\n</details>\n\n![](https://cml.dev/watermark.png#23f076ef0187a4009d3c62b14a02e146baf0e35f \"CML watermark\")\n", "<details>\n<summary>Show benchmarks</summary>\n\nPyArrow==6.0.0\n\n<details>\n<summary>Show updated benchmarks!</summary>\n\n### Benchmark: benchmark_array_xd.json\n\n| metric | read_batch_formatted_as_numpy after write_array2d | read_batch_formatted_as_numpy after write_flattened_sequence | read_batch_formatted_as_numpy after write_nested_sequence | read_batch_unformated after write_array2d | read_batch_unformated after write_flattened_sequence | read_batch_unformated after write_nested_sequence | read_col_formatted_as_numpy after write_array2d | read_col_formatted_as_numpy after write_flattened_sequence | read_col_formatted_as_numpy after write_nested_sequence | read_col_unformated after write_array2d | read_col_unformated after write_flattened_sequence | read_col_unformated after write_nested_sequence | read_formatted_as_numpy after write_array2d | read_formatted_as_numpy after write_flattened_sequence | read_formatted_as_numpy after write_nested_sequence | read_unformated after write_array2d | read_unformated after write_flattened_sequence | read_unformated after write_nested_sequence | write_array2d | write_flattened_sequence | write_nested_sequence |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.010077 / 0.011353 (-0.001276) | 0.005650 / 0.011008 (-0.005359) | 0.101285 / 0.038508 (0.062777) | 0.039571 / 0.023109 (0.016462) | 0.291855 / 0.275898 (0.015957) | 0.363582 / 0.323480 (0.040102) | 0.008513 / 0.007986 (0.000527) | 0.004472 / 0.004328 (0.000144) | 0.077314 / 0.004250 (0.073064) | 0.050707 / 0.037052 (0.013654) | 0.317282 / 0.258489 (0.058792) | 0.342348 / 0.293841 (0.048507) | 0.042951 / 0.128546 (-0.085595) | 0.012295 / 0.075646 (-0.063351) | 0.337269 / 0.419271 (-0.082003) | 0.048953 / 0.043533 (0.005420) | 0.292547 / 0.255139 (0.037408) | 0.325436 / 0.283200 (0.042236) | 0.111859 / 0.141683 (-0.029824) | 1.501958 / 1.452155 (0.049804) | 1.522281 / 1.492716 (0.029565) |\n\n### Benchmark: benchmark_getitem\\_100B.json\n\n| metric | get_batch_of\\_1024\\_random_rows | get_batch_of\\_1024\\_rows | get_first_row | get_last_row |\n|--------|---|---|---|---|\n| new / old (diff) | 0.011775 / 0.018006 (-0.006231) | 0.513283 / 0.000490 (0.512793) | 0.002941 / 0.000200 (0.002741) | 0.000099 / 0.000054 (0.000044) |\n\n### Benchmark: benchmark_indices_mapping.json\n\n| metric | select | shard | shuffle | sort | train_test_split |\n|--------|---|---|---|---|---|\n| new / old (diff) | 0.028702 / 0.037411 (-0.008710) | 0.108465 / 0.014526 (0.093940) | 0.121806 / 0.176557 (-0.054750) | 0.158424 / 0.737135 (-0.578712) | 0.128077 / 0.296338 (-0.168262) |\n\n### Benchmark: benchmark_iterating.json\n\n| metric | read 5000 | read 50000 | read_batch 50000 10 | read_batch 50000 100 | read_batch 50000 1000 | read_formatted numpy 5000 | read_formatted pandas 5000 | read_formatted tensorflow 5000 | read_formatted torch 5000 | read_formatted_batch numpy 5000 10 | read_formatted_batch numpy 5000 1000 | shuffled read 5000 | shuffled read 50000 | shuffled read_batch 50000 10 | shuffled read_batch 50000 100 | shuffled read_batch 50000 1000 | shuffled read_formatted numpy 5000 | shuffled read_formatted_batch numpy 5000 10 | shuffled read_formatted_batch numpy 5000 1000 |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.395392 / 0.215209 (0.180183) | 3.944138 / 2.077655 (1.866483) | 1.773698 / 1.504120 (0.269578) | 1.588907 / 1.541195 (0.047712) | 1.697794 / 1.468490 (0.229304) | 0.690281 / 4.584777 (-3.894496) | 3.819661 / 3.745712 (0.073948) | 3.228006 / 5.269862 (-2.041856) | 1.755625 / 4.565676 (-2.810052) | 0.083169 / 0.424275 (-0.341106) | 0.012337 / 0.007607 (0.004730) | 0.504730 / 0.226044 (0.278686) | 5.016916 / 2.268929 (2.747988) | 2.245484 / 55.444624 (-53.199141) | 1.911682 / 6.876477 (-4.964795) | 1.957659 / 2.142072 (-0.184413) | 0.818361 / 4.805227 (-3.986866) | 0.162386 / 6.500664 (-6.338279) | 0.062461 / 0.075469 (-0.013008) |\n\n### Benchmark: benchmark_map_filter.json\n\n| metric | filter | map fast-tokenizer batched | map identity | map identity batched | map no-op batched | map no-op batched numpy | map no-op batched pandas | map no-op batched pytorch | map no-op batched tensorflow |\n|--------|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 1.197654 / 1.841788 (-0.644134) | 15.465611 / 8.074308 (7.391303) | 14.409126 / 10.191392 (4.217734) | 0.171776 / 0.680424 (-0.508647) | 0.028749 / 0.534201 (-0.505452) | 0.439666 / 0.579283 (-0.139618) | 0.445159 / 0.434364 (0.010795) | 0.543992 / 0.540337 (0.003655) | 0.643911 / 1.386936 (-0.743025) |\n\n</details>\nPyArrow==latest\n\n<details>\n<summary>Show updated benchmarks!</summary>\n\n### Benchmark: benchmark_array_xd.json\n\n| metric | read_batch_formatted_as_numpy after write_array2d | read_batch_formatted_as_numpy after write_flattened_sequence | read_batch_formatted_as_numpy after write_nested_sequence | read_batch_unformated after write_array2d | read_batch_unformated after write_flattened_sequence | read_batch_unformated after write_nested_sequence | read_col_formatted_as_numpy after write_array2d | read_col_formatted_as_numpy after write_flattened_sequence | read_col_formatted_as_numpy after write_nested_sequence | read_col_unformated after write_array2d | read_col_unformated after write_flattened_sequence | read_col_unformated after write_nested_sequence | read_formatted_as_numpy after write_array2d | read_formatted_as_numpy after write_flattened_sequence | read_formatted_as_numpy after write_nested_sequence | read_unformated after write_array2d | read_unformated after write_flattened_sequence | read_unformated after write_nested_sequence | write_array2d | write_flattened_sequence | write_nested_sequence |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.007036 / 0.011353 (-0.004317) | 0.005273 / 0.011008 (-0.005735) | 0.075314 / 0.038508 (0.036806) | 0.033075 / 0.023109 (0.009966) | 0.350133 / 0.275898 (0.074235) | 0.399366 / 0.323480 (0.075886) | 0.005945 / 0.007986 (-0.002041) | 0.004276 / 0.004328 (-0.000052) | 0.074975 / 0.004250 (0.070725) | 0.051758 / 0.037052 (0.014706) | 0.355077 / 0.258489 (0.096588) | 0.430296 / 0.293841 (0.136455) | 0.036257 / 0.128546 (-0.092290) | 0.012376 / 0.075646 (-0.063270) | 0.087441 / 0.419271 (-0.331830) | 0.049066 / 0.043533 (0.005534) | 0.339867 / 0.255139 (0.084728) | 0.384379 / 0.283200 (0.101179) | 0.104843 / 0.141683 (-0.036840) | 1.498897 / 1.452155 (0.046742) | 1.551400 / 1.492716 (0.058684) |\n\n### Benchmark: benchmark_getitem\\_100B.json\n\n| metric | get_batch_of\\_1024\\_random_rows | get_batch_of\\_1024\\_rows | get_first_row | get_last_row |\n|--------|---|---|---|---|\n| new / old (diff) | 0.334504 / 0.018006 (0.316498) | 0.516551 / 0.000490 (0.516061) | 0.000450 / 0.000200 (0.000250) | 0.000057 / 0.000054 (0.000003) |\n\n### Benchmark: benchmark_indices_mapping.json\n\n| metric | select | shard | shuffle | sort | train_test_split |\n|--------|---|---|---|---|---|\n| new / old (diff) | 0.029313 / 0.037411 (-0.008099) | 0.110667 / 0.014526 (0.096141) | 0.124001 / 0.176557 (-0.052556) | 0.159154 / 0.737135 (-0.577981) | 0.129503 / 0.296338 (-0.166836) |\n\n### Benchmark: benchmark_iterating.json\n\n| metric | read 5000 | read 50000 | read_batch 50000 10 | read_batch 50000 100 | read_batch 50000 1000 | read_formatted numpy 5000 | read_formatted pandas 5000 | read_formatted tensorflow 5000 | read_formatted torch 5000 | read_formatted_batch numpy 5000 10 | read_formatted_batch numpy 5000 1000 | shuffled read 5000 | shuffled read 50000 | shuffled read_batch 50000 10 | shuffled read_batch 50000 100 | shuffled read_batch 50000 1000 | shuffled read_formatted numpy 5000 | shuffled read_formatted_batch numpy 5000 10 | shuffled read_formatted_batch numpy 5000 1000 |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.416749 / 0.215209 (0.201540) | 4.171163 / 2.077655 (2.093508) | 1.981071 / 1.504120 (0.476951) | 1.788303 / 1.541195 (0.247108) | 1.912118 / 1.468490 (0.443628) | 0.708764 / 4.584777 (-3.876013) | 3.815222 / 3.745712 (0.069510) | 2.121633 / 5.269862 (-3.148229) | 1.347866 / 4.565676 (-3.217811) | 0.086340 / 0.424275 (-0.337935) | 0.012646 / 0.007607 (0.005039) | 0.525286 / 0.226044 (0.299241) | 5.254922 / 2.268929 (2.985994) | 2.488743 / 55.444624 (-52.955881) | 2.128069 / 6.876477 (-4.748408) | 2.180358 / 2.142072 (0.038286) | 0.841011 / 4.805227 (-3.964216) | 0.168732 / 6.500664 (-6.331932) | 0.065559 / 0.075469 (-0.009910) |\n\n### Benchmark: benchmark_map_filter.json\n\n| metric | filter | map fast-tokenizer batched | map identity | map identity batched | map no-op batched | map no-op batched numpy | map no-op batched pandas | map no-op batched pytorch | map no-op batched tensorflow |\n|--------|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 1.270518 / 1.841788 (-0.571270) | 15.557563 / 8.074308 (7.483255) | 13.660757 / 10.191392 (3.469365) | 0.185636 / 0.680424 (-0.494788) | 0.018152 / 0.534201 (-0.516049) | 0.423553 / 0.579283 (-0.155730) | 0.412718 / 0.434364 (-0.021646) | 0.528455 / 0.540337 (-0.011882) | 0.635274 / 1.386936 (-0.751662) |\n\n</details>\n</details>\n\n![](https://cml.dev/watermark.png#d40f05ef827c52344a2c6e83f7c8d13bb6b660d3 \"CML watermark\")\n", "<details>\n<summary>Show benchmarks</summary>\n\nPyArrow==6.0.0\n\n<details>\n<summary>Show updated benchmarks!</summary>\n\n### Benchmark: benchmark_array_xd.json\n\n| metric | read_batch_formatted_as_numpy after write_array2d | read_batch_formatted_as_numpy after write_flattened_sequence | read_batch_formatted_as_numpy after write_nested_sequence | read_batch_unformated after write_array2d | read_batch_unformated after write_flattened_sequence | read_batch_unformated after write_nested_sequence | read_col_formatted_as_numpy after write_array2d | read_col_formatted_as_numpy after write_flattened_sequence | read_col_formatted_as_numpy after write_nested_sequence | read_col_unformated after write_array2d | read_col_unformated after write_flattened_sequence | read_col_unformated after write_nested_sequence | read_formatted_as_numpy after write_array2d | read_formatted_as_numpy after write_flattened_sequence | read_formatted_as_numpy after write_nested_sequence | read_unformated after write_array2d | read_unformated after write_flattened_sequence | read_unformated after write_nested_sequence | write_array2d | write_flattened_sequence | write_nested_sequence |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.011194 / 0.011353 (-0.000159) | 0.006344 / 0.011008 (-0.004664) | 0.122013 / 0.038508 (0.083505) | 0.044323 / 0.023109 (0.021214) | 0.356665 / 0.275898 (0.080767) | 0.439871 / 0.323480 (0.116391) | 0.010694 / 0.007986 (0.002709) | 0.004648 / 0.004328 (0.000320) | 0.091140 / 0.004250 (0.086890) | 0.052457 / 0.037052 (0.015404) | 0.369282 / 0.258489 (0.110793) | 0.403279 / 0.293841 (0.109438) | 0.054075 / 0.128546 (-0.074472) | 0.014484 / 0.075646 (-0.061162) | 0.407932 / 0.419271 (-0.011340) | 0.060681 / 0.043533 (0.017148) | 0.350889 / 0.255139 (0.095750) | 0.392041 / 0.283200 (0.108841) | 0.121252 / 0.141683 (-0.020431) | 1.809527 / 1.452155 (0.357373) | 1.835141 / 1.492716 (0.342425) |\n\n### Benchmark: benchmark_getitem\\_100B.json\n\n| metric | get_batch_of\\_1024\\_random_rows | get_batch_of\\_1024\\_rows | get_first_row | get_last_row |\n|--------|---|---|---|---|\n| new / old (diff) | 0.227372 / 0.018006 (0.209366) | 0.481908 / 0.000490 (0.481418) | 0.007262 / 0.000200 (0.007062) | 0.000148 / 0.000054 (0.000093) |\n\n### Benchmark: benchmark_indices_mapping.json\n\n| metric | select | shard | shuffle | sort | train_test_split |\n|--------|---|---|---|---|---|\n| new / old (diff) | 0.031039 / 0.037411 (-0.006372) | 0.133947 / 0.014526 (0.119421) | 0.141935 / 0.176557 (-0.034622) | 0.197854 / 0.737135 (-0.539281) | 0.152393 / 0.296338 (-0.143945) |\n\n### Benchmark: benchmark_iterating.json\n\n| metric | read 5000 | read 50000 | read_batch 50000 10 | read_batch 50000 100 | read_batch 50000 1000 | read_formatted numpy 5000 | read_formatted pandas 5000 | read_formatted tensorflow 5000 | read_formatted torch 5000 | read_formatted_batch numpy 5000 10 | read_formatted_batch numpy 5000 1000 | shuffled read 5000 | shuffled read 50000 | shuffled read_batch 50000 10 | shuffled read_batch 50000 100 | shuffled read_batch 50000 1000 | shuffled read_formatted numpy 5000 | shuffled read_formatted_batch numpy 5000 10 | shuffled read_formatted_batch numpy 5000 1000 |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.517400 / 0.215209 (0.302191) | 4.899972 / 2.077655 (2.822317) | 2.171023 / 1.504120 (0.666903) | 2.008706 / 1.541195 (0.467511) | 1.988777 / 1.468490 (0.520287) | 0.859872 / 4.584777 (-3.724905) | 4.673923 / 3.745712 (0.928211) | 2.703189 / 5.269862 (-2.566672) | 1.891680 / 4.565676 (-2.673997) | 0.109601 / 0.424275 (-0.314674) | 0.014622 / 0.007607 (0.007015) | 0.618990 / 0.226044 (0.392946) | 6.255608 / 2.268929 (3.986679) | 2.822199 / 55.444624 (-52.622425) | 2.457684 / 6.876477 (-4.418793) | 2.500041 / 2.142072 (0.357968) | 1.054529 / 4.805227 (-3.750698) | 0.209501 / 6.500664 (-6.291163) | 0.074929 / 0.075469 (-0.000540) |\n\n### Benchmark: benchmark_map_filter.json\n\n| metric | filter | map fast-tokenizer batched | map identity | map identity batched | map no-op batched | map no-op batched numpy | map no-op batched pandas | map no-op batched pytorch | map no-op batched tensorflow |\n|--------|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 1.532780 / 1.841788 (-0.309008) | 19.159455 / 8.074308 (11.085147) | 17.817063 / 10.191392 (7.625671) | 0.194078 / 0.680424 (-0.486346) | 0.038211 / 0.534201 (-0.495990) | 0.537366 / 0.579283 (-0.041917) | 0.538995 / 0.434364 (0.104631) | 0.679431 / 0.540337 (0.139094) | 0.801960 / 1.386936 (-0.584976) |\n\n</details>\nPyArrow==latest\n\n<details>\n<summary>Show updated benchmarks!</summary>\n\n### Benchmark: benchmark_array_xd.json\n\n| metric | read_batch_formatted_as_numpy after write_array2d | read_batch_formatted_as_numpy after write_flattened_sequence | read_batch_formatted_as_numpy after write_nested_sequence | read_batch_unformated after write_array2d | read_batch_unformated after write_flattened_sequence | read_batch_unformated after write_nested_sequence | read_col_formatted_as_numpy after write_array2d | read_col_formatted_as_numpy after write_flattened_sequence | read_col_formatted_as_numpy after write_nested_sequence | read_col_unformated after write_array2d | read_col_unformated after write_flattened_sequence | read_col_unformated after write_nested_sequence | read_formatted_as_numpy after write_array2d | read_formatted_as_numpy after write_flattened_sequence | read_formatted_as_numpy after write_nested_sequence | read_unformated after write_array2d | read_unformated after write_flattened_sequence | read_unformated after write_nested_sequence | write_array2d | write_flattened_sequence | write_nested_sequence |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.008729 / 0.011353 (-0.002624) | 0.005711 / 0.011008 (-0.005297) | 0.091570 / 0.038508 (0.053062) | 0.039805 / 0.023109 (0.016696) | 0.413507 / 0.275898 (0.137609) | 0.456342 / 0.323480 (0.132862) | 0.006201 / 0.007986 (-0.001785) | 0.009700 / 0.004328 (0.005372) | 0.089146 / 0.004250 (0.084896) | 0.057543 / 0.037052 (0.020490) | 0.420806 / 0.258489 (0.162317) | 0.471962 / 0.293841 (0.178121) | 0.043940 / 0.128546 (-0.084606) | 0.014457 / 0.075646 (-0.061190) | 0.106674 / 0.419271 (-0.312598) | 0.058930 / 0.043533 (0.015397) | 0.419111 / 0.255139 (0.163972) | 0.452974 / 0.283200 (0.169774) | 0.124573 / 0.141683 (-0.017110) | 1.864753 / 1.452155 (0.412599) | 1.935387 / 1.492716 (0.442670) |\n\n### Benchmark: benchmark_getitem\\_100B.json\n\n| metric | get_batch_of\\_1024\\_random_rows | get_batch_of\\_1024\\_rows | get_first_row | get_last_row |\n|--------|---|---|---|---|\n| new / old (diff) | 0.275657 / 0.018006 (0.257651) | 0.498096 / 0.000490 (0.497606) | 0.000480 / 0.000200 (0.000280) | 0.000066 / 0.000054 (0.000012) |\n\n### Benchmark: benchmark_indices_mapping.json\n\n| metric | select | shard | shuffle | sort | train_test_split |\n|--------|---|---|---|---|---|\n| new / old (diff) | 0.034377 / 0.037411 (-0.003035) | 0.138050 / 0.014526 (0.123524) | 0.153718 / 0.176557 (-0.022838) | 0.201445 / 0.737135 (-0.535690) | 0.160346 / 0.296338 (-0.135992) |\n\n### Benchmark: benchmark_iterating.json\n\n| metric | read 5000 | read 50000 | read_batch 50000 10 | read_batch 50000 100 | read_batch 50000 1000 | read_formatted numpy 5000 | read_formatted pandas 5000 | read_formatted tensorflow 5000 | read_formatted torch 5000 | read_formatted_batch numpy 5000 10 | read_formatted_batch numpy 5000 1000 | shuffled read 5000 | shuffled read 50000 | shuffled read_batch 50000 10 | shuffled read_batch 50000 100 | shuffled read_batch 50000 1000 | shuffled read_formatted numpy 5000 | shuffled read_formatted_batch numpy 5000 10 | shuffled read_formatted_batch numpy 5000 1000 |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.540670 / 0.215209 (0.325461) | 5.376291 / 2.077655 (3.298636) | 2.581799 / 1.504120 (1.077679) | 2.328858 / 1.541195 (0.787663) | 2.446458 / 1.468490 (0.977968) | 0.923005 / 4.584777 (-3.661772) | 4.815977 / 3.745712 (1.070265) | 4.205725 / 5.269862 (-1.064137) | 2.400466 / 4.565676 (-2.165211) | 0.107207 / 0.424275 (-0.317068) | 0.015427 / 0.007607 (0.007819) | 0.657267 / 0.226044 (0.431222) | 6.491256 / 2.268929 (4.222327) | 3.179099 / 55.444624 (-52.265525) | 2.722434 / 6.876477 (-4.154042) | 2.788202 / 2.142072 (0.646129) | 1.060016 / 4.805227 (-3.745211) | 0.206899 / 6.500664 (-6.293766) | 0.077868 / 0.075469 (0.002399) |\n\n### Benchmark: benchmark_map_filter.json\n\n| metric | filter | map fast-tokenizer batched | map identity | map identity batched | map no-op batched | map no-op batched numpy | map no-op batched pandas | map no-op batched pytorch | map no-op batched tensorflow |\n|--------|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 1.567894 / 1.841788 (-0.273893) | 19.314330 / 8.074308 (11.240022) | 17.597614 / 10.191392 (7.406222) | 0.195777 / 0.680424 (-0.484647) | 0.022160 / 0.534201 (-0.512041) | 0.530592 / 0.579283 (-0.048691) | 0.508591 / 0.434364 (0.074227) | 0.619794 / 0.540337 (0.079457) | 0.749773 / 1.386936 (-0.637163) |\n\n</details>\n</details>\n\n![](https://cml.dev/watermark.png#8637141a67639c510294620306c9bb25d31d34ef \"CML watermark\")\n", "<details>\n<summary>Show benchmarks</summary>\n\nPyArrow==6.0.0\n\n<details>\n<summary>Show updated benchmarks!</summary>\n\n### Benchmark: benchmark_array_xd.json\n\n| metric | read_batch_formatted_as_numpy after write_array2d | read_batch_formatted_as_numpy after write_flattened_sequence | read_batch_formatted_as_numpy after write_nested_sequence | read_batch_unformated after write_array2d | read_batch_unformated after write_flattened_sequence | read_batch_unformated after write_nested_sequence | read_col_formatted_as_numpy after write_array2d | read_col_formatted_as_numpy after write_flattened_sequence | read_col_formatted_as_numpy after write_nested_sequence | read_col_unformated after write_array2d | read_col_unformated after write_flattened_sequence | read_col_unformated after write_nested_sequence | read_formatted_as_numpy after write_array2d | read_formatted_as_numpy after write_flattened_sequence | read_formatted_as_numpy after write_nested_sequence | read_unformated after write_array2d | read_unformated after write_flattened_sequence | read_unformated after write_nested_sequence | write_array2d | write_flattened_sequence | write_nested_sequence |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.012431 / 0.011353 (0.001078) | 0.006526 / 0.011008 (-0.004482) | 0.132266 / 0.038508 (0.093757) | 0.043199 / 0.023109 (0.020089) | 0.405230 / 0.275898 (0.129332) | 0.494643 / 0.323480 (0.171163) | 0.009927 / 0.007986 (0.001941) | 0.005227 / 0.004328 (0.000899) | 0.110914 / 0.004250 (0.106664) | 0.047815 / 0.037052 (0.010763) | 0.419099 / 0.258489 (0.160610) | 0.463405 / 0.293841 (0.169564) | 0.057858 / 0.128546 (-0.070688) | 0.018918 / 0.075646 (-0.056728) | 0.450584 / 0.419271 (0.031313) | 0.060457 / 0.043533 (0.016924) | 0.408234 / 0.255139 (0.153095) | 0.433722 / 0.283200 (0.150523) | 0.119403 / 0.141683 (-0.022280) | 1.966742 / 1.452155 (0.514587) | 1.980685 / 1.492716 (0.487969) |\n\n### Benchmark: benchmark_getitem\\_100B.json\n\n| metric | get_batch_of\\_1024\\_random_rows | get_batch_of\\_1024\\_rows | get_first_row | get_last_row |\n|--------|---|---|---|---|\n| new / old (diff) | 0.292853 / 0.018006 (0.274847) | 0.619697 / 0.000490 (0.619207) | 0.002135 / 0.000200 (0.001935) | 0.000117 / 0.000054 (0.000062) |\n\n### Benchmark: benchmark_indices_mapping.json\n\n| metric | select | shard | shuffle | sort | train_test_split |\n|--------|---|---|---|---|---|\n| new / old (diff) | 0.031283 / 0.037411 (-0.006129) | 0.128649 / 0.014526 (0.114123) | 0.150116 / 0.176557 (-0.026441) | 0.187605 / 0.737135 (-0.549530) | 0.153334 / 0.296338 (-0.143005) |\n\n### Benchmark: benchmark_iterating.json\n\n| metric | read 5000 | read 50000 | read_batch 50000 10 | read_batch 50000 100 | read_batch 50000 1000 | read_formatted numpy 5000 | read_formatted pandas 5000 | read_formatted tensorflow 5000 | read_formatted torch 5000 | read_formatted_batch numpy 5000 10 | read_formatted_batch numpy 5000 1000 | shuffled read 5000 | shuffled read 50000 | shuffled read_batch 50000 10 | shuffled read_batch 50000 100 | shuffled read_batch 50000 1000 | shuffled read_formatted numpy 5000 | shuffled read_formatted_batch numpy 5000 10 | shuffled read_formatted_batch numpy 5000 1000 |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.659660 / 0.215209 (0.444451) | 6.459749 / 2.077655 (4.382094) | 2.764566 / 1.504120 (1.260446) | 2.362630 / 1.541195 (0.821435) | 2.426421 / 1.468490 (0.957931) | 1.282407 / 4.584777 (-3.302370) | 5.668865 / 3.745712 (1.923153) | 3.236255 / 5.269862 (-2.033606) | 2.248836 / 4.565676 (-2.316841) | 0.145861 / 0.424275 (-0.278414) | 0.015707 / 0.007607 (0.008100) | 0.805218 / 0.226044 (0.579174) | 8.146831 / 2.268929 (5.877903) | 3.506283 / 55.444624 (-51.938341) | 2.736682 / 6.876477 (-4.139795) | 2.959039 / 2.142072 (0.816967) | 1.528428 / 4.805227 (-3.276799) | 0.270980 / 6.500664 (-6.229684) | 0.086824 / 0.075469 (0.011355) |\n\n### Benchmark: benchmark_map_filter.json\n\n| metric | filter | map fast-tokenizer batched | map identity | map identity batched | map no-op batched | map no-op batched numpy | map no-op batched pandas | map no-op batched pytorch | map no-op batched tensorflow |\n|--------|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 1.682506 / 1.841788 (-0.159282) | 18.844103 / 8.074308 (10.769795) | 21.008471 / 10.191392 (10.817079) | 0.258372 / 0.680424 (-0.422052) | 0.046505 / 0.534201 (-0.487696) | 0.574760 / 0.579283 (-0.004523) | 0.663745 / 0.434364 (0.229381) | 0.702411 / 0.540337 (0.162074) | 0.824024 / 1.386936 (-0.562912) |\n\n</details>\nPyArrow==latest\n\n<details>\n<summary>Show updated benchmarks!</summary>\n\n### Benchmark: benchmark_array_xd.json\n\n| metric | read_batch_formatted_as_numpy after write_array2d | read_batch_formatted_as_numpy after write_flattened_sequence | read_batch_formatted_as_numpy after write_nested_sequence | read_batch_unformated after write_array2d | read_batch_unformated after write_flattened_sequence | read_batch_unformated after write_nested_sequence | read_col_formatted_as_numpy after write_array2d | read_col_formatted_as_numpy after write_flattened_sequence | read_col_formatted_as_numpy after write_nested_sequence | read_col_unformated after write_array2d | read_col_unformated after write_flattened_sequence | read_col_unformated after write_nested_sequence | read_formatted_as_numpy after write_array2d | read_formatted_as_numpy after write_flattened_sequence | read_formatted_as_numpy after write_nested_sequence | read_unformated after write_array2d | read_unformated after write_flattened_sequence | read_unformated after write_nested_sequence | write_array2d | write_flattened_sequence | write_nested_sequence |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.010016 / 0.011353 (-0.001337) | 0.007459 / 0.011008 (-0.003549) | 0.103954 / 0.038508 (0.065446) | 0.036363 / 0.023109 (0.013254) | 0.464079 / 0.275898 (0.188181) | 0.504730 / 0.323480 (0.181250) | 0.007865 / 0.007986 (-0.000121) | 0.005210 / 0.004328 (0.000882) | 0.105018 / 0.004250 (0.100767) | 0.062191 / 0.037052 (0.025139) | 0.483304 / 0.258489 (0.224815) | 0.547030 / 0.293841 (0.253189) | 0.055436 / 0.128546 (-0.073110) | 0.021073 / 0.075646 (-0.054573) | 0.120952 / 0.419271 (-0.298319) | 0.075593 / 0.043533 (0.032060) | 0.459930 / 0.255139 (0.204791) | 0.486924 / 0.283200 (0.203724) | 0.129465 / 0.141683 (-0.012218) | 1.902322 / 1.452155 (0.450167) | 1.980809 / 1.492716 (0.488092) |\n\n### Benchmark: benchmark_getitem\\_100B.json\n\n| metric | get_batch_of\\_1024\\_random_rows | get_batch_of\\_1024\\_rows | get_first_row | get_last_row |\n|--------|---|---|---|---|\n| new / old (diff) | 0.259263 / 0.018006 (0.241257) | 0.596703 / 0.000490 (0.596213) | 0.004520 / 0.000200 (0.004320) | 0.000124 / 0.000054 (0.000070) |\n\n### Benchmark: benchmark_indices_mapping.json\n\n| metric | select | shard | shuffle | sort | train_test_split |\n|--------|---|---|---|---|---|\n| new / old (diff) | 0.032802 / 0.037411 (-0.004609) | 0.138751 / 0.014526 (0.124225) | 0.147106 / 0.176557 (-0.029451) | 0.194791 / 0.737135 (-0.542345) | 0.152643 / 0.296338 (-0.143696) |\n\n### Benchmark: benchmark_iterating.json\n\n| metric | read 5000 | read 50000 | read_batch 50000 10 | read_batch 50000 100 | read_batch 50000 1000 | read_formatted numpy 5000 | read_formatted pandas 5000 | read_formatted tensorflow 5000 | read_formatted torch 5000 | read_formatted_batch numpy 5000 10 | read_formatted_batch numpy 5000 1000 | shuffled read 5000 | shuffled read 50000 | shuffled read_batch 50000 10 | shuffled read_batch 50000 100 | shuffled read_batch 50000 1000 | shuffled read_formatted numpy 5000 | shuffled read_formatted_batch numpy 5000 10 | shuffled read_formatted_batch numpy 5000 1000 |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.678455 / 0.215209 (0.463246) | 6.673643 / 2.077655 (4.595989) | 2.943368 / 1.504120 (1.439248) | 2.591223 / 1.541195 (1.050029) | 2.741097 / 1.468490 (1.272607) | 1.261178 / 4.584777 (-3.323599) | 5.773853 / 3.745712 (2.028141) | 3.171559 / 5.269862 (-2.098303) | 2.124898 / 4.565676 (-2.440779) | 0.161849 / 0.424275 (-0.262426) | 0.015498 / 0.007607 (0.007891) | 0.857984 / 0.226044 (0.631940) | 8.456946 / 2.268929 (6.188018) | 3.818787 / 55.444624 (-51.625837) | 3.009953 / 6.876477 (-3.866523) | 3.113006 / 2.142072 (0.970934) | 1.477299 / 4.805227 (-3.327929) | 0.267207 / 6.500664 (-6.233457) | 0.087590 / 0.075469 (0.012121) |\n\n### Benchmark: benchmark_map_filter.json\n\n| metric | filter | map fast-tokenizer batched | map identity | map identity batched | map no-op batched | map no-op batched numpy | map no-op batched pandas | map no-op batched pytorch | map no-op batched tensorflow |\n|--------|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 1.757389 / 1.841788 (-0.084398) | 19.287690 / 8.074308 (11.213381) | 21.601991 / 10.191392 (11.410599) | 0.260464 / 0.680424 (-0.419960) | 0.028552 / 0.534201 (-0.505649) | 0.558934 / 0.579283 (-0.020349) | 0.673651 / 0.434364 (0.239287) | 0.714448 / 0.540337 (0.174111) | 0.857608 / 1.386936 (-0.529328) |\n\n</details>\n</details>\n\n![](https://cml.dev/watermark.png#2d3bd0134de444ffd10c4a39873dbf9aa3732c08 \"CML watermark\")\n", "Ready for review @mariosasko, LMKWYT :)\r\n\r\nSorry it tooks me a few tries to fix the CI - I ended up not trying to use the latest `torch` version in the CI.", "<details>\n<summary>Show benchmarks</summary>\n\nPyArrow==6.0.0\n\n<details>\n<summary>Show updated benchmarks!</summary>\n\n### Benchmark: benchmark_array_xd.json\n\n| metric | read_batch_formatted_as_numpy after write_array2d | read_batch_formatted_as_numpy after write_flattened_sequence | read_batch_formatted_as_numpy after write_nested_sequence | read_batch_unformated after write_array2d | read_batch_unformated after write_flattened_sequence | read_batch_unformated after write_nested_sequence | read_col_formatted_as_numpy after write_array2d | read_col_formatted_as_numpy after write_flattened_sequence | read_col_formatted_as_numpy after write_nested_sequence | read_col_unformated after write_array2d | read_col_unformated after write_flattened_sequence | read_col_unformated after write_nested_sequence | read_formatted_as_numpy after write_array2d | read_formatted_as_numpy after write_flattened_sequence | read_formatted_as_numpy after write_nested_sequence | read_unformated after write_array2d | read_unformated after write_flattened_sequence | read_unformated after write_nested_sequence | write_array2d | write_flattened_sequence | write_nested_sequence |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.009474 / 0.011353 (-0.001878) | 0.005507 / 0.011008 (-0.005501) | 0.101219 / 0.038508 (0.062711) | 0.035591 / 0.023109 (0.012481) | 0.305841 / 0.275898 (0.029943) | 0.339135 / 0.323480 (0.015656) | 0.007920 / 0.007986 (-0.000066) | 0.004252 / 0.004328 (-0.000077) | 0.076912 / 0.004250 (0.072662) | 0.041923 / 0.037052 (0.004871) | 0.301405 / 0.258489 (0.042916) | 0.356488 / 0.293841 (0.062647) | 0.039342 / 0.128546 (-0.089204) | 0.012711 / 0.075646 (-0.062935) | 0.334193 / 0.419271 (-0.085079) | 0.049112 / 0.043533 (0.005579) | 0.301484 / 0.255139 (0.046345) | 0.315306 / 0.283200 (0.032106) | 0.102959 / 0.141683 (-0.038724) | 1.420677 / 1.452155 (-0.031478) | 1.549493 / 1.492716 (0.056777) |\n\n### Benchmark: benchmark_getitem\\_100B.json\n\n| metric | get_batch_of\\_1024\\_random_rows | get_batch_of\\_1024\\_rows | get_first_row | get_last_row |\n|--------|---|---|---|---|\n| new / old (diff) | 0.284639 / 0.018006 (0.266633) | 0.501226 / 0.000490 (0.500736) | 0.004328 / 0.000200 (0.004128) | 0.000091 / 0.000054 (0.000036) |\n\n### Benchmark: benchmark_indices_mapping.json\n\n| metric | select | shard | shuffle | sort | train_test_split |\n|--------|---|---|---|---|---|\n| new / old (diff) | 0.027034 / 0.037411 (-0.010377) | 0.108066 / 0.014526 (0.093540) | 0.122106 / 0.176557 (-0.054451) | 0.162908 / 0.737135 (-0.574227) | 0.127233 / 0.296338 (-0.169105) |\n\n### Benchmark: benchmark_iterating.json\n\n| metric | read 5000 | read 50000 | read_batch 50000 10 | read_batch 50000 100 | read_batch 50000 1000 | read_formatted numpy 5000 | read_formatted pandas 5000 | read_formatted tensorflow 5000 | read_formatted torch 5000 | read_formatted_batch numpy 5000 10 | read_formatted_batch numpy 5000 1000 | shuffled read 5000 | shuffled read 50000 | shuffled read_batch 50000 10 | shuffled read_batch 50000 100 | shuffled read_batch 50000 1000 | shuffled read_formatted numpy 5000 | shuffled read_formatted_batch numpy 5000 10 | shuffled read_formatted_batch numpy 5000 1000 |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.394023 / 0.215209 (0.178813) | 3.932729 / 2.077655 (1.855075) | 1.771195 / 1.504120 (0.267075) | 1.582788 / 1.541195 (0.041594) | 1.703219 / 1.468490 (0.234728) | 0.702629 / 4.584777 (-3.882148) | 3.780187 / 3.745712 (0.034475) | 2.180433 / 5.269862 (-3.089428) | 1.504806 / 4.565676 (-3.060871) | 0.085289 / 0.424275 (-0.338986) | 0.012580 / 0.007607 (0.004973) | 0.515408 / 0.226044 (0.289363) | 5.010613 / 2.268929 (2.741685) | 2.256648 / 55.444624 (-53.187976) | 1.914971 / 6.876477 (-4.961505) | 2.038436 / 2.142072 (-0.103636) | 0.846240 / 4.805227 (-3.958987) | 0.164920 / 6.500664 (-6.335744) | 0.063899 / 0.075469 (-0.011570) |\n\n### Benchmark: benchmark_map_filter.json\n\n| metric | filter | map fast-tokenizer batched | map identity | map identity batched | map no-op batched | map no-op batched numpy | map no-op batched pandas | map no-op batched pytorch | map no-op batched tensorflow |\n|--------|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 1.224160 / 1.841788 (-0.617627) | 15.089995 / 8.074308 (7.015687) | 14.777003 / 10.191392 (4.585611) | 0.169873 / 0.680424 (-0.510551) | 0.029233 / 0.534201 (-0.504968) | 0.445424 / 0.579283 (-0.133859) | 0.439194 / 0.434364 (0.004830) | 0.536370 / 0.540337 (-0.003968) | 0.636694 / 1.386936 (-0.750242) |\n\n</details>\nPyArrow==latest\n\n<details>\n<summary>Show updated benchmarks!</summary>\n\n### Benchmark: benchmark_array_xd.json\n\n| metric | read_batch_formatted_as_numpy after write_array2d | read_batch_formatted_as_numpy after write_flattened_sequence | read_batch_formatted_as_numpy after write_nested_sequence | read_batch_unformated after write_array2d | read_batch_unformated after write_flattened_sequence | read_batch_unformated after write_nested_sequence | read_col_formatted_as_numpy after write_array2d | read_col_formatted_as_numpy after write_flattened_sequence | read_col_formatted_as_numpy after write_nested_sequence | read_col_unformated after write_array2d | read_col_unformated after write_flattened_sequence | read_col_unformated after write_nested_sequence | read_formatted_as_numpy after write_array2d | read_formatted_as_numpy after write_flattened_sequence | read_formatted_as_numpy after write_nested_sequence | read_unformated after write_array2d | read_unformated after write_flattened_sequence | read_unformated after write_nested_sequence | write_array2d | write_flattened_sequence | write_nested_sequence |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.008230 / 0.011353 (-0.003122) | 0.005499 / 0.011008 (-0.005509) | 0.076108 / 0.038508 (0.037600) | 0.037444 / 0.023109 (0.014335) | 0.364420 / 0.275898 (0.088522) | 0.412308 / 0.323480 (0.088828) | 0.006704 / 0.007986 (-0.001282) | 0.004359 / 0.004328 (0.000031) | 0.075080 / 0.004250 (0.070830) | 0.057698 / 0.037052 (0.020646) | 0.366088 / 0.258489 (0.107599) | 0.409583 / 0.293841 (0.115742) | 0.037882 / 0.128546 (-0.090664) | 0.012421 / 0.075646 (-0.063225) | 0.087701 / 0.419271 (-0.331571) | 0.050669 / 0.043533 (0.007136) | 0.351139 / 0.255139 (0.096000) | 0.384340 / 0.283200 (0.101140) | 0.108097 / 0.141683 (-0.033586) | 1.445010 / 1.452155 (-0.007145) | 1.559570 / 1.492716 (0.066853) |\n\n### Benchmark: benchmark_getitem\\_100B.json\n\n| metric | get_batch_of\\_1024\\_random_rows | get_batch_of\\_1024\\_rows | get_first_row | get_last_row |\n|--------|---|---|---|---|\n| new / old (diff) | 0.324114 / 0.018006 (0.306108) | 0.549134 / 0.000490 (0.548644) | 0.003544 / 0.000200 (0.003344) | 0.000097 / 0.000054 (0.000042) |\n\n### Benchmark: benchmark_indices_mapping.json\n\n| metric | select | shard | shuffle | sort | train_test_split |\n|--------|---|---|---|---|---|\n| new / old (diff) | 0.030646 / 0.037411 (-0.006765) | 0.108573 / 0.014526 (0.094047) | 0.125291 / 0.176557 (-0.051266) | 0.174798 / 0.737135 (-0.562338) | 0.128000 / 0.296338 (-0.168338) |\n\n### Benchmark: benchmark_iterating.json\n\n| metric | read 5000 | read 50000 | read_batch 50000 10 | read_batch 50000 100 | read_batch 50000 1000 | read_formatted numpy 5000 | read_formatted pandas 5000 | read_formatted tensorflow 5000 | read_formatted torch 5000 | read_formatted_batch numpy 5000 10 | read_formatted_batch numpy 5000 1000 | shuffled read 5000 | shuffled read 50000 | shuffled read_batch 50000 10 | shuffled read_batch 50000 100 | shuffled read_batch 50000 1000 | shuffled read_formatted numpy 5000 | shuffled read_formatted_batch numpy 5000 10 | shuffled read_formatted_batch numpy 5000 1000 |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.428881 / 0.215209 (0.213672) | 4.282320 / 2.077655 (2.204665) | 2.061462 / 1.504120 (0.557342) | 1.858477 / 1.541195 (0.317283) | 1.971646 / 1.468490 (0.503156) | 0.723631 / 4.584777 (-3.861146) | 3.822376 / 3.745712 (0.076664) | 2.174427 / 5.269862 (-3.095434) | 1.386066 / 4.565676 (-3.179611) | 0.088391 / 0.424275 (-0.335884) | 0.012948 / 0.007607 (0.005341) | 0.524423 / 0.226044 (0.298378) | 5.249389 / 2.268929 (2.980460) | 2.528662 / 55.444624 (-52.915962) | 2.245329 / 6.876477 (-4.631147) | 2.402733 / 2.142072 (0.260660) | 0.868864 / 4.805227 (-3.936364) | 0.174066 / 6.500664 (-6.326598) | 0.066165 / 0.075469 (-0.009304) |\n\n### Benchmark: benchmark_map_filter.json\n\n| metric | filter | map fast-tokenizer batched | map identity | map identity batched | map no-op batched | map no-op batched numpy | map no-op batched pandas | map no-op batched pytorch | map no-op batched tensorflow |\n|--------|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 1.296922 / 1.841788 (-0.544865) | 15.814109 / 8.074308 (7.739801) | 14.086059 / 10.191392 (3.894667) | 0.190952 / 0.680424 (-0.489472) | 0.017679 / 0.534201 (-0.516522) | 0.428872 / 0.579283 (-0.150411) | 0.435399 / 0.434364 (0.001035) | 0.540856 / 0.540337 (0.000519) | 0.648904 / 1.386936 (-0.738032) |\n\n</details>\n</details>\n\n![](https://cml.dev/watermark.png#f401758c5019ede4404994d5d59220125984874d \"CML watermark\")\n" ]

[]

[]

[ "Update `datasets` or downgrade `huggingface-hub` ;)\r\n\r\nThe `huggingface-hub` lib did a breaking change a few months ago, and you're using an old version of `datasets` that does't support it", "Awesome thanks a lot! Everything works just fine with `datasets==2.9.0` :-) ", "Getting same error with latest versions.\r\n\r\n\r\n```shell\r\n---------------------------------------------------------------------------\r\nTypeError Traceback (most recent call last)\r\nCell In[99], line 1\r\n----> 1 dataset.push_to_hub(\"mirfan899/kids_phoneme_asr\")\r\n\r\nFile /opt/conda/lib/python3.10/site-packages/datasets/arrow_dataset.py:3538, in Dataset.push_to_hub(self, repo_id, split, private, token, branch, shard_size, embed_external_files)\r\n 3493 def push_to_hub(\r\n 3494 self,\r\n 3495 repo_id: str,\r\n (...)\r\n 3501 embed_external_files: bool = True,\r\n 3502 ):\r\n 3503 \"\"\"Pushes the dataset to the hub.\r\n 3504 The dataset is pushed using HTTP requests and does not need to have neither git or git-lfs installed.\r\n 3505 \r\n (...)\r\n 3536 ```\r\n 3537 \"\"\"\r\n-> 3538 repo_id, split, uploaded_size, dataset_nbytes = self._push_parquet_shards_to_hub(\r\n 3539 repo_id=repo_id,\r\n 3540 split=split,\r\n 3541 private=private,\r\n 3542 token=token,\r\n 3543 branch=branch,\r\n 3544 shard_size=shard_size,\r\n 3545 embed_external_files=embed_external_files,\r\n 3546 )\r\n 3547 organization, dataset_name = repo_id.split(\"/\")\r\n 3548 info_to_dump = self.info.copy()\r\n\r\nFile /opt/conda/lib/python3.10/site-packages/datasets/arrow_dataset.py:3474, in Dataset._push_parquet_shards_to_hub(self, repo_id, split, private, token, branch, shard_size, embed_external_files)\r\n 3472 shard.to_parquet(buffer)\r\n 3473 uploaded_size += buffer.tell()\r\n-> 3474 _retry(\r\n 3475 api.upload_file,\r\n 3476 func_kwargs=dict(\r\n 3477 path_or_fileobj=buffer.getvalue(),\r\n 3478 path_in_repo=path_in_repo(index),\r\n 3479 repo_id=repo_id,\r\n 3480 token=token,\r\n 3481 repo_type=\"dataset\",\r\n 3482 revision=branch,\r\n 3483 identical_ok=True,\r\n 3484 ),\r\n 3485 exceptions=HTTPError,\r\n 3486 status_codes=[504],\r\n 3487 base_wait_time=2.0,\r\n 3488 max_retries=5,\r\n 3489 max_wait_time=20.0,\r\n 3490 )\r\n 3491 return repo_id, split, uploaded_size, dataset_nbytes\r\n\r\nFile /opt/conda/lib/python3.10/site-packages/datasets/utils/file_utils.py:330, in _retry(func, func_args, func_kwargs, exceptions, status_codes, max_retries, base_wait_time, max_wait_time)\r\n 328 while True:\r\n 329 try:\r\n--> 330 return func(*func_args, **func_kwargs)\r\n 331 except exceptions as err:\r\n 332 if retry >= max_retries or (status_codes and err.response.status_code not in status_codes):\r\n\r\nFile /opt/conda/lib/python3.10/site-packages/huggingface_hub/utils/_validators.py:120, in validate_hf_hub_args.<locals>._inner_fn(*args, **kwargs)\r\n 117 if check_use_auth_token:\r\n 118 kwargs = smoothly_deprecate_use_auth_token(fn_name=fn.__name__, has_token=has_token, kwargs=kwargs)\r\n--> 120 return fn(*args, **kwargs)\r\n\r\nTypeError: HfApi.upload_file() got an unexpected keyword argument 'identical_ok'\r\n```", "Feel free to update `datasets` and `huggingface-hub`, it should fix it :)", "I went ahead and upgraded both datasets and hub and still getting the same error\r\n", "Which version do you have ? It's been a while since it has been fixed" ]

[]

[]

[ "The docs for this PR live [here](https://moon-ci-docs.huggingface.co/docs/datasets/pr_5510). All of your documentation changes will be reflected on that endpoint.", "To the maintainer, sorry about the repeated run requests for formatting. Missed the `make style` outlined in contributing guidelines. ", "Anything I can do to get the workflow to run? @lhoestq ", "cc @mariosasko \r\n\r\n> Anything I can do to get the workflow to run?\r\n\r\nYou can merge `main` into your branch to fix code formatting (we switched from isort+flake8 to ruff this week), and then run `make style`", "I believe that should be good. @mariosasko" ]

[]

[]

[ "The docs for this PR live [here](https://moon-ci-docs.huggingface.co/docs/datasets/pr_5509). All of your documentation changes will be reflected on that endpoint.", "Hi! I've commented on the original issue to provide some context. Feel free to share your opinion there." ]

[]

[]

[ "Hey, I'm a research engineer working on language modelling wanting to contribute to open source. I was wondering if I could give it a shot?", "Hi! This issue was fixed in https://github.com/huggingface/datasets/pull/4972, so please install `datasets>=2.5.0` to avoid it." ]

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

[ { "avatar_url": "https://avatars.githubusercontent.com/u/47462742?v=4", "events_url": "https://api.github.com/users/mariosasko/events{/privacy}", "followers_url": "https://api.github.com/users/mariosasko/followers", "following_url": "https://api.github.com/users/mariosasko/following{/other_user}", "gists_url": "https://api.github.com/users/mariosasko/gists{/gist_id}", "gravatar_id": "", "html_url": "https://github.com/mariosasko", "id": 47462742, "login": "mariosasko", "node_id": "MDQ6VXNlcjQ3NDYyNzQy", "organizations_url": "https://api.github.com/users/mariosasko/orgs", "received_events_url": "https://api.github.com/users/mariosasko/received_events", "repos_url": "https://api.github.com/users/mariosasko/repos", "site_admin": false, "starred_url": "https://api.github.com/users/mariosasko/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/mariosasko/subscriptions", "type": "User", "url": "https://api.github.com/users/mariosasko" } ]

[]

[]

[]

[ "Hi ! `datasets` doesn't do batching - the PyTorch DataLoader does and is created by the `Trainer`. Do you pass other arguments to training_args with respect to data loading ?\r\n\r\nAlso we recently released `.to_iterable_dataset` that does pretty much what you implemented, but using contiguous shards to get a better speed:\r\n```python\r\nif use_iterable_dataset:\r\n num_shards = 100\r\n dataset = dataset.to_iterable_dataset(num_shards=num_shards)\r\n```", "This is the full set of training args passed. No training args were changed when switching dataset types.\r\n\r\n```python\r\ntraining_args = TrainingArguments(\r\n output_dir=\"./checkpoints\",\r\n overwrite_output_dir=True,\r\n num_train_epochs=1,\r\n per_device_train_batch_size=256,\r\n save_steps=2000,\r\n save_total_limit=4,\r\n prediction_loss_only=True,\r\n report_to='none',\r\n gradient_accumulation_steps=6,\r\n fp16=True,\r\n max_steps=60000,\r\n lr_scheduler_type='linear',\r\n warmup_ratio=0.1,\r\n logging_steps=100,\r\n weight_decay=0.01,\r\n adam_beta1=0.9,\r\n adam_beta2=0.98,\r\n adam_epsilon=1e-6,\r\n learning_rate=1e-4\r\n)\r\n```", "I think the issue comes from `transformers`: https://github.com/huggingface/transformers/issues/21444", "Makes sense. Given that it's a `transformers` issue and already being tracked, I'll close this out." ]

[]

[]

[ "This change seems to come from a few months ago in the PyTorch side. That's good news and it means we may not need to pass a batch_sampler as soon as we add `Dataset.__getitems__` to get the optimal speed :)\r\n\r\nThanks for reporting ! Would you like to open a PR to add `__getitems__` and remove this outdated documentation ?", "Yeah I figured this was the sort of thing that probably once worked. I can confirm that you no longer need the batch sampler, just `batch_size=n` in the `DataLoader`.\r\n\r\nI'll pass on the PR, I'm flat out right now, sorry." ]

[]

[]

[ "_The documentation is not available anymore as the PR was closed or merged._", "<details>\n<summary>Show benchmarks</summary>\n\nPyArrow==6.0.0\n\n<details>\n<summary>Show updated benchmarks!</summary>\n\n### Benchmark: benchmark_array_xd.json\n\n| metric | read_batch_formatted_as_numpy after write_array2d | read_batch_formatted_as_numpy after write_flattened_sequence | read_batch_formatted_as_numpy after write_nested_sequence | read_batch_unformated after write_array2d | read_batch_unformated after write_flattened_sequence | read_batch_unformated after write_nested_sequence | read_col_formatted_as_numpy after write_array2d | read_col_formatted_as_numpy after write_flattened_sequence | read_col_formatted_as_numpy after write_nested_sequence | read_col_unformated after write_array2d | read_col_unformated after write_flattened_sequence | read_col_unformated after write_nested_sequence | read_formatted_as_numpy after write_array2d | read_formatted_as_numpy after write_flattened_sequence | read_formatted_as_numpy after write_nested_sequence | read_unformated after write_array2d | read_unformated after write_flattened_sequence | read_unformated after write_nested_sequence | write_array2d | write_flattened_sequence | write_nested_sequence |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.008606 / 0.011353 (-0.002747) | 0.004659 / 0.011008 (-0.006349) | 0.101311 / 0.038508 (0.062802) | 0.029664 / 0.023109 (0.006555) | 0.321850 / 0.275898 (0.045952) | 0.380497 / 0.323480 (0.057017) | 0.007003 / 0.007986 (-0.000982) | 0.003393 / 0.004328 (-0.000936) | 0.078704 / 0.004250 (0.074453) | 0.035810 / 0.037052 (-0.001242) | 0.327271 / 0.258489 (0.068782) | 0.369302 / 0.293841 (0.075461) | 0.033625 / 0.128546 (-0.094921) | 0.011563 / 0.075646 (-0.064084) | 0.323950 / 0.419271 (-0.095322) | 0.040660 / 0.043533 (-0.002872) | 0.327211 / 0.255139 (0.072072) | 0.350325 / 0.283200 (0.067125) | 0.085427 / 0.141683 (-0.056256) | 1.464370 / 1.452155 (0.012216) | 1.490355 / 1.492716 (-0.002362) |\n\n### Benchmark: benchmark_getitem\\_100B.json\n\n| metric | get_batch_of\\_1024\\_random_rows | get_batch_of\\_1024\\_rows | get_first_row | get_last_row |\n|--------|---|---|---|---|\n| new / old (diff) | 0.202879 / 0.018006 (0.184873) | 0.419836 / 0.000490 (0.419346) | 0.000303 / 0.000200 (0.000103) | 0.000063 / 0.000054 (0.000008) |\n\n### Benchmark: benchmark_indices_mapping.json\n\n| metric | select | shard | shuffle | sort | train_test_split |\n|--------|---|---|---|---|---|\n| new / old (diff) | 0.023336 / 0.037411 (-0.014075) | 0.096817 / 0.014526 (0.082291) | 0.103990 / 0.176557 (-0.072567) | 0.137749 / 0.737135 (-0.599386) | 0.108236 / 0.296338 (-0.188102) |\n\n### Benchmark: benchmark_iterating.json\n\n| metric | read 5000 | read 50000 | read_batch 50000 10 | read_batch 50000 100 | read_batch 50000 1000 | read_formatted numpy 5000 | read_formatted pandas 5000 | read_formatted tensorflow 5000 | read_formatted torch 5000 | read_formatted_batch numpy 5000 10 | read_formatted_batch numpy 5000 1000 | shuffled read 5000 | shuffled read 50000 | shuffled read_batch 50000 10 | shuffled read_batch 50000 100 | shuffled read_batch 50000 1000 | shuffled read_formatted numpy 5000 | shuffled read_formatted_batch numpy 5000 10 | shuffled read_formatted_batch numpy 5000 1000 |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.420801 / 0.215209 (0.205592) | 4.205308 / 2.077655 (2.127653) | 2.050363 / 1.504120 (0.546243) | 1.877390 / 1.541195 (0.336195) | 2.031060 / 1.468490 (0.562570) | 0.687950 / 4.584777 (-3.896827) | 3.363202 / 3.745712 (-0.382510) | 1.869482 / 5.269862 (-3.400379) | 1.159131 / 4.565676 (-3.406545) | 0.082374 / 0.424275 (-0.341901) | 0.012425 / 0.007607 (0.004818) | 0.519775 / 0.226044 (0.293731) | 5.244612 / 2.268929 (2.975684) | 2.371314 / 55.444624 (-53.073311) | 2.052713 / 6.876477 (-4.823764) | 2.190015 / 2.142072 (0.047942) | 0.803806 / 4.805227 (-4.001421) | 0.148110 / 6.500664 (-6.352554) | 0.064174 / 0.075469 (-0.011295) |\n\n### Benchmark: benchmark_map_filter.json\n\n| metric | filter | map fast-tokenizer batched | map identity | map identity batched | map no-op batched | map no-op batched numpy | map no-op batched pandas | map no-op batched pytorch | map no-op batched tensorflow |\n|--------|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 1.250424 / 1.841788 (-0.591364) | 13.487870 / 8.074308 (5.413561) | 13.080736 / 10.191392 (2.889344) | 0.147715 / 0.680424 (-0.532709) | 0.028409 / 0.534201 (-0.505792) | 0.397531 / 0.579283 (-0.181752) | 0.399458 / 0.434364 (-0.034905) | 0.461467 / 0.540337 (-0.078871) | 0.541639 / 1.386936 (-0.845297) |\n\n</details>\nPyArrow==latest\n\n<details>\n<summary>Show updated benchmarks!</summary>\n\n### Benchmark: benchmark_array_xd.json\n\n| metric | read_batch_formatted_as_numpy after write_array2d | read_batch_formatted_as_numpy after write_flattened_sequence | read_batch_formatted_as_numpy after write_nested_sequence | read_batch_unformated after write_array2d | read_batch_unformated after write_flattened_sequence | read_batch_unformated after write_nested_sequence | read_col_formatted_as_numpy after write_array2d | read_col_formatted_as_numpy after write_flattened_sequence | read_col_formatted_as_numpy after write_nested_sequence | read_col_unformated after write_array2d | read_col_unformated after write_flattened_sequence | read_col_unformated after write_nested_sequence | read_formatted_as_numpy after write_array2d | read_formatted_as_numpy after write_flattened_sequence | read_formatted_as_numpy after write_nested_sequence | read_unformated after write_array2d | read_unformated after write_flattened_sequence | read_unformated after write_nested_sequence | write_array2d | write_flattened_sequence | write_nested_sequence |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.006753 / 0.011353 (-0.004600) | 0.004573 / 0.011008 (-0.006435) | 0.076122 / 0.038508 (0.037614) | 0.027529 / 0.023109 (0.004419) | 0.341291 / 0.275898 (0.065393) | 0.376889 / 0.323480 (0.053409) | 0.005032 / 0.007986 (-0.002953) | 0.003447 / 0.004328 (-0.000882) | 0.075186 / 0.004250 (0.070936) | 0.038516 / 0.037052 (0.001463) | 0.340927 / 0.258489 (0.082438) | 0.386626 / 0.293841 (0.092785) | 0.031929 / 0.128546 (-0.096617) | 0.011759 / 0.075646 (-0.063888) | 0.085616 / 0.419271 (-0.333656) | 0.042858 / 0.043533 (-0.000674) | 0.341881 / 0.255139 (0.086742) | 0.367502 / 0.283200 (0.084303) | 0.090788 / 0.141683 (-0.050895) | 1.472871 / 1.452155 (0.020716) | 1.577825 / 1.492716 (0.085109) |\n\n### Benchmark: benchmark_getitem\\_100B.json\n\n| metric | get_batch_of\\_1024\\_random_rows | get_batch_of\\_1024\\_rows | get_first_row | get_last_row |\n|--------|---|---|---|---|\n| new / old (diff) | 0.233137 / 0.018006 (0.215131) | 0.415016 / 0.000490 (0.414526) | 0.000379 / 0.000200 (0.000179) | 0.000059 / 0.000054 (0.000004) |\n\n### Benchmark: benchmark_indices_mapping.json\n\n| metric | select | shard | shuffle | sort | train_test_split |\n|--------|---|---|---|---|---|\n| new / old (diff) | 0.024966 / 0.037411 (-0.012445) | 0.102794 / 0.014526 (0.088268) | 0.107543 / 0.176557 (-0.069014) | 0.143133 / 0.737135 (-0.594002) | 0.111494 / 0.296338 (-0.184845) |\n\n### Benchmark: benchmark_iterating.json\n\n| metric | read 5000 | read 50000 | read_batch 50000 10 | read_batch 50000 100 | read_batch 50000 1000 | read_formatted numpy 5000 | read_formatted pandas 5000 | read_formatted tensorflow 5000 | read_formatted torch 5000 | read_formatted_batch numpy 5000 10 | read_formatted_batch numpy 5000 1000 | shuffled read 5000 | shuffled read 50000 | shuffled read_batch 50000 10 | shuffled read_batch 50000 100 | shuffled read_batch 50000 1000 | shuffled read_formatted numpy 5000 | shuffled read_formatted_batch numpy 5000 10 | shuffled read_formatted_batch numpy 5000 1000 |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.438354 / 0.215209 (0.223145) | 4.382244 / 2.077655 (2.304589) | 2.056340 / 1.504120 (0.552220) | 1.851524 / 1.541195 (0.310330) | 1.933147 / 1.468490 (0.464657) | 0.701446 / 4.584777 (-3.883331) | 3.396893 / 3.745712 (-0.348819) | 2.837516 / 5.269862 (-2.432346) | 1.538298 / 4.565676 (-3.027379) | 0.083449 / 0.424275 (-0.340826) | 0.012793 / 0.007607 (0.005186) | 0.539661 / 0.226044 (0.313616) | 5.428415 / 2.268929 (3.159487) | 2.527582 / 55.444624 (-52.917042) | 2.172795 / 6.876477 (-4.703682) | 2.220011 / 2.142072 (0.077938) | 0.814338 / 4.805227 (-3.990889) | 0.153468 / 6.500664 (-6.347196) | 0.069056 / 0.075469 (-0.006413) |\n\n### Benchmark: benchmark_map_filter.json\n\n| metric | filter | map fast-tokenizer batched | map identity | map identity batched | map no-op batched | map no-op batched numpy | map no-op batched pandas | map no-op batched pytorch | map no-op batched tensorflow |\n|--------|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 1.278434 / 1.841788 (-0.563354) | 14.284924 / 8.074308 (6.210616) | 13.486596 / 10.191392 (3.295203) | 0.138457 / 0.680424 (-0.541967) | 0.016609 / 0.534201 (-0.517592) | 0.382828 / 0.579283 (-0.196455) | 0.387604 / 0.434364 (-0.046760) | 0.478801 / 0.540337 (-0.061536) | 0.565352 / 1.386936 (-0.821584) |\n\n</details>\n</details>\n\n![](https://cml.dev/watermark.png#c39ba501daab763b9972f44f229c66d900d20bee \"CML watermark\")\n", "> Thanks! I modified the test a bit to make it more consistent with the rest of the \"extractor\" tests.\r\n\r\nAppreciate the assist on the tests! 🚀 " ]

[]

[]

[ "_The documentation is not available anymore as the PR was closed or merged._", "> Thanks! I've left some comments.\r\n> \r\n> We should also add some tests, mainly to make sure `reverse` behaves as expected. Let me know if you need help with that.\r\n\r\nThanks for the offer! I couldn't find any guidelines on how huggingface goes about testing, so it would indeed be great to get a few pointers on that. I assume I should expand on the `test_sort` function in `test_arrow_dataset.py` but since I am not very familiar with the `datasets` package, it isn't immediately for which cases I should test (i.e., expand on).", "@MichlF \r\n\r\nResolving a comment means that the comment has been addressed with the code change, so since this is not the case here, can you please \"unresolve\" the comments and address them adequately? \r\n\r\n> I assume I should expand on the `test_sort` function in `test_arrow_dataset.py`\r\n\r\nYes, that's correct. I think one test to check sorting on multiple keys and another one to check if an error is raised when `len(reverse)!=len(column_names)` should be enough.\r\n", "> Yes, that's correct. I think one test to check sorting on multiple keys and another one to check if an error is raised when `len(reverse)!=len(column_names)` should be enough.\r\n\r\nI have added the tests in https://github.com/huggingface/datasets/pull/5502/commits/0efa259732e822e94d67b96a70031a3daccedfc1 by keeping them in the same format of the tests of the old `sort` function. Let me know if they can be improved.\r\n", "<details>\n<summary>Show benchmarks</summary>\n\nPyArrow==6.0.0\n\n<details>\n<summary>Show updated benchmarks!</summary>\n\n### Benchmark: benchmark_array_xd.json\n\n| metric | read_batch_formatted_as_numpy after write_array2d | read_batch_formatted_as_numpy after write_flattened_sequence | read_batch_formatted_as_numpy after write_nested_sequence | read_batch_unformated after write_array2d | read_batch_unformated after write_flattened_sequence | read_batch_unformated after write_nested_sequence | read_col_formatted_as_numpy after write_array2d | read_col_formatted_as_numpy after write_flattened_sequence | read_col_formatted_as_numpy after write_nested_sequence | read_col_unformated after write_array2d | read_col_unformated after write_flattened_sequence | read_col_unformated after write_nested_sequence | read_formatted_as_numpy after write_array2d | read_formatted_as_numpy after write_flattened_sequence | read_formatted_as_numpy after write_nested_sequence | read_unformated after write_array2d | read_unformated after write_flattened_sequence | read_unformated after write_nested_sequence | write_array2d | write_flattened_sequence | write_nested_sequence |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.010170 / 0.011353 (-0.001183) | 0.005891 / 0.011008 (-0.005117) | 0.100416 / 0.038508 (0.061908) | 0.041309 / 0.023109 (0.018200) | 0.300813 / 0.275898 (0.024915) | 0.376679 / 0.323480 (0.053199) | 0.008806 / 0.007986 (0.000821) | 0.005964 / 0.004328 (0.001636) | 0.075862 / 0.004250 (0.071611) | 0.050370 / 0.037052 (0.013318) | 0.313365 / 0.258489 (0.054876) | 0.351184 / 0.293841 (0.057343) | 0.039556 / 0.128546 (-0.088991) | 0.012462 / 0.075646 (-0.063185) | 0.337141 / 0.419271 (-0.082130) | 0.049678 / 0.043533 (0.006145) | 0.298547 / 0.255139 (0.043408) | 0.317547 / 0.283200 (0.034347) | 0.113595 / 0.141683 (-0.028088) | 1.448467 / 1.452155 (-0.003688) | 1.501303 / 1.492716 (0.008587) |\n\n### Benchmark: benchmark_getitem\\_100B.json\n\n| metric | get_batch_of\\_1024\\_random_rows | get_batch_of\\_1024\\_rows | get_first_row | get_last_row |\n|--------|---|---|---|---|\n| new / old (diff) | 0.011005 / 0.018006 (-0.007002) | 0.527430 / 0.000490 (0.526940) | 0.005073 / 0.000200 (0.004873) | 0.000100 / 0.000054 (0.000045) |\n\n### Benchmark: benchmark_indices_mapping.json\n\n| metric | select | shard | shuffle | sort | train_test_split |\n|--------|---|---|---|---|---|\n| new / old (diff) | 0.030377 / 0.037411 (-0.007034) | 0.116932 / 0.014526 (0.102406) | 0.124047 / 0.176557 (-0.052509) | 0.192358 / 0.737135 (-0.544777) | 0.130528 / 0.296338 (-0.165811) |\n\n### Benchmark: benchmark_iterating.json\n\n| metric | read 5000 | read 50000 | read_batch 50000 10 | read_batch 50000 100 | read_batch 50000 1000 | read_formatted numpy 5000 | read_formatted pandas 5000 | read_formatted tensorflow 5000 | read_formatted torch 5000 | read_formatted_batch numpy 5000 10 | read_formatted_batch numpy 5000 1000 | shuffled read 5000 | shuffled read 50000 | shuffled read_batch 50000 10 | shuffled read_batch 50000 100 | shuffled read_batch 50000 1000 | shuffled read_formatted numpy 5000 | shuffled read_formatted_batch numpy 5000 10 | shuffled read_formatted_batch numpy 5000 1000 |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.401158 / 0.215209 (0.185949) | 4.005854 / 2.077655 (1.928200) | 1.810365 / 1.504120 (0.306245) | 1.626490 / 1.541195 (0.085295) | 1.752591 / 1.468490 (0.284101) | 0.709065 / 4.584777 (-3.875712) | 3.893356 / 3.745712 (0.147643) | 3.655180 / 5.269862 (-1.614682) | 1.873660 / 4.565676 (-2.692017) | 0.085860 / 0.424275 (-0.338415) | 0.012671 / 0.007607 (0.005063) | 0.512804 / 0.226044 (0.286759) | 5.103426 / 2.268929 (2.834497) | 2.336148 / 55.444624 (-53.108477) | 2.000140 / 6.876477 (-4.876336) | 2.095155 / 2.142072 (-0.046918) | 0.848612 / 4.805227 (-3.956615) | 0.171840 / 6.500664 (-6.328824) | 0.064144 / 0.075469 (-0.011325) |\n\n### Benchmark: benchmark_map_filter.json\n\n| metric | filter | map fast-tokenizer batched | map identity | map identity batched | map no-op batched | map no-op batched numpy | map no-op batched pandas | map no-op batched pytorch | map no-op batched tensorflow |\n|--------|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 1.222106 / 1.841788 (-0.619682) | 15.828559 / 8.074308 (7.754251) | 14.995298 / 10.191392 (4.803906) | 0.172783 / 0.680424 (-0.507641) | 0.029296 / 0.534201 (-0.504905) | 0.447469 / 0.579283 (-0.131814) | 0.658615 / 0.434364 (0.224251) | 1.527607 / 0.540337 (0.987270) | 1.830018 / 1.386936 (0.443082) |\n\n</details>\nPyArrow==latest\n\n<details>\n<summary>Show updated benchmarks!</summary>\n\n### Benchmark: benchmark_array_xd.json\n\n| metric | read_batch_formatted_as_numpy after write_array2d | read_batch_formatted_as_numpy after write_flattened_sequence | read_batch_formatted_as_numpy after write_nested_sequence | read_batch_unformated after write_array2d | read_batch_unformated after write_flattened_sequence | read_batch_unformated after write_nested_sequence | read_col_formatted_as_numpy after write_array2d | read_col_formatted_as_numpy after write_flattened_sequence | read_col_formatted_as_numpy after write_nested_sequence | read_col_unformated after write_array2d | read_col_unformated after write_flattened_sequence | read_col_unformated after write_nested_sequence | read_formatted_as_numpy after write_array2d | read_formatted_as_numpy after write_flattened_sequence | read_formatted_as_numpy after write_nested_sequence | read_unformated after write_array2d | read_unformated after write_flattened_sequence | read_unformated after write_nested_sequence | write_array2d | write_flattened_sequence | write_nested_sequence |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.007922 / 0.011353 (-0.003431) | 0.005369 / 0.011008 (-0.005639) | 0.076580 / 0.038508 (0.038071) | 0.038770 / 0.023109 (0.015661) | 0.338995 / 0.275898 (0.063097) | 0.380865 / 0.323480 (0.057385) | 0.006489 / 0.007986 (-0.001497) | 0.004421 / 0.004328 (0.000093) | 0.074143 / 0.004250 (0.069893) | 0.054224 / 0.037052 (0.017171) | 0.348887 / 0.258489 (0.090397) | 0.395044 / 0.293841 (0.101203) | 0.037040 / 0.128546 (-0.091507) | 0.012547 / 0.075646 (-0.063099) | 0.087521 / 0.419271 (-0.331751) | 0.049918 / 0.043533 (0.006385) | 0.342428 / 0.255139 (0.087289) | 0.362216 / 0.283200 (0.079016) | 0.107204 / 0.141683 (-0.034479) | 1.509206 / 1.452155 (0.057052) | 1.596010 / 1.492716 (0.103293) |\n\n### Benchmark: benchmark_getitem\\_100B.json\n\n| metric | get_batch_of\\_1024\\_random_rows | get_batch_of\\_1024\\_rows | get_first_row | get_last_row |\n|--------|---|---|---|---|\n| new / old (diff) | 0.246795 / 0.018006 (0.228788) | 0.505998 / 0.000490 (0.505509) | 0.000446 / 0.000200 (0.000246) | 0.000064 / 0.000054 (0.000009) |\n\n### Benchmark: benchmark_indices_mapping.json\n\n| metric | select | shard | shuffle | sort | train_test_split |\n|--------|---|---|---|---|---|\n| new / old (diff) | 0.031591 / 0.037411 (-0.005821) | 0.117595 / 0.014526 (0.103069) | 0.132500 / 0.176557 (-0.044056) | 0.202244 / 0.737135 (-0.534891) | 0.136624 / 0.296338 (-0.159715) |\n\n### Benchmark: benchmark_iterating.json\n\n| metric | read 5000 | read 50000 | read_batch 50000 10 | read_batch 50000 100 | read_batch 50000 1000 | read_formatted numpy 5000 | read_formatted pandas 5000 | read_formatted tensorflow 5000 | read_formatted torch 5000 | read_formatted_batch numpy 5000 10 | read_formatted_batch numpy 5000 1000 | shuffled read 5000 | shuffled read 50000 | shuffled read_batch 50000 10 | shuffled read_batch 50000 100 | shuffled read_batch 50000 1000 | shuffled read_formatted numpy 5000 | shuffled read_formatted_batch numpy 5000 10 | shuffled read_formatted_batch numpy 5000 1000 |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.428235 / 0.215209 (0.213026) | 4.262691 / 2.077655 (2.185036) | 2.057348 / 1.504120 (0.553228) | 1.928559 / 1.541195 (0.387364) | 2.120838 / 1.468490 (0.652347) | 0.706300 / 4.584777 (-3.878477) | 3.951828 / 3.745712 (0.206115) | 2.144218 / 5.269862 (-3.125644) | 1.359500 / 4.565676 (-3.206177) | 0.085404 / 0.424275 (-0.338872) | 0.012363 / 0.007607 (0.004756) | 0.529985 / 0.226044 (0.303941) | 5.295831 / 2.268929 (3.026903) | 2.522602 / 55.444624 (-52.922022) | 2.182850 / 6.876477 (-4.693627) | 2.270187 / 2.142072 (0.128114) | 0.841676 / 4.805227 (-3.963551) | 0.168366 / 6.500664 (-6.332298) | 0.065371 / 0.075469 (-0.010098) |\n\n### Benchmark: benchmark_map_filter.json\n\n| metric | filter | map fast-tokenizer batched | map identity | map identity batched | map no-op batched | map no-op batched numpy | map no-op batched pandas | map no-op batched pytorch | map no-op batched tensorflow |\n|--------|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 1.261464 / 1.841788 (-0.580324) | 17.010125 / 8.074308 (8.935817) | 14.304453 / 10.191392 (4.113061) | 0.177782 / 0.680424 (-0.502642) | 0.017762 / 0.534201 (-0.516439) | 0.427283 / 0.579283 (-0.152000) | 0.455176 / 0.434364 (0.020812) | 0.525962 / 0.540337 (-0.014375) | 0.625583 / 1.386936 (-0.761353) |\n\n</details>\n</details>\n\n![](https://cml.dev/watermark.png#3b2aba6637dc61f145acda40e4e7b028c3947d72 \"CML watermark\")\n" ]

[]

[]

[ "The docs for this PR live [here](https://moon-ci-docs.huggingface.co/docs/datasets/pr_5501). All of your documentation changes will be reflected on that endpoint.", "<details>\n<summary>Show benchmarks</summary>\n\nPyArrow==6.0.0\n\n<details>\n<summary>Show updated benchmarks!</summary>\n\n### Benchmark: benchmark_array_xd.json\n\n| metric | read_batch_formatted_as_numpy after write_array2d | read_batch_formatted_as_numpy after write_flattened_sequence | read_batch_formatted_as_numpy after write_nested_sequence | read_batch_unformated after write_array2d | read_batch_unformated after write_flattened_sequence | read_batch_unformated after write_nested_sequence | read_col_formatted_as_numpy after write_array2d | read_col_formatted_as_numpy after write_flattened_sequence | read_col_formatted_as_numpy after write_nested_sequence | read_col_unformated after write_array2d | read_col_unformated after write_flattened_sequence | read_col_unformated after write_nested_sequence | read_formatted_as_numpy after write_array2d | read_formatted_as_numpy after write_flattened_sequence | read_formatted_as_numpy after write_nested_sequence | read_unformated after write_array2d | read_unformated after write_flattened_sequence | read_unformated after write_nested_sequence | write_array2d | write_flattened_sequence | write_nested_sequence |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.008407 / 0.011353 (-0.002946) | 0.004651 / 0.011008 (-0.006357) | 0.100367 / 0.038508 (0.061859) | 0.029107 / 0.023109 (0.005998) | 0.302798 / 0.275898 (0.026900) | 0.354379 / 0.323480 (0.030899) | 0.006985 / 0.007986 (-0.001001) | 0.003365 / 0.004328 (-0.000963) | 0.078312 / 0.004250 (0.074062) | 0.034205 / 0.037052 (-0.002847) | 0.310431 / 0.258489 (0.051941) | 0.346239 / 0.293841 (0.052398) | 0.033800 / 0.128546 (-0.094747) | 0.011515 / 0.075646 (-0.064131) | 0.323588 / 0.419271 (-0.095684) | 0.040766 / 0.043533 (-0.002767) | 0.300914 / 0.255139 (0.045775) | 0.332983 / 0.283200 (0.049784) | 0.087500 / 0.141683 (-0.054182) | 1.469505 / 1.452155 (0.017350) | 1.505119 / 1.492716 (0.012403) |\n\n### Benchmark: benchmark_getitem\\_100B.json\n\n| metric | get_batch_of\\_1024\\_random_rows | get_batch_of\\_1024\\_rows | get_first_row | get_last_row |\n|--------|---|---|---|---|\n| new / old (diff) | 0.187319 / 0.018006 (0.169313) | 0.405498 / 0.000490 (0.405008) | 0.001000 / 0.000200 (0.000800) | 0.000069 / 0.000054 (0.000015) |\n\n### Benchmark: benchmark_indices_mapping.json\n\n| metric | select | shard | shuffle | sort | train_test_split |\n|--------|---|---|---|---|---|\n| new / old (diff) | 0.022583 / 0.037411 (-0.014828) | 0.098096 / 0.014526 (0.083570) | 0.104272 / 0.176557 (-0.072284) | 0.142801 / 0.737135 (-0.594335) | 0.109749 / 0.296338 (-0.186590) |\n\n### Benchmark: benchmark_iterating.json\n\n| metric | read 5000 | read 50000 | read_batch 50000 10 | read_batch 50000 100 | read_batch 50000 1000 | read_formatted numpy 5000 | read_formatted pandas 5000 | read_formatted tensorflow 5000 | read_formatted torch 5000 | read_formatted_batch numpy 5000 10 | read_formatted_batch numpy 5000 1000 | shuffled read 5000 | shuffled read 50000 | shuffled read_batch 50000 10 | shuffled read_batch 50000 100 | shuffled read_batch 50000 1000 | shuffled read_formatted numpy 5000 | shuffled read_formatted_batch numpy 5000 10 | shuffled read_formatted_batch numpy 5000 1000 |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.423343 / 0.215209 (0.208134) | 4.215116 / 2.077655 (2.137461) | 1.899714 / 1.504120 (0.395594) | 1.689579 / 1.541195 (0.148384) | 1.710292 / 1.468490 (0.241801) | 0.690976 / 4.584777 (-3.893801) | 3.432501 / 3.745712 (-0.313212) | 1.899600 / 5.269862 (-3.370261) | 1.279801 / 4.565676 (-3.285876) | 0.082763 / 0.424275 (-0.341512) | 0.012545 / 0.007607 (0.004938) | 0.531381 / 0.226044 (0.305336) | 5.320077 / 2.268929 (3.051148) | 2.370705 / 55.444624 (-53.073919) | 2.007089 / 6.876477 (-4.869388) | 2.062412 / 2.142072 (-0.079661) | 0.814998 / 4.805227 (-3.990229) | 0.149822 / 6.500664 (-6.350842) | 0.064399 / 0.075469 (-0.011070) |\n\n### Benchmark: benchmark_map_filter.json\n\n| metric | filter | map fast-tokenizer batched | map identity | map identity batched | map no-op batched | map no-op batched numpy | map no-op batched pandas | map no-op batched pytorch | map no-op batched tensorflow |\n|--------|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 1.226196 / 1.841788 (-0.615591) | 13.823443 / 8.074308 (5.749134) | 13.813667 / 10.191392 (3.622275) | 0.161289 / 0.680424 (-0.519135) | 0.028569 / 0.534201 (-0.505632) | 0.390360 / 0.579283 (-0.188923) | 0.396217 / 0.434364 (-0.038147) | 0.483120 / 0.540337 (-0.057217) | 0.570041 / 1.386936 (-0.816895) |\n\n</details>\nPyArrow==latest\n\n<details>\n<summary>Show updated benchmarks!</summary>\n\n### Benchmark: benchmark_array_xd.json\n\n| metric | read_batch_formatted_as_numpy after write_array2d | read_batch_formatted_as_numpy after write_flattened_sequence | read_batch_formatted_as_numpy after write_nested_sequence | read_batch_unformated after write_array2d | read_batch_unformated after write_flattened_sequence | read_batch_unformated after write_nested_sequence | read_col_formatted_as_numpy after write_array2d | read_col_formatted_as_numpy after write_flattened_sequence | read_col_formatted_as_numpy after write_nested_sequence | read_col_unformated after write_array2d | read_col_unformated after write_flattened_sequence | read_col_unformated after write_nested_sequence | read_formatted_as_numpy after write_array2d | read_formatted_as_numpy after write_flattened_sequence | read_formatted_as_numpy after write_nested_sequence | read_unformated after write_array2d | read_unformated after write_flattened_sequence | read_unformated after write_nested_sequence | write_array2d | write_flattened_sequence | write_nested_sequence |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.006422 / 0.011353 (-0.004931) | 0.004528 / 0.011008 (-0.006481) | 0.076043 / 0.038508 (0.037535) | 0.027631 / 0.023109 (0.004522) | 0.340622 / 0.275898 (0.064724) | 0.376694 / 0.323480 (0.053214) | 0.004993 / 0.007986 (-0.002992) | 0.003403 / 0.004328 (-0.000926) | 0.074521 / 0.004250 (0.070270) | 0.037568 / 0.037052 (0.000516) | 0.343423 / 0.258489 (0.084934) | 0.387729 / 0.293841 (0.093888) | 0.031790 / 0.128546 (-0.096757) | 0.011767 / 0.075646 (-0.063879) | 0.085182 / 0.419271 (-0.334090) | 0.042867 / 0.043533 (-0.000666) | 0.341269 / 0.255139 (0.086130) | 0.368460 / 0.283200 (0.085261) | 0.090153 / 0.141683 (-0.051530) | 1.536490 / 1.452155 (0.084335) | 1.596403 / 1.492716 (0.103686) |\n\n### Benchmark: benchmark_getitem\\_100B.json\n\n| metric | get_batch_of\\_1024\\_random_rows | get_batch_of\\_1024\\_rows | get_first_row | get_last_row |\n|--------|---|---|---|---|\n| new / old (diff) | 0.222373 / 0.018006 (0.204367) | 0.396145 / 0.000490 (0.395655) | 0.000384 / 0.000200 (0.000184) | 0.000062 / 0.000054 (0.000008) |\n\n### Benchmark: benchmark_indices_mapping.json\n\n| metric | select | shard | shuffle | sort | train_test_split |\n|--------|---|---|---|---|---|\n| new / old (diff) | 0.024801 / 0.037411 (-0.012610) | 0.099711 / 0.014526 (0.085185) | 0.106094 / 0.176557 (-0.070463) | 0.147819 / 0.737135 (-0.589316) | 0.110065 / 0.296338 (-0.186274) |\n\n### Benchmark: benchmark_iterating.json\n\n| metric | read 5000 | read 50000 | read_batch 50000 10 | read_batch 50000 100 | read_batch 50000 1000 | read_formatted numpy 5000 | read_formatted pandas 5000 | read_formatted tensorflow 5000 | read_formatted torch 5000 | read_formatted_batch numpy 5000 10 | read_formatted_batch numpy 5000 1000 | shuffled read 5000 | shuffled read 50000 | shuffled read_batch 50000 10 | shuffled read_batch 50000 100 | shuffled read_batch 50000 1000 | shuffled read_formatted numpy 5000 | shuffled read_formatted_batch numpy 5000 10 | shuffled read_formatted_batch numpy 5000 1000 |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.442863 / 0.215209 (0.227654) | 4.420043 / 2.077655 (2.342388) | 2.070136 / 1.504120 (0.566016) | 1.862363 / 1.541195 (0.321168) | 1.910890 / 1.468490 (0.442400) | 0.702570 / 4.584777 (-3.882207) | 3.435855 / 3.745712 (-0.309857) | 1.871290 / 5.269862 (-3.398572) | 1.169321 / 4.565676 (-3.396355) | 0.083674 / 0.424275 (-0.340601) | 0.012823 / 0.007607 (0.005216) | 0.539330 / 0.226044 (0.313285) | 5.403317 / 2.268929 (3.134389) | 2.536508 / 55.444624 (-52.908117) | 2.179629 / 6.876477 (-4.696847) | 2.207586 / 2.142072 (0.065514) | 0.812256 / 4.805227 (-3.992972) | 0.152915 / 6.500664 (-6.347749) | 0.068431 / 0.075469 (-0.007038) |\n\n### Benchmark: benchmark_map_filter.json\n\n| metric | filter | map fast-tokenizer batched | map identity | map identity batched | map no-op batched | map no-op batched numpy | map no-op batched pandas | map no-op batched pytorch | map no-op batched tensorflow |\n|--------|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 1.294982 / 1.841788 (-0.546806) | 13.912811 / 8.074308 (5.838503) | 13.415658 / 10.191392 (3.224266) | 0.149531 / 0.680424 (-0.530893) | 0.016785 / 0.534201 (-0.517416) | 0.381055 / 0.579283 (-0.198228) | 0.392084 / 0.434364 (-0.042280) | 0.472614 / 0.540337 (-0.067724) | 0.559799 / 1.386936 (-0.827137) |\n\n</details>\n</details>\n\n![](https://cml.dev/watermark.png#6ef20f9b71acbb387caab2d297d8c22ba3db3633 \"CML watermark\")\n", "We simply do GET requests to hf.co to download files from the Hub right now. We may switch to hfh when we update how we do caching \r\n\r\nYou can try on any dataset hosted on the hub like `imagenet-1k`", "<details>\n<summary>Show benchmarks</summary>\n\nPyArrow==6.0.0\n\n<details>\n<summary>Show updated benchmarks!</summary>\n\n### Benchmark: benchmark_array_xd.json\n\n| metric | read_batch_formatted_as_numpy after write_array2d | read_batch_formatted_as_numpy after write_flattened_sequence | read_batch_formatted_as_numpy after write_nested_sequence | read_batch_unformated after write_array2d | read_batch_unformated after write_flattened_sequence | read_batch_unformated after write_nested_sequence | read_col_formatted_as_numpy after write_array2d | read_col_formatted_as_numpy after write_flattened_sequence | read_col_formatted_as_numpy after write_nested_sequence | read_col_unformated after write_array2d | read_col_unformated after write_flattened_sequence | read_col_unformated after write_nested_sequence | read_formatted_as_numpy after write_array2d | read_formatted_as_numpy after write_flattened_sequence | read_formatted_as_numpy after write_nested_sequence | read_unformated after write_array2d | read_unformated after write_flattened_sequence | read_unformated after write_nested_sequence | write_array2d | write_flattened_sequence | write_nested_sequence |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.010931 / 0.011353 (-0.000422) | 0.005730 / 0.011008 (-0.005278) | 0.116653 / 0.038508 (0.078145) | 0.041439 / 0.023109 (0.018330) | 0.359559 / 0.275898 (0.083661) | 0.408398 / 0.323480 (0.084918) | 0.009193 / 0.007986 (0.001208) | 0.006024 / 0.004328 (0.001695) | 0.087743 / 0.004250 (0.083492) | 0.048636 / 0.037052 (0.011584) | 0.363133 / 0.258489 (0.104643) | 0.407144 / 0.293841 (0.113303) | 0.044610 / 0.128546 (-0.083936) | 0.014075 / 0.075646 (-0.061571) | 0.396506 / 0.419271 (-0.022766) | 0.057014 / 0.043533 (0.013482) | 0.358254 / 0.255139 (0.103115) | 0.399887 / 0.283200 (0.116687) | 0.115337 / 0.141683 (-0.026346) | 1.731655 / 1.452155 (0.279500) | 1.813276 / 1.492716 (0.320560) |\n\n### Benchmark: benchmark_getitem\\_100B.json\n\n| metric | get_batch_of\\_1024\\_random_rows | get_batch_of\\_1024\\_rows | get_first_row | get_last_row |\n|--------|---|---|---|---|\n| new / old (diff) | 0.210197 / 0.018006 (0.192191) | 0.475887 / 0.000490 (0.475397) | 0.003323 / 0.000200 (0.003123) | 0.000100 / 0.000054 (0.000045) |\n\n### Benchmark: benchmark_indices_mapping.json\n\n| metric | select | shard | shuffle | sort | train_test_split |\n|--------|---|---|---|---|---|\n| new / old (diff) | 0.031686 / 0.037411 (-0.005725) | 0.131167 / 0.014526 (0.116641) | 0.137919 / 0.176557 (-0.038637) | 0.184843 / 0.737135 (-0.552293) | 0.144998 / 0.296338 (-0.151340) |\n\n### Benchmark: benchmark_iterating.json\n\n| metric | read 5000 | read 50000 | read_batch 50000 10 | read_batch 50000 100 | read_batch 50000 1000 | read_formatted numpy 5000 | read_formatted pandas 5000 | read_formatted tensorflow 5000 | read_formatted torch 5000 | read_formatted_batch numpy 5000 10 | read_formatted_batch numpy 5000 1000 | shuffled read 5000 | shuffled read 50000 | shuffled read_batch 50000 10 | shuffled read_batch 50000 100 | shuffled read_batch 50000 1000 | shuffled read_formatted numpy 5000 | shuffled read_formatted_batch numpy 5000 10 | shuffled read_formatted_batch numpy 5000 1000 |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.471371 / 0.215209 (0.256162) | 4.693739 / 2.077655 (2.616084) | 2.251567 / 1.504120 (0.747447) | 1.993653 / 1.541195 (0.452458) | 2.053236 / 1.468490 (0.584746) | 0.809226 / 4.584777 (-3.775551) | 4.494120 / 3.745712 (0.748408) | 2.436921 / 5.269862 (-2.832940) | 1.541973 / 4.565676 (-3.023704) | 0.098401 / 0.424275 (-0.325874) | 0.014329 / 0.007607 (0.006722) | 0.597813 / 0.226044 (0.371769) | 5.964035 / 2.268929 (3.695107) | 2.709283 / 55.444624 (-52.735341) | 2.323537 / 6.876477 (-4.552940) | 2.401707 / 2.142072 (0.259635) | 0.976379 / 4.805227 (-3.828848) | 0.194638 / 6.500664 (-6.306026) | 0.076904 / 0.075469 (0.001435) |\n\n### Benchmark: benchmark_map_filter.json\n\n| metric | filter | map fast-tokenizer batched | map identity | map identity batched | map no-op batched | map no-op batched numpy | map no-op batched pandas | map no-op batched pytorch | map no-op batched tensorflow |\n|--------|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 1.516877 / 1.841788 (-0.324911) | 18.228010 / 8.074308 (10.153702) | 16.631750 / 10.191392 (6.440358) | 0.176030 / 0.680424 (-0.504394) | 0.033769 / 0.534201 (-0.500432) | 0.520511 / 0.579283 (-0.058773) | 0.531764 / 0.434364 (0.097400) | 0.648658 / 0.540337 (0.108321) | 0.779124 / 1.386936 (-0.607812) |\n\n</details>\nPyArrow==latest\n\n<details>\n<summary>Show updated benchmarks!</summary>\n\n### Benchmark: benchmark_array_xd.json\n\n| metric | read_batch_formatted_as_numpy after write_array2d | read_batch_formatted_as_numpy after write_flattened_sequence | read_batch_formatted_as_numpy after write_nested_sequence | read_batch_unformated after write_array2d | read_batch_unformated after write_flattened_sequence | read_batch_unformated after write_nested_sequence | read_col_formatted_as_numpy after write_array2d | read_col_formatted_as_numpy after write_flattened_sequence | read_col_formatted_as_numpy after write_nested_sequence | read_col_unformated after write_array2d | read_col_unformated after write_flattened_sequence | read_col_unformated after write_nested_sequence | read_formatted_as_numpy after write_array2d | read_formatted_as_numpy after write_flattened_sequence | read_formatted_as_numpy after write_nested_sequence | read_unformated after write_array2d | read_unformated after write_flattened_sequence | read_unformated after write_nested_sequence | write_array2d | write_flattened_sequence | write_nested_sequence |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.008635 / 0.011353 (-0.002718) | 0.005785 / 0.011008 (-0.005223) | 0.087042 / 0.038508 (0.048534) | 0.039632 / 0.023109 (0.016523) | 0.419719 / 0.275898 (0.143821) | 0.463860 / 0.323480 (0.140380) | 0.006621 / 0.007986 (-0.001364) | 0.004655 / 0.004328 (0.000327) | 0.087003 / 0.004250 (0.082753) | 0.057122 / 0.037052 (0.020069) | 0.417820 / 0.258489 (0.159331) | 0.485981 / 0.293841 (0.192140) | 0.042606 / 0.128546 (-0.085940) | 0.014369 / 0.075646 (-0.061278) | 0.101939 / 0.419271 (-0.317333) | 0.058303 / 0.043533 (0.014770) | 0.415053 / 0.255139 (0.159914) | 0.439914 / 0.283200 (0.156714) | 0.134628 / 0.141683 (-0.007055) | 1.765464 / 1.452155 (0.313309) | 1.843963 / 1.492716 (0.351247) |\n\n### Benchmark: benchmark_getitem\\_100B.json\n\n| metric | get_batch_of\\_1024\\_random_rows | get_batch_of\\_1024\\_rows | get_first_row | get_last_row |\n|--------|---|---|---|---|\n| new / old (diff) | 0.307156 / 0.018006 (0.289150) | 0.476657 / 0.000490 (0.476167) | 0.019718 / 0.000200 (0.019518) | 0.000160 / 0.000054 (0.000105) |\n\n### Benchmark: benchmark_indices_mapping.json\n\n| metric | select | shard | shuffle | sort | train_test_split |\n|--------|---|---|---|---|---|\n| new / old (diff) | 0.035286 / 0.037411 (-0.002125) | 0.138094 / 0.014526 (0.123568) | 0.144768 / 0.176557 (-0.031789) | 0.191386 / 0.737135 (-0.545750) | 0.151988 / 0.296338 (-0.144350) |\n\n### Benchmark: benchmark_iterating.json\n\n| metric | read 5000 | read 50000 | read_batch 50000 10 | read_batch 50000 100 | read_batch 50000 1000 | read_formatted numpy 5000 | read_formatted pandas 5000 | read_formatted tensorflow 5000 | read_formatted torch 5000 | read_formatted_batch numpy 5000 10 | read_formatted_batch numpy 5000 1000 | shuffled read 5000 | shuffled read 50000 | shuffled read_batch 50000 10 | shuffled read_batch 50000 100 | shuffled read_batch 50000 1000 | shuffled read_formatted numpy 5000 | shuffled read_formatted_batch numpy 5000 10 | shuffled read_formatted_batch numpy 5000 1000 |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.504733 / 0.215209 (0.289523) | 5.027048 / 2.077655 (2.949394) | 2.441571 / 1.504120 (0.937451) | 2.198242 / 1.541195 (0.657047) | 2.298473 / 1.468490 (0.829983) | 0.848048 / 4.584777 (-3.736729) | 4.613102 / 3.745712 (0.867390) | 2.522824 / 5.269862 (-2.747037) | 1.610159 / 4.565676 (-2.955517) | 0.105197 / 0.424275 (-0.319078) | 0.015195 / 0.007607 (0.007588) | 0.626976 / 0.226044 (0.400932) | 6.268459 / 2.268929 (3.999530) | 3.014387 / 55.444624 (-52.430237) | 2.554102 / 6.876477 (-4.322375) | 2.656051 / 2.142072 (0.513979) | 1.027978 / 4.805227 (-3.777249) | 0.200686 / 6.500664 (-6.299978) | 0.077104 / 0.075469 (0.001635) |\n\n### Benchmark: benchmark_map_filter.json\n\n| metric | filter | map fast-tokenizer batched | map identity | map identity batched | map no-op batched | map no-op batched numpy | map no-op batched pandas | map no-op batched pytorch | map no-op batched tensorflow |\n|--------|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 1.485228 / 1.841788 (-0.356560) | 18.319949 / 8.074308 (10.245641) | 15.855739 / 10.191392 (5.664347) | 0.204365 / 0.680424 (-0.476059) | 0.023824 / 0.534201 (-0.510377) | 0.505000 / 0.579283 (-0.074283) | 0.502866 / 0.434364 (0.068502) | 0.629574 / 0.540337 (0.089237) | 0.746602 / 1.386936 (-0.640334) |\n\n</details>\n</details>\n\n![](https://cml.dev/watermark.png#900d429d3601657f766737b8670f855033078d57 \"CML watermark\")\n" ]

[]

[]

[ "I update the `datatsets` version and it works." ]

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

[]

[ "Hi ! To skip the verification step that checks if newer data exist, you can enable offline mode with `HF_DATASETS_OFFLINE=1`.\r\n\r\nAlthough I agree this step should be much faster for datasets hosted on the HF Hub - we could just compare the commit hash from the local data and the remote git repository. We're not been leveraging the git commit hashes, since the library was built before we even had git repositories for each dataset on HF.", "Thanks @lhoestq, for memory when I recorded those times I had `HF_DATASETS_OFFLINE` set." ]

[]

[]

[ "Hi! Instead of a single boolean, your filter function should return an iterable (of booleans) in the batched mode like so:\r\n```python\r\ntrain_dataset = train_dataset.filter(\r\n function=lambda batch: [image is not None for image in batch[\"image\"]], \r\n batched=True,\r\n batch_size=10)\r\n```\r\n\r\nPS: You can make this operation much faster by operating directly on the arrow data to skip the decoding part:\r\n```python\r\ntrain_dataset = train_dataset.with_format(\"arrow\")\r\ntrain_dataset = train_dataset.filter(\r\n function=lambda table: table[\"image\"].is_valid().to_pylist(), \r\n batched=True,\r\n batch_size=100)\r\ntrain_dataset = train_dataset.with_format(None)\r\n```", "Thank a lot!" ]

[]

[]

[ "_The documentation is not available anymore as the PR was closed or merged._", "<details>\n<summary>Show benchmarks</summary>\n\nPyArrow==6.0.0\n\n<details>\n<summary>Show updated benchmarks!</summary>\n\n### Benchmark: benchmark_array_xd.json\n\n| metric | read_batch_formatted_as_numpy after write_array2d | read_batch_formatted_as_numpy after write_flattened_sequence | read_batch_formatted_as_numpy after write_nested_sequence | read_batch_unformated after write_array2d | read_batch_unformated after write_flattened_sequence | read_batch_unformated after write_nested_sequence | read_col_formatted_as_numpy after write_array2d | read_col_formatted_as_numpy after write_flattened_sequence | read_col_formatted_as_numpy after write_nested_sequence | read_col_unformated after write_array2d | read_col_unformated after write_flattened_sequence | read_col_unformated after write_nested_sequence | read_formatted_as_numpy after write_array2d | read_formatted_as_numpy after write_flattened_sequence | read_formatted_as_numpy after write_nested_sequence | read_unformated after write_array2d | read_unformated after write_flattened_sequence | read_unformated after write_nested_sequence | write_array2d | write_flattened_sequence | write_nested_sequence |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.009491 / 0.011353 (-0.001862) | 0.004690 / 0.011008 (-0.006319) | 0.111904 / 0.038508 (0.073396) | 0.030781 / 0.023109 (0.007671) | 0.309442 / 0.275898 (0.033544) | 0.389511 / 0.323480 (0.066031) | 0.007277 / 0.007986 (-0.000709) | 0.004364 / 0.004328 (0.000036) | 0.074501 / 0.004250 (0.070250) | 0.036799 / 0.037052 (-0.000254) | 0.320279 / 0.258489 (0.061790) | 0.353887 / 0.293841 (0.060046) | 0.047969 / 0.128546 (-0.080577) | 0.017281 / 0.075646 (-0.058366) | 0.339655 / 0.419271 (-0.079617) | 0.049317 / 0.043533 (0.005784) | 0.321221 / 0.255139 (0.066082) | 0.354743 / 0.283200 (0.071544) | 0.098634 / 0.141683 (-0.043049) | 1.408640 / 1.452155 (-0.043515) | 1.488361 / 1.492716 (-0.004356) |\n\n### Benchmark: benchmark_getitem\\_100B.json\n\n| metric | get_batch_of\\_1024\\_random_rows | get_batch_of\\_1024\\_rows | get_first_row | get_last_row |\n|--------|---|---|---|---|\n| new / old (diff) | 0.233677 / 0.018006 (0.215671) | 0.604424 / 0.000490 (0.603934) | 0.003834 / 0.000200 (0.003634) | 0.000103 / 0.000054 (0.000049) |\n\n### Benchmark: benchmark_indices_mapping.json\n\n| metric | select | shard | shuffle | sort | train_test_split |\n|--------|---|---|---|---|---|\n| new / old (diff) | 0.022682 / 0.037411 (-0.014729) | 0.103800 / 0.014526 (0.089274) | 0.113868 / 0.176557 (-0.062689) | 0.155111 / 0.737135 (-0.582025) | 0.111862 / 0.296338 (-0.184476) |\n\n### Benchmark: benchmark_iterating.json\n\n| metric | read 5000 | read 50000 | read_batch 50000 10 | read_batch 50000 100 | read_batch 50000 1000 | read_formatted numpy 5000 | read_formatted pandas 5000 | read_formatted tensorflow 5000 | read_formatted torch 5000 | read_formatted_batch numpy 5000 10 | read_formatted_batch numpy 5000 1000 | shuffled read 5000 | shuffled read 50000 | shuffled read_batch 50000 10 | shuffled read_batch 50000 100 | shuffled read_batch 50000 1000 | shuffled read_formatted numpy 5000 | shuffled read_formatted_batch numpy 5000 10 | shuffled read_formatted_batch numpy 5000 1000 |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.474992 / 0.215209 (0.259783) | 4.755325 / 2.077655 (2.677670) | 1.889754 / 1.504120 (0.385634) | 1.597009 / 1.541195 (0.055814) | 1.639570 / 1.468490 (0.171080) | 0.970681 / 4.584777 (-3.614096) | 4.782567 / 3.745712 (1.036855) | 4.350465 / 5.269862 (-0.919397) | 2.413533 / 4.565676 (-2.152144) | 0.115510 / 0.424275 (-0.308765) | 0.011663 / 0.007607 (0.004055) | 0.626450 / 0.226044 (0.400406) | 6.238147 / 2.268929 (3.969218) | 2.603070 / 55.444624 (-52.841555) | 2.030378 / 6.876477 (-4.846099) | 1.996883 / 2.142072 (-0.145190) | 1.206436 / 4.805227 (-3.598792) | 0.203018 / 6.500664 (-6.297646) | 0.060550 / 0.075469 (-0.014919) |\n\n### Benchmark: benchmark_map_filter.json\n\n| metric | filter | map fast-tokenizer batched | map identity | map identity batched | map no-op batched | map no-op batched numpy | map no-op batched pandas | map no-op batched pytorch | map no-op batched tensorflow |\n|--------|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 1.259850 / 1.841788 (-0.581937) | 14.079936 / 8.074308 (6.005628) | 16.036329 / 10.191392 (5.844937) | 0.221546 / 0.680424 (-0.458878) | 0.042416 / 0.534201 (-0.491785) | 0.438851 / 0.579283 (-0.140432) | 0.507053 / 0.434364 (0.072689) | 0.518672 / 0.540337 (-0.021665) | 0.585278 / 1.386936 (-0.801659) |\n\n</details>\nPyArrow==latest\n\n<details>\n<summary>Show updated benchmarks!</summary>\n\n### Benchmark: benchmark_array_xd.json\n\n| metric | read_batch_formatted_as_numpy after write_array2d | read_batch_formatted_as_numpy after write_flattened_sequence | read_batch_formatted_as_numpy after write_nested_sequence | read_batch_unformated after write_array2d | read_batch_unformated after write_flattened_sequence | read_batch_unformated after write_nested_sequence | read_col_formatted_as_numpy after write_array2d | read_col_formatted_as_numpy after write_flattened_sequence | read_col_formatted_as_numpy after write_nested_sequence | read_col_unformated after write_array2d | read_col_unformated after write_flattened_sequence | read_col_unformated after write_nested_sequence | read_formatted_as_numpy after write_array2d | read_formatted_as_numpy after write_flattened_sequence | read_formatted_as_numpy after write_nested_sequence | read_unformated after write_array2d | read_unformated after write_flattened_sequence | read_unformated after write_nested_sequence | write_array2d | write_flattened_sequence | write_nested_sequence |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.010718 / 0.011353 (-0.000635) | 0.005469 / 0.011008 (-0.005539) | 0.075624 / 0.038508 (0.037116) | 0.029103 / 0.023109 (0.005994) | 0.353294 / 0.275898 (0.077395) | 0.353674 / 0.323480 (0.030194) | 0.005678 / 0.007986 (-0.002308) | 0.004610 / 0.004328 (0.000282) | 0.075213 / 0.004250 (0.070963) | 0.040032 / 0.037052 (0.002980) | 0.344363 / 0.258489 (0.085874) | 0.376861 / 0.293841 (0.083020) | 0.043718 / 0.128546 (-0.084828) | 0.016057 / 0.075646 (-0.059589) | 0.087746 / 0.419271 (-0.331526) | 0.051380 / 0.043533 (0.007848) | 0.336904 / 0.255139 (0.081765) | 0.357636 / 0.283200 (0.074436) | 0.089425 / 0.141683 (-0.052258) | 1.377462 / 1.452155 (-0.074692) | 1.448844 / 1.492716 (-0.043872) |\n\n### Benchmark: benchmark_getitem\\_100B.json\n\n| metric | get_batch_of\\_1024\\_random_rows | get_batch_of\\_1024\\_rows | get_first_row | get_last_row |\n|--------|---|---|---|---|\n| new / old (diff) | 0.259038 / 0.018006 (0.241031) | 0.512284 / 0.000490 (0.511794) | 0.005666 / 0.000200 (0.005466) | 0.000123 / 0.000054 (0.000068) |\n\n### Benchmark: benchmark_indices_mapping.json\n\n| metric | select | shard | shuffle | sort | train_test_split |\n|--------|---|---|---|---|---|\n| new / old (diff) | 0.023669 / 0.037411 (-0.013742) | 0.097979 / 0.014526 (0.083453) | 0.117947 / 0.176557 (-0.058610) | 0.140764 / 0.737135 (-0.596372) | 0.114700 / 0.296338 (-0.181638) |\n\n### Benchmark: benchmark_iterating.json\n\n| metric | read 5000 | read 50000 | read_batch 50000 10 | read_batch 50000 100 | read_batch 50000 1000 | read_formatted numpy 5000 | read_formatted pandas 5000 | read_formatted tensorflow 5000 | read_formatted torch 5000 | read_formatted_batch numpy 5000 10 | read_formatted_batch numpy 5000 1000 | shuffled read 5000 | shuffled read 50000 | shuffled read_batch 50000 10 | shuffled read_batch 50000 100 | shuffled read_batch 50000 1000 | shuffled read_formatted numpy 5000 | shuffled read_formatted_batch numpy 5000 10 | shuffled read_formatted_batch numpy 5000 1000 |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.528844 / 0.215209 (0.313635) | 5.073828 / 2.077655 (2.996173) | 2.088738 / 1.504120 (0.584618) | 1.855820 / 1.541195 (0.314626) | 1.838639 / 1.468490 (0.370149) | 0.968228 / 4.584777 (-3.616549) | 4.589792 / 3.745712 (0.844079) | 2.586149 / 5.269862 (-2.683712) | 1.714241 / 4.565676 (-2.851435) | 0.124502 / 0.424275 (-0.299774) | 0.012115 / 0.007607 (0.004507) | 0.679539 / 0.226044 (0.453494) | 6.541335 / 2.268929 (4.272407) | 2.749153 / 55.444624 (-52.695471) | 2.124164 / 6.876477 (-4.752313) | 2.181249 / 2.142072 (0.039177) | 1.196846 / 4.805227 (-3.608381) | 0.213352 / 6.500664 (-6.287312) | 0.075021 / 0.075469 (-0.000448) |\n\n### Benchmark: benchmark_map_filter.json\n\n| metric | filter | map fast-tokenizer batched | map identity | map identity batched | map no-op batched | map no-op batched numpy | map no-op batched pandas | map no-op batched pytorch | map no-op batched tensorflow |\n|--------|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 1.254301 / 1.841788 (-0.587487) | 14.494254 / 8.074308 (6.419946) | 16.619679 / 10.191392 (6.428287) | 0.205158 / 0.680424 (-0.475266) | 0.022181 / 0.534201 (-0.512019) | 0.422928 / 0.579283 (-0.156355) | 0.539825 / 0.434364 (0.105461) | 0.523165 / 0.540337 (-0.017173) | 0.615014 / 1.386936 (-0.771922) |\n\n</details>\n</details>\n\n![](https://cml.dev/watermark.png#e4d8a3d43569d61e73f7ab12ff3a6b48466afa8d \"CML watermark\")\n", "<details>\n<summary>Show benchmarks</summary>\n\nPyArrow==6.0.0\n\n<details>\n<summary>Show updated benchmarks!</summary>\n\n### Benchmark: benchmark_array_xd.json\n\n| metric | read_batch_formatted_as_numpy after write_array2d | read_batch_formatted_as_numpy after write_flattened_sequence | read_batch_formatted_as_numpy after write_nested_sequence | read_batch_unformated after write_array2d | read_batch_unformated after write_flattened_sequence | read_batch_unformated after write_nested_sequence | read_col_formatted_as_numpy after write_array2d | read_col_formatted_as_numpy after write_flattened_sequence | read_col_formatted_as_numpy after write_nested_sequence | read_col_unformated after write_array2d | read_col_unformated after write_flattened_sequence | read_col_unformated after write_nested_sequence | read_formatted_as_numpy after write_array2d | read_formatted_as_numpy after write_flattened_sequence | read_formatted_as_numpy after write_nested_sequence | read_unformated after write_array2d | read_unformated after write_flattened_sequence | read_unformated after write_nested_sequence | write_array2d | write_flattened_sequence | write_nested_sequence |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.011522 / 0.011353 (0.000169) | 0.006906 / 0.011008 (-0.004102) | 0.114692 / 0.038508 (0.076184) | 0.037686 / 0.023109 (0.014577) | 0.393662 / 0.275898 (0.117764) | 0.377730 / 0.323480 (0.054250) | 0.008212 / 0.007986 (0.000226) | 0.005470 / 0.004328 (0.001142) | 0.086962 / 0.004250 (0.082712) | 0.039085 / 0.037052 (0.002033) | 0.357565 / 0.258489 (0.099076) | 0.404384 / 0.293841 (0.110543) | 0.055523 / 0.128546 (-0.073023) | 0.018277 / 0.075646 (-0.057369) | 0.389812 / 0.419271 (-0.029459) | 0.058706 / 0.043533 (0.015173) | 0.344735 / 0.255139 (0.089597) | 0.395734 / 0.283200 (0.112535) | 0.096098 / 0.141683 (-0.045584) | 1.546654 / 1.452155 (0.094499) | 1.665314 / 1.492716 (0.172597) |\n\n### Benchmark: benchmark_getitem\\_100B.json\n\n| metric | get_batch_of\\_1024\\_random_rows | get_batch_of\\_1024\\_rows | get_first_row | get_last_row |\n|--------|---|---|---|---|\n| new / old (diff) | 0.255893 / 0.018006 (0.237887) | 0.589563 / 0.000490 (0.589074) | 0.005890 / 0.000200 (0.005690) | 0.000123 / 0.000054 (0.000069) |\n\n### Benchmark: benchmark_indices_mapping.json\n\n| metric | select | shard | shuffle | sort | train_test_split |\n|--------|---|---|---|---|---|\n| new / old (diff) | 0.029167 / 0.037411 (-0.008245) | 0.113561 / 0.014526 (0.099036) | 0.125361 / 0.176557 (-0.051195) | 0.182225 / 0.737135 (-0.554910) | 0.125147 / 0.296338 (-0.171192) |\n\n### Benchmark: benchmark_iterating.json\n\n| metric | read 5000 | read 50000 | read_batch 50000 10 | read_batch 50000 100 | read_batch 50000 1000 | read_formatted numpy 5000 | read_formatted pandas 5000 | read_formatted tensorflow 5000 | read_formatted torch 5000 | read_formatted_batch numpy 5000 10 | read_formatted_batch numpy 5000 1000 | shuffled read 5000 | shuffled read 50000 | shuffled read_batch 50000 10 | shuffled read_batch 50000 100 | shuffled read_batch 50000 1000 | shuffled read_formatted numpy 5000 | shuffled read_formatted_batch numpy 5000 10 | shuffled read_formatted_batch numpy 5000 1000 |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.596859 / 0.215209 (0.381650) | 5.797725 / 2.077655 (3.720071) | 2.238420 / 1.504120 (0.734300) | 1.933177 / 1.541195 (0.391982) | 2.030750 / 1.468490 (0.562260) | 1.122655 / 4.584777 (-3.462122) | 5.247913 / 3.745712 (1.502201) | 2.792742 / 5.269862 (-2.477120) | 1.861487 / 4.565676 (-2.704190) | 0.133009 / 0.424275 (-0.291266) | 0.013219 / 0.007607 (0.005612) | 0.696905 / 0.226044 (0.470861) | 6.961298 / 2.268929 (4.692369) | 2.895352 / 55.444624 (-52.549273) | 2.353677 / 6.876477 (-4.522799) | 2.458804 / 2.142072 (0.316731) | 1.271905 / 4.805227 (-3.533322) | 0.224850 / 6.500664 (-6.275814) | 0.083773 / 0.075469 (0.008304) |\n\n### Benchmark: benchmark_map_filter.json\n\n| metric | filter | map fast-tokenizer batched | map identity | map identity batched | map no-op batched | map no-op batched numpy | map no-op batched pandas | map no-op batched pytorch | map no-op batched tensorflow |\n|--------|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 1.502425 / 1.841788 (-0.339363) | 16.959241 / 8.074308 (8.884933) | 19.865569 / 10.191392 (9.674177) | 0.228608 / 0.680424 (-0.451816) | 0.044035 / 0.534201 (-0.490166) | 0.545172 / 0.579283 (-0.034112) | 0.677193 / 0.434364 (0.242829) | 0.608988 / 0.540337 (0.068650) | 0.719210 / 1.386936 (-0.667726) |\n\n</details>\nPyArrow==latest\n\n<details>\n<summary>Show updated benchmarks!</summary>\n\n### Benchmark: benchmark_array_xd.json\n\n| metric | read_batch_formatted_as_numpy after write_array2d | read_batch_formatted_as_numpy after write_flattened_sequence | read_batch_formatted_as_numpy after write_nested_sequence | read_batch_unformated after write_array2d | read_batch_unformated after write_flattened_sequence | read_batch_unformated after write_nested_sequence | read_col_formatted_as_numpy after write_array2d | read_col_formatted_as_numpy after write_flattened_sequence | read_col_formatted_as_numpy after write_nested_sequence | read_col_unformated after write_array2d | read_col_unformated after write_flattened_sequence | read_col_unformated after write_nested_sequence | read_formatted_as_numpy after write_array2d | read_formatted_as_numpy after write_flattened_sequence | read_formatted_as_numpy after write_nested_sequence | read_unformated after write_array2d | read_unformated after write_flattened_sequence | read_unformated after write_nested_sequence | write_array2d | write_flattened_sequence | write_nested_sequence |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.008297 / 0.011353 (-0.003056) | 0.005729 / 0.011008 (-0.005280) | 0.084762 / 0.038508 (0.046254) | 0.030622 / 0.023109 (0.007512) | 0.408017 / 0.275898 (0.132119) | 0.432114 / 0.323480 (0.108634) | 0.006965 / 0.007986 (-0.001021) | 0.004830 / 0.004328 (0.000502) | 0.087375 / 0.004250 (0.083124) | 0.048110 / 0.037052 (0.011058) | 0.414978 / 0.258489 (0.156489) | 0.446136 / 0.293841 (0.152295) | 0.064351 / 0.128546 (-0.064195) | 0.018273 / 0.075646 (-0.057374) | 0.114853 / 0.419271 (-0.304418) | 0.056962 / 0.043533 (0.013429) | 0.427791 / 0.255139 (0.172652) | 0.428829 / 0.283200 (0.145629) | 0.108004 / 0.141683 (-0.033679) | 1.639285 / 1.452155 (0.187130) | 1.652106 / 1.492716 (0.159390) |\n\n### Benchmark: benchmark_getitem\\_100B.json\n\n| metric | get_batch_of\\_1024\\_random_rows | get_batch_of\\_1024\\_rows | get_first_row | get_last_row |\n|--------|---|---|---|---|\n| new / old (diff) | 0.359744 / 0.018006 (0.341738) | 0.596060 / 0.000490 (0.595570) | 0.025448 / 0.000200 (0.025248) | 0.000158 / 0.000054 (0.000104) |\n\n### Benchmark: benchmark_indices_mapping.json\n\n| metric | select | shard | shuffle | sort | train_test_split |\n|--------|---|---|---|---|---|\n| new / old (diff) | 0.026348 / 0.037411 (-0.011064) | 0.119153 / 0.014526 (0.104628) | 0.129304 / 0.176557 (-0.047253) | 0.195670 / 0.737135 (-0.541465) | 0.135559 / 0.296338 (-0.160780) |\n\n### Benchmark: benchmark_iterating.json\n\n| metric | read 5000 | read 50000 | read_batch 50000 10 | read_batch 50000 100 | read_batch 50000 1000 | read_formatted numpy 5000 | read_formatted pandas 5000 | read_formatted tensorflow 5000 | read_formatted torch 5000 | read_formatted_batch numpy 5000 10 | read_formatted_batch numpy 5000 1000 | shuffled read 5000 | shuffled read 50000 | shuffled read_batch 50000 10 | shuffled read_batch 50000 100 | shuffled read_batch 50000 1000 | shuffled read_formatted numpy 5000 | shuffled read_formatted_batch numpy 5000 10 | shuffled read_formatted_batch numpy 5000 1000 |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.588963 / 0.215209 (0.373754) | 5.682957 / 2.077655 (3.605302) | 2.380178 / 1.504120 (0.876059) | 2.131299 / 1.541195 (0.590104) | 2.167839 / 1.468490 (0.699349) | 1.126418 / 4.584777 (-3.458359) | 5.289104 / 3.745712 (1.543392) | 2.952128 / 5.269862 (-2.317734) | 1.922974 / 4.565676 (-2.642702) | 0.143874 / 0.424275 (-0.280401) | 0.015399 / 0.007607 (0.007792) | 0.815675 / 0.226044 (0.589631) | 7.320146 / 2.268929 (5.051217) | 3.453670 / 55.444624 (-51.990954) | 2.579133 / 6.876477 (-4.297344) | 2.532331 / 2.142072 (0.390258) | 1.345881 / 4.805227 (-3.459347) | 0.242448 / 6.500664 (-6.258216) | 0.070007 / 0.075469 (-0.005462) |\n\n### Benchmark: benchmark_map_filter.json\n\n| metric | filter | map fast-tokenizer batched | map identity | map identity batched | map no-op batched | map no-op batched numpy | map no-op batched pandas | map no-op batched pytorch | map no-op batched tensorflow |\n|--------|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 1.433173 / 1.841788 (-0.408614) | 17.127287 / 8.074308 (9.052979) | 17.953878 / 10.191392 (7.762486) | 0.220035 / 0.680424 (-0.460389) | 0.028660 / 0.534201 (-0.505541) | 0.496233 / 0.579283 (-0.083050) | 0.591587 / 0.434364 (0.157223) | 0.635204 / 0.540337 (0.094867) | 0.702143 / 1.386936 (-0.684793) |\n\n</details>\n</details>\n\n![](https://cml.dev/watermark.png#7cfac43b980ab9e4a69c2328f085770996323005 \"CML watermark\")\n" ]

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

[]

[ "Hi! Sure, feel free to open a PR, so we can see the API you have in mind.", "I would like to give it a go! #self-assign" ]

[ { "color": "d73a4a", "default": true, "description": "Something isn't working", "id": 1935892857, "name": "bug", "node_id": "MDU6TGFiZWwxOTM1ODkyODU3", "url": "https://api.github.com/repos/huggingface/datasets/labels/bug" }, { "color": "7057ff", "default": true, "description": "Good for newcomers", "id": 1935892877, "name": "good first issue", "node_id": "MDU6TGFiZWwxOTM1ODkyODc3", "url": "https://api.github.com/repos/huggingface/datasets/labels/good%20first%20issue" } ]

[]

[ "Hi! This is indeed a bug in our zero-copy logic.\r\n\r\nTo fix it, instead of the line:\r\nhttps://github.com/huggingface/datasets/blob/7cfac43b980ab9e4a69c2328f085770996323005/src/datasets/features/features.py#L702\r\n\r\nwe should have:\r\n```python\r\nreturn pa.types.is_primitive(pa_type) and not (pa.types.is_boolean(pa_type) or pa.types.is_temporal(pa_type))\r\n```", "@mariosasko submitted a small PR [here](https://github.com/huggingface/datasets/pull/5504)" ]

[ { "color": "0075ca", "default": true, "description": "Improvements or additions to documentation", "id": 1935892861, "name": "documentation", "node_id": "MDU6TGFiZWwxOTM1ODkyODYx", "url": "https://api.github.com/repos/huggingface/datasets/labels/documentation" } ]

[]

[ "Totally agree, the docs should be in sync with our code.\r\n\r\nIndeed to avoid confusing users, I think we should have updated the docs at the same time as this PR:\r\n- #5167", "@albertvillanova yeah sure I should have, but I forgot back then, sorry for that 😶", "No, @polinaeterna, nothing to be sorry about.\r\n\r\nMy comment was for all of us datasets team, as a reminder: when making a PR, but also when reviewing some other's PR, we should not forget to update the corresponding docstring and doc pages. It is something we can improve if we help each other in reminding about it... :hugs: ", "@polinaeterna I think we can close this issue now as we no longer use `torchaudio` for decoding." ]

[]

[]

[ "_The documentation is not available anymore as the PR was closed or merged._", "The docs for this PR live [here](https://moon-ci-docs.huggingface.co/docs/datasets/pr_5493). All of your documentation changes will be reflected on that endpoint.", "<details>\n<summary>Show benchmarks</summary>\n\nPyArrow==6.0.0\n\n<details>\n<summary>Show updated benchmarks!</summary>\n\n### Benchmark: benchmark_array_xd.json\n\n| metric | read_batch_formatted_as_numpy after write_array2d | read_batch_formatted_as_numpy after write_flattened_sequence | read_batch_formatted_as_numpy after write_nested_sequence | read_batch_unformated after write_array2d | read_batch_unformated after write_flattened_sequence | read_batch_unformated after write_nested_sequence | read_col_formatted_as_numpy after write_array2d | read_col_formatted_as_numpy after write_flattened_sequence | read_col_formatted_as_numpy after write_nested_sequence | read_col_unformated after write_array2d | read_col_unformated after write_flattened_sequence | read_col_unformated after write_nested_sequence | read_formatted_as_numpy after write_array2d | read_formatted_as_numpy after write_flattened_sequence | read_formatted_as_numpy after write_nested_sequence | read_unformated after write_array2d | read_unformated after write_flattened_sequence | read_unformated after write_nested_sequence | write_array2d | write_flattened_sequence | write_nested_sequence |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.008956 / 0.011353 (-0.002397) | 0.004590 / 0.011008 (-0.006418) | 0.101305 / 0.038508 (0.062797) | 0.030347 / 0.023109 (0.007237) | 0.302492 / 0.275898 (0.026594) | 0.335986 / 0.323480 (0.012506) | 0.007272 / 0.007986 (-0.000714) | 0.004303 / 0.004328 (-0.000025) | 0.078592 / 0.004250 (0.074341) | 0.035545 / 0.037052 (-0.001507) | 0.316052 / 0.258489 (0.057563) | 0.342523 / 0.293841 (0.048682) | 0.034128 / 0.128546 (-0.094419) | 0.011475 / 0.075646 (-0.064171) | 0.325272 / 0.419271 (-0.093999) | 0.041815 / 0.043533 (-0.001717) | 0.303093 / 0.255139 (0.047955) | 0.331987 / 0.283200 (0.048788) | 0.087264 / 0.141683 (-0.054419) | 1.476284 / 1.452155 (0.024129) | 1.562034 / 1.492716 (0.069318) |\n\n### Benchmark: benchmark_getitem\\_100B.json\n\n| metric | get_batch_of\\_1024\\_random_rows | get_batch_of\\_1024\\_rows | get_first_row | get_last_row |\n|--------|---|---|---|---|\n| new / old (diff) | 0.206502 / 0.018006 (0.188496) | 0.409893 / 0.000490 (0.409404) | 0.002479 / 0.000200 (0.002279) | 0.000073 / 0.000054 (0.000018) |\n\n### Benchmark: benchmark_indices_mapping.json\n\n| metric | select | shard | shuffle | sort | train_test_split |\n|--------|---|---|---|---|---|\n| new / old (diff) | 0.022891 / 0.037411 (-0.014520) | 0.100209 / 0.014526 (0.085683) | 0.105576 / 0.176557 (-0.070981) | 0.141035 / 0.737135 (-0.596100) | 0.109733 / 0.296338 (-0.186606) |\n\n### Benchmark: benchmark_iterating.json\n\n| metric | read 5000 | read 50000 | read_batch 50000 10 | read_batch 50000 100 | read_batch 50000 1000 | read_formatted numpy 5000 | read_formatted pandas 5000 | read_formatted tensorflow 5000 | read_formatted torch 5000 | read_formatted_batch numpy 5000 10 | read_formatted_batch numpy 5000 1000 | shuffled read 5000 | shuffled read 50000 | shuffled read_batch 50000 10 | shuffled read_batch 50000 100 | shuffled read_batch 50000 1000 | shuffled read_formatted numpy 5000 | shuffled read_formatted_batch numpy 5000 10 | shuffled read_formatted_batch numpy 5000 1000 |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.413791 / 0.215209 (0.198582) | 4.125890 / 2.077655 (2.048235) | 1.833023 / 1.504120 (0.328903) | 1.631325 / 1.541195 (0.090130) | 1.708406 / 1.468490 (0.239916) | 0.690100 / 4.584777 (-3.894677) | 3.379058 / 3.745712 (-0.366654) | 2.019044 / 5.269862 (-3.250818) | 1.323332 / 4.565676 (-3.242344) | 0.082709 / 0.424275 (-0.341566) | 0.012434 / 0.007607 (0.004827) | 0.527139 / 0.226044 (0.301095) | 5.271529 / 2.268929 (3.002601) | 2.297311 / 55.444624 (-53.147314) | 1.949021 / 6.876477 (-4.927456) | 2.001098 / 2.142072 (-0.140975) | 0.811591 / 4.805227 (-3.993636) | 0.149028 / 6.500664 (-6.351637) | 0.066233 / 0.075469 (-0.009236) |\n\n### Benchmark: benchmark_map_filter.json\n\n| metric | filter | map fast-tokenizer batched | map identity | map identity batched | map no-op batched | map no-op batched numpy | map no-op batched pandas | map no-op batched pytorch | map no-op batched tensorflow |\n|--------|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 1.254276 / 1.841788 (-0.587512) | 13.638485 / 8.074308 (5.564177) | 13.943274 / 10.191392 (3.751882) | 0.147426 / 0.680424 (-0.532997) | 0.028602 / 0.534201 (-0.505599) | 0.398080 / 0.579283 (-0.181203) | 0.402178 / 0.434364 (-0.032186) | 0.477045 / 0.540337 (-0.063292) | 0.567731 / 1.386936 (-0.819205) |\n\n</details>\nPyArrow==latest\n\n<details>\n<summary>Show updated benchmarks!</summary>\n\n### Benchmark: benchmark_array_xd.json\n\n| metric | read_batch_formatted_as_numpy after write_array2d | read_batch_formatted_as_numpy after write_flattened_sequence | read_batch_formatted_as_numpy after write_nested_sequence | read_batch_unformated after write_array2d | read_batch_unformated after write_flattened_sequence | read_batch_unformated after write_nested_sequence | read_col_formatted_as_numpy after write_array2d | read_col_formatted_as_numpy after write_flattened_sequence | read_col_formatted_as_numpy after write_nested_sequence | read_col_unformated after write_array2d | read_col_unformated after write_flattened_sequence | read_col_unformated after write_nested_sequence | read_formatted_as_numpy after write_array2d | read_formatted_as_numpy after write_flattened_sequence | read_formatted_as_numpy after write_nested_sequence | read_unformated after write_array2d | read_unformated after write_flattened_sequence | read_unformated after write_nested_sequence | write_array2d | write_flattened_sequence | write_nested_sequence |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.006936 / 0.011353 (-0.004417) | 0.004614 / 0.011008 (-0.006394) | 0.079779 / 0.038508 (0.041271) | 0.027941 / 0.023109 (0.004832) | 0.347224 / 0.275898 (0.071326) | 0.378183 / 0.323480 (0.054703) | 0.005249 / 0.007986 (-0.002737) | 0.004907 / 0.004328 (0.000579) | 0.078678 / 0.004250 (0.074428) | 0.041912 / 0.037052 (0.004860) | 0.347838 / 0.258489 (0.089349) | 0.386760 / 0.293841 (0.092919) | 0.032680 / 0.128546 (-0.095867) | 0.014321 / 0.075646 (-0.061325) | 0.087924 / 0.419271 (-0.331347) | 0.045060 / 0.043533 (0.001527) | 0.340986 / 0.255139 (0.085847) | 0.368689 / 0.283200 (0.085489) | 0.093274 / 0.141683 (-0.048409) | 1.474435 / 1.452155 (0.022281) | 1.569753 / 1.492716 (0.077037) |\n\n### Benchmark: benchmark_getitem\\_100B.json\n\n| metric | get_batch_of\\_1024\\_random_rows | get_batch_of\\_1024\\_rows | get_first_row | get_last_row |\n|--------|---|---|---|---|\n| new / old (diff) | 0.206789 / 0.018006 (0.188783) | 0.416518 / 0.000490 (0.416028) | 0.000404 / 0.000200 (0.000204) | 0.000059 / 0.000054 (0.000005) |\n\n### Benchmark: benchmark_indices_mapping.json\n\n| metric | select | shard | shuffle | sort | train_test_split |\n|--------|---|---|---|---|---|\n| new / old (diff) | 0.026207 / 0.037411 (-0.011205) | 0.101914 / 0.014526 (0.087388) | 0.108585 / 0.176557 (-0.067972) | 0.150438 / 0.737135 (-0.586697) | 0.110744 / 0.296338 (-0.185594) |\n\n### Benchmark: benchmark_iterating.json\n\n| metric | read 5000 | read 50000 | read_batch 50000 10 | read_batch 50000 100 | read_batch 50000 1000 | read_formatted numpy 5000 | read_formatted pandas 5000 | read_formatted tensorflow 5000 | read_formatted torch 5000 | read_formatted_batch numpy 5000 10 | read_formatted_batch numpy 5000 1000 | shuffled read 5000 | shuffled read 50000 | shuffled read_batch 50000 10 | shuffled read_batch 50000 100 | shuffled read_batch 50000 1000 | shuffled read_formatted numpy 5000 | shuffled read_formatted_batch numpy 5000 10 | shuffled read_formatted_batch numpy 5000 1000 |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.443571 / 0.215209 (0.228362) | 4.433139 / 2.077655 (2.355485) | 2.109525 / 1.504120 (0.605405) | 1.901484 / 1.541195 (0.360290) | 1.968812 / 1.468490 (0.500322) | 0.704334 / 4.584777 (-3.880443) | 3.392028 / 3.745712 (-0.353684) | 3.072693 / 5.269862 (-2.197168) | 1.552227 / 4.565676 (-3.013449) | 0.083741 / 0.424275 (-0.340534) | 0.012627 / 0.007607 (0.005020) | 0.544706 / 0.226044 (0.318662) | 5.462743 / 2.268929 (3.193815) | 2.551265 / 55.444624 (-52.893360) | 2.208075 / 6.876477 (-4.668401) | 2.259092 / 2.142072 (0.117020) | 0.810687 / 4.805227 (-3.994540) | 0.152347 / 6.500664 (-6.348317) | 0.068346 / 0.075469 (-0.007123) |\n\n### Benchmark: benchmark_map_filter.json\n\n| metric | filter | map fast-tokenizer batched | map identity | map identity batched | map no-op batched | map no-op batched numpy | map no-op batched pandas | map no-op batched pytorch | map no-op batched tensorflow |\n|--------|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 1.269716 / 1.841788 (-0.572072) | 14.215698 / 8.074308 (6.141390) | 13.691773 / 10.191392 (3.500381) | 0.152620 / 0.680424 (-0.527804) | 0.017219 / 0.534201 (-0.516982) | 0.382533 / 0.579283 (-0.196750) | 0.388994 / 0.434364 (-0.045370) | 0.479400 / 0.540337 (-0.060938) | 0.572699 / 1.386936 (-0.814237) |\n\n</details>\n</details>\n\n![](https://cml.dev/watermark.png#f2d90f14cd6e756abeb27045940a6756104cc2d6 \"CML watermark\")\n" ]

[ { "color": "a2eeef", "default": true, "description": "New feature or request", "id": 1935892871, "name": "enhancement", "node_id": "MDU6TGFiZWwxOTM1ODkyODcx", "url": "https://api.github.com/repos/huggingface/datasets/labels/enhancement" }, { "color": "7057ff", "default": true, "description": "Good for newcomers", "id": 1935892877, "name": "good first issue", "node_id": "MDU6TGFiZWwxOTM1ODkyODc3", "url": "https://api.github.com/repos/huggingface/datasets/labels/good%20first%20issue" } ]

[ { "avatar_url": "https://avatars.githubusercontent.com/u/32437151?v=4", "events_url": "https://api.github.com/users/nateraw/events{/privacy}", "followers_url": "https://api.github.com/users/nateraw/followers", "following_url": "https://api.github.com/users/nateraw/following{/other_user}", "gists_url": "https://api.github.com/users/nateraw/gists{/gist_id}", "gravatar_id": "", "html_url": "https://github.com/nateraw", "id": 32437151, "login": "nateraw", "node_id": "MDQ6VXNlcjMyNDM3MTUx", "organizations_url": "https://api.github.com/users/nateraw/orgs", "received_events_url": "https://api.github.com/users/nateraw/received_events", "repos_url": "https://api.github.com/users/nateraw/repos", "site_admin": false, "starred_url": "https://api.github.com/users/nateraw/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/nateraw/subscriptions", "type": "User", "url": "https://api.github.com/users/nateraw" }, { "avatar_url": "https://avatars.githubusercontent.com/u/38854604?v=4", "events_url": "https://api.github.com/users/AJDERS/events{/privacy}", "followers_url": "https://api.github.com/users/AJDERS/followers", "following_url": "https://api.github.com/users/AJDERS/following{/other_user}", "gists_url": "https://api.github.com/users/AJDERS/gists{/gist_id}", "gravatar_id": "", "html_url": "https://github.com/AJDERS", "id": 38854604, "login": "AJDERS", "node_id": "MDQ6VXNlcjM4ODU0NjA0", "organizations_url": "https://api.github.com/users/AJDERS/orgs", "received_events_url": "https://api.github.com/users/AJDERS/received_events", "repos_url": "https://api.github.com/users/AJDERS/repos", "site_admin": false, "starred_url": "https://api.github.com/users/AJDERS/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/AJDERS/subscriptions", "type": "User", "url": "https://api.github.com/users/AJDERS" } ]

[ "Assigned to myself and will get to it in the next week, but if someone finds this issue annoying and wants to submit a PR before I do, just ping me here and I'll reassign :). ", "I would like to be assigned to this issue, @nateraw . #self-assign" ]

[]

[]

[ "_The documentation is not available anymore as the PR was closed or merged._", "<details>\n<summary>Show benchmarks</summary>\n\nPyArrow==6.0.0\n\n<details>\n<summary>Show updated benchmarks!</summary>\n\n### Benchmark: benchmark_array_xd.json\n\n| metric | read_batch_formatted_as_numpy after write_array2d | read_batch_formatted_as_numpy after write_flattened_sequence | read_batch_formatted_as_numpy after write_nested_sequence | read_batch_unformated after write_array2d | read_batch_unformated after write_flattened_sequence | read_batch_unformated after write_nested_sequence | read_col_formatted_as_numpy after write_array2d | read_col_formatted_as_numpy after write_flattened_sequence | read_col_formatted_as_numpy after write_nested_sequence | read_col_unformated after write_array2d | read_col_unformated after write_flattened_sequence | read_col_unformated after write_nested_sequence | read_formatted_as_numpy after write_array2d | read_formatted_as_numpy after write_flattened_sequence | read_formatted_as_numpy after write_nested_sequence | read_unformated after write_array2d | read_unformated after write_flattened_sequence | read_unformated after write_nested_sequence | write_array2d | write_flattened_sequence | write_nested_sequence |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.008726 / 0.011353 (-0.002627) | 0.004589 / 0.011008 (-0.006419) | 0.101078 / 0.038508 (0.062570) | 0.029732 / 0.023109 (0.006622) | 0.298309 / 0.275898 (0.022411) | 0.367800 / 0.323480 (0.044320) | 0.007025 / 0.007986 (-0.000961) | 0.003513 / 0.004328 (-0.000815) | 0.079531 / 0.004250 (0.075281) | 0.035588 / 0.037052 (-0.001465) | 0.307850 / 0.258489 (0.049361) | 0.351603 / 0.293841 (0.057762) | 0.033593 / 0.128546 (-0.094954) | 0.011669 / 0.075646 (-0.063977) | 0.323025 / 0.419271 (-0.096246) | 0.042047 / 0.043533 (-0.001486) | 0.300565 / 0.255139 (0.045426) | 0.329362 / 0.283200 (0.046163) | 0.089001 / 0.141683 (-0.052682) | 1.472799 / 1.452155 (0.020644) | 1.488902 / 1.492716 (-0.003814) |\n\n### Benchmark: benchmark_getitem\\_100B.json\n\n| metric | get_batch_of\\_1024\\_random_rows | get_batch_of\\_1024\\_rows | get_first_row | get_last_row |\n|--------|---|---|---|---|\n| new / old (diff) | 0.012491 / 0.018006 (-0.005515) | 0.408245 / 0.000490 (0.407755) | 0.003878 / 0.000200 (0.003678) | 0.000078 / 0.000054 (0.000023) |\n\n### Benchmark: benchmark_indices_mapping.json\n\n| metric | select | shard | shuffle | sort | train_test_split |\n|--------|---|---|---|---|---|\n| new / old (diff) | 0.023698 / 0.037411 (-0.013713) | 0.100442 / 0.014526 (0.085916) | 0.108233 / 0.176557 (-0.068323) | 0.145308 / 0.737135 (-0.591827) | 0.113121 / 0.296338 (-0.183218) |\n\n### Benchmark: benchmark_iterating.json\n\n| metric | read 5000 | read 50000 | read_batch 50000 10 | read_batch 50000 100 | read_batch 50000 1000 | read_formatted numpy 5000 | read_formatted pandas 5000 | read_formatted tensorflow 5000 | read_formatted torch 5000 | read_formatted_batch numpy 5000 10 | read_formatted_batch numpy 5000 1000 | shuffled read 5000 | shuffled read 50000 | shuffled read_batch 50000 10 | shuffled read_batch 50000 100 | shuffled read_batch 50000 1000 | shuffled read_formatted numpy 5000 | shuffled read_formatted_batch numpy 5000 10 | shuffled read_formatted_batch numpy 5000 1000 |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.420490 / 0.215209 (0.205281) | 4.179838 / 2.077655 (2.102183) | 2.156007 / 1.504120 (0.651887) | 1.911358 / 1.541195 (0.370163) | 1.867961 / 1.468490 (0.399471) | 0.685254 / 4.584777 (-3.899523) | 3.382386 / 3.745712 (-0.363326) | 3.285657 / 5.269862 (-1.984205) | 1.693878 / 4.565676 (-2.871798) | 0.081680 / 0.424275 (-0.342595) | 0.012182 / 0.007607 (0.004575) | 0.526021 / 0.226044 (0.299977) | 5.276217 / 2.268929 (3.007289) | 2.541518 / 55.444624 (-52.903106) | 2.313452 / 6.876477 (-4.563025) | 2.340000 / 2.142072 (0.197928) | 0.807099 / 4.805227 (-3.998128) | 0.147587 / 6.500664 (-6.353077) | 0.064280 / 0.075469 (-0.011189) |\n\n### Benchmark: benchmark_map_filter.json\n\n| metric | filter | map fast-tokenizer batched | map identity | map identity batched | map no-op batched | map no-op batched numpy | map no-op batched pandas | map no-op batched pytorch | map no-op batched tensorflow |\n|--------|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 1.223466 / 1.841788 (-0.618321) | 13.911365 / 8.074308 (5.837057) | 14.261550 / 10.191392 (4.070158) | 0.135922 / 0.680424 (-0.544502) | 0.028832 / 0.534201 (-0.505368) | 0.393142 / 0.579283 (-0.186141) | 0.400507 / 0.434364 (-0.033857) | 0.471792 / 0.540337 (-0.068546) | 0.558278 / 1.386936 (-0.828658) |\n\n</details>\nPyArrow==latest\n\n<details>\n<summary>Show updated benchmarks!</summary>\n\n### Benchmark: benchmark_array_xd.json\n\n| metric | read_batch_formatted_as_numpy after write_array2d | read_batch_formatted_as_numpy after write_flattened_sequence | read_batch_formatted_as_numpy after write_nested_sequence | read_batch_unformated after write_array2d | read_batch_unformated after write_flattened_sequence | read_batch_unformated after write_nested_sequence | read_col_formatted_as_numpy after write_array2d | read_col_formatted_as_numpy after write_flattened_sequence | read_col_formatted_as_numpy after write_nested_sequence | read_col_unformated after write_array2d | read_col_unformated after write_flattened_sequence | read_col_unformated after write_nested_sequence | read_formatted_as_numpy after write_array2d | read_formatted_as_numpy after write_flattened_sequence | read_formatted_as_numpy after write_nested_sequence | read_unformated after write_array2d | read_unformated after write_flattened_sequence | read_unformated after write_nested_sequence | write_array2d | write_flattened_sequence | write_nested_sequence |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.006644 / 0.011353 (-0.004709) | 0.004531 / 0.011008 (-0.006478) | 0.076285 / 0.038508 (0.037777) | 0.027249 / 0.023109 (0.004140) | 0.343137 / 0.275898 (0.067239) | 0.378498 / 0.323480 (0.055018) | 0.004950 / 0.007986 (-0.003036) | 0.003422 / 0.004328 (-0.000907) | 0.075662 / 0.004250 (0.071412) | 0.039692 / 0.037052 (0.002640) | 0.343402 / 0.258489 (0.084913) | 0.385067 / 0.293841 (0.091226) | 0.032382 / 0.128546 (-0.096164) | 0.011577 / 0.075646 (-0.064069) | 0.085534 / 0.419271 (-0.333738) | 0.052139 / 0.043533 (0.008606) | 0.342176 / 0.255139 (0.087037) | 0.367298 / 0.283200 (0.084098) | 0.096088 / 0.141683 (-0.045595) | 1.470770 / 1.452155 (0.018615) | 1.567316 / 1.492716 (0.074600) |\n\n### Benchmark: benchmark_getitem\\_100B.json\n\n| metric | get_batch_of\\_1024\\_random_rows | get_batch_of\\_1024\\_rows | get_first_row | get_last_row |\n|--------|---|---|---|---|\n| new / old (diff) | 0.217664 / 0.018006 (0.199657) | 0.397807 / 0.000490 (0.397317) | 0.006864 / 0.000200 (0.006664) | 0.000099 / 0.000054 (0.000044) |\n\n### Benchmark: benchmark_indices_mapping.json\n\n| metric | select | shard | shuffle | sort | train_test_split |\n|--------|---|---|---|---|---|\n| new / old (diff) | 0.025064 / 0.037411 (-0.012348) | 0.100906 / 0.014526 (0.086380) | 0.107444 / 0.176557 (-0.069113) | 0.143679 / 0.737135 (-0.593457) | 0.112460 / 0.296338 (-0.183879) |\n\n### Benchmark: benchmark_iterating.json\n\n| metric | read 5000 | read 50000 | read_batch 50000 10 | read_batch 50000 100 | read_batch 50000 1000 | read_formatted numpy 5000 | read_formatted pandas 5000 | read_formatted tensorflow 5000 | read_formatted torch 5000 | read_formatted_batch numpy 5000 10 | read_formatted_batch numpy 5000 1000 | shuffled read 5000 | shuffled read 50000 | shuffled read_batch 50000 10 | shuffled read_batch 50000 100 | shuffled read_batch 50000 1000 | shuffled read_formatted numpy 5000 | shuffled read_formatted_batch numpy 5000 10 | shuffled read_formatted_batch numpy 5000 1000 |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.442634 / 0.215209 (0.227425) | 4.410687 / 2.077655 (2.333032) | 2.067445 / 1.504120 (0.563325) | 1.860569 / 1.541195 (0.319374) | 1.943523 / 1.468490 (0.475033) | 0.694585 / 4.584777 (-3.890192) | 3.375906 / 3.745712 (-0.369806) | 3.483334 / 5.269862 (-1.786528) | 1.437700 / 4.565676 (-3.127977) | 0.083138 / 0.424275 (-0.341137) | 0.012979 / 0.007607 (0.005372) | 0.536414 / 0.226044 (0.310370) | 5.379872 / 2.268929 (3.110943) | 2.517907 / 55.444624 (-52.926717) | 2.164772 / 6.876477 (-4.711705) | 2.212839 / 2.142072 (0.070767) | 0.799675 / 4.805227 (-4.005553) | 0.150253 / 6.500664 (-6.350411) | 0.067033 / 0.075469 (-0.008436) |\n\n### Benchmark: benchmark_map_filter.json\n\n| metric | filter | map fast-tokenizer batched | map identity | map identity batched | map no-op batched | map no-op batched numpy | map no-op batched pandas | map no-op batched pytorch | map no-op batched tensorflow |\n|--------|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 1.295592 / 1.841788 (-0.546196) | 14.372932 / 8.074308 (6.298623) | 13.618423 / 10.191392 (3.427031) | 0.141212 / 0.680424 (-0.539212) | 0.016933 / 0.534201 (-0.517268) | 0.385664 / 0.579283 (-0.193619) | 0.386919 / 0.434364 (-0.047445) | 0.477022 / 0.540337 (-0.063315) | 0.565158 / 1.386936 (-0.821778) |\n\n</details>\n</details>\n\n![](https://cml.dev/watermark.png#38c715cc787a81d0fd894205b4b24aca2f45f84b \"CML watermark\")\n" ]

[]

[]

[ "_The documentation is not available anymore as the PR was closed or merged._", "<details>\n<summary>Show benchmarks</summary>\n\nPyArrow==6.0.0\n\n<details>\n<summary>Show updated benchmarks!</summary>\n\n### Benchmark: benchmark_array_xd.json\n\n| metric | read_batch_formatted_as_numpy after write_array2d | read_batch_formatted_as_numpy after write_flattened_sequence | read_batch_formatted_as_numpy after write_nested_sequence | read_batch_unformated after write_array2d | read_batch_unformated after write_flattened_sequence | read_batch_unformated after write_nested_sequence | read_col_formatted_as_numpy after write_array2d | read_col_formatted_as_numpy after write_flattened_sequence | read_col_formatted_as_numpy after write_nested_sequence | read_col_unformated after write_array2d | read_col_unformated after write_flattened_sequence | read_col_unformated after write_nested_sequence | read_formatted_as_numpy after write_array2d | read_formatted_as_numpy after write_flattened_sequence | read_formatted_as_numpy after write_nested_sequence | read_unformated after write_array2d | read_unformated after write_flattened_sequence | read_unformated after write_nested_sequence | write_array2d | write_flattened_sequence | write_nested_sequence |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.008581 / 0.011353 (-0.002772) | 0.004519 / 0.011008 (-0.006490) | 0.099721 / 0.038508 (0.061213) | 0.029217 / 0.023109 (0.006107) | 0.298229 / 0.275898 (0.022331) | 0.332605 / 0.323480 (0.009125) | 0.006880 / 0.007986 (-0.001106) | 0.003324 / 0.004328 (-0.001005) | 0.078143 / 0.004250 (0.073892) | 0.034262 / 0.037052 (-0.002790) | 0.304162 / 0.258489 (0.045673) | 0.342351 / 0.293841 (0.048510) | 0.033387 / 0.128546 (-0.095159) | 0.011397 / 0.075646 (-0.064249) | 0.321527 / 0.419271 (-0.097744) | 0.040886 / 0.043533 (-0.002647) | 0.299968 / 0.255139 (0.044829) | 0.322484 / 0.283200 (0.039285) | 0.083832 / 0.141683 (-0.057851) | 1.482241 / 1.452155 (0.030086) | 1.548438 / 1.492716 (0.055721) |\n\n### Benchmark: benchmark_getitem\\_100B.json\n\n| metric | get_batch_of\\_1024\\_random_rows | get_batch_of\\_1024\\_rows | get_first_row | get_last_row |\n|--------|---|---|---|---|\n| new / old (diff) | 0.191002 / 0.018006 (0.172996) | 0.403423 / 0.000490 (0.402933) | 0.002493 / 0.000200 (0.002293) | 0.000074 / 0.000054 (0.000020) |\n\n### Benchmark: benchmark_indices_mapping.json\n\n| metric | select | shard | shuffle | sort | train_test_split |\n|--------|---|---|---|---|---|\n| new / old (diff) | 0.023720 / 0.037411 (-0.013691) | 0.100806 / 0.014526 (0.086281) | 0.105314 / 0.176557 (-0.071242) | 0.141490 / 0.737135 (-0.595645) | 0.108695 / 0.296338 (-0.187644) |\n\n### Benchmark: benchmark_iterating.json\n\n| metric | read 5000 | read 50000 | read_batch 50000 10 | read_batch 50000 100 | read_batch 50000 1000 | read_formatted numpy 5000 | read_formatted pandas 5000 | read_formatted tensorflow 5000 | read_formatted torch 5000 | read_formatted_batch numpy 5000 10 | read_formatted_batch numpy 5000 1000 | shuffled read 5000 | shuffled read 50000 | shuffled read_batch 50000 10 | shuffled read_batch 50000 100 | shuffled read_batch 50000 1000 | shuffled read_formatted numpy 5000 | shuffled read_formatted_batch numpy 5000 10 | shuffled read_formatted_batch numpy 5000 1000 |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.412250 / 0.215209 (0.197041) | 4.124830 / 2.077655 (2.047175) | 1.851948 / 1.504120 (0.347828) | 1.651597 / 1.541195 (0.110403) | 1.712486 / 1.468490 (0.243996) | 0.696634 / 4.584777 (-3.888143) | 3.304220 / 3.745712 (-0.441492) | 1.862776 / 5.269862 (-3.407086) | 1.159452 / 4.565676 (-3.406224) | 0.082930 / 0.424275 (-0.341345) | 0.012586 / 0.007607 (0.004979) | 0.524499 / 0.226044 (0.298455) | 5.249235 / 2.268929 (2.980307) | 2.293187 / 55.444624 (-53.151437) | 1.950101 / 6.876477 (-4.926376) | 2.008274 / 2.142072 (-0.133799) | 0.811641 / 4.805227 (-3.993586) | 0.148785 / 6.500664 (-6.351879) | 0.064461 / 0.075469 (-0.011008) |\n\n### Benchmark: benchmark_map_filter.json\n\n| metric | filter | map fast-tokenizer batched | map identity | map identity batched | map no-op batched | map no-op batched numpy | map no-op batched pandas | map no-op batched pytorch | map no-op batched tensorflow |\n|--------|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 1.232227 / 1.841788 (-0.609561) | 13.235896 / 8.074308 (5.161588) | 13.837420 / 10.191392 (3.646028) | 0.135586 / 0.680424 (-0.544838) | 0.028935 / 0.534201 (-0.505266) | 0.397064 / 0.579283 (-0.182220) | 0.393814 / 0.434364 (-0.040549) | 0.480450 / 0.540337 (-0.059887) | 0.561159 / 1.386936 (-0.825777) |\n\n</details>\nPyArrow==latest\n\n<details>\n<summary>Show updated benchmarks!</summary>\n\n### Benchmark: benchmark_array_xd.json\n\n| metric | read_batch_formatted_as_numpy after write_array2d | read_batch_formatted_as_numpy after write_flattened_sequence | read_batch_formatted_as_numpy after write_nested_sequence | read_batch_unformated after write_array2d | read_batch_unformated after write_flattened_sequence | read_batch_unformated after write_nested_sequence | read_col_formatted_as_numpy after write_array2d | read_col_formatted_as_numpy after write_flattened_sequence | read_col_formatted_as_numpy after write_nested_sequence | read_col_unformated after write_array2d | read_col_unformated after write_flattened_sequence | read_col_unformated after write_nested_sequence | read_formatted_as_numpy after write_array2d | read_formatted_as_numpy after write_flattened_sequence | read_formatted_as_numpy after write_nested_sequence | read_unformated after write_array2d | read_unformated after write_flattened_sequence | read_unformated after write_nested_sequence | write_array2d | write_flattened_sequence | write_nested_sequence |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.006696 / 0.011353 (-0.004657) | 0.004528 / 0.011008 (-0.006480) | 0.077335 / 0.038508 (0.038827) | 0.027181 / 0.023109 (0.004072) | 0.345379 / 0.275898 (0.069481) | 0.372544 / 0.323480 (0.049064) | 0.006808 / 0.007986 (-0.001178) | 0.003284 / 0.004328 (-0.001045) | 0.077379 / 0.004250 (0.073129) | 0.039954 / 0.037052 (0.002901) | 0.348094 / 0.258489 (0.089605) | 0.382315 / 0.293841 (0.088474) | 0.031694 / 0.128546 (-0.096852) | 0.011714 / 0.075646 (-0.063933) | 0.086425 / 0.419271 (-0.332846) | 0.041778 / 0.043533 (-0.001754) | 0.342161 / 0.255139 (0.087022) | 0.363798 / 0.283200 (0.080599) | 0.091315 / 0.141683 (-0.050368) | 1.462066 / 1.452155 (0.009912) | 1.541417 / 1.492716 (0.048700) |\n\n### Benchmark: benchmark_getitem\\_100B.json\n\n| metric | get_batch_of\\_1024\\_random_rows | get_batch_of\\_1024\\_rows | get_first_row | get_last_row |\n|--------|---|---|---|---|\n| new / old (diff) | 0.235840 / 0.018006 (0.217834) | 0.397096 / 0.000490 (0.396606) | 0.004597 / 0.000200 (0.004397) | 0.000079 / 0.000054 (0.000025) |\n\n### Benchmark: benchmark_indices_mapping.json\n\n| metric | select | shard | shuffle | sort | train_test_split |\n|--------|---|---|---|---|---|\n| new / old (diff) | 0.024296 / 0.037411 (-0.013115) | 0.099167 / 0.014526 (0.084641) | 0.108257 / 0.176557 (-0.068299) | 0.143434 / 0.737135 (-0.593701) | 0.111933 / 0.296338 (-0.184406) |\n\n### Benchmark: benchmark_iterating.json\n\n| metric | read 5000 | read 50000 | read_batch 50000 10 | read_batch 50000 100 | read_batch 50000 1000 | read_formatted numpy 5000 | read_formatted pandas 5000 | read_formatted tensorflow 5000 | read_formatted torch 5000 | read_formatted_batch numpy 5000 10 | read_formatted_batch numpy 5000 1000 | shuffled read 5000 | shuffled read 50000 | shuffled read_batch 50000 10 | shuffled read_batch 50000 100 | shuffled read_batch 50000 1000 | shuffled read_formatted numpy 5000 | shuffled read_formatted_batch numpy 5000 10 | shuffled read_formatted_batch numpy 5000 1000 |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.440306 / 0.215209 (0.225096) | 4.374065 / 2.077655 (2.296410) | 2.072653 / 1.504120 (0.568533) | 1.864829 / 1.541195 (0.323635) | 1.927970 / 1.468490 (0.459479) | 0.710118 / 4.584777 (-3.874659) | 3.391216 / 3.745712 (-0.354496) | 1.888847 / 5.269862 (-3.381015) | 1.178740 / 4.565676 (-3.386936) | 0.083950 / 0.424275 (-0.340325) | 0.012567 / 0.007607 (0.004960) | 0.540557 / 0.226044 (0.314513) | 5.437621 / 2.268929 (3.168692) | 2.531165 / 55.444624 (-52.913460) | 2.181450 / 6.876477 (-4.695027) | 2.209108 / 2.142072 (0.067035) | 0.814236 / 4.805227 (-3.990991) | 0.153000 / 6.500664 (-6.347664) | 0.066769 / 0.075469 (-0.008700) |\n\n### Benchmark: benchmark_map_filter.json\n\n| metric | filter | map fast-tokenizer batched | map identity | map identity batched | map no-op batched | map no-op batched numpy | map no-op batched pandas | map no-op batched pytorch | map no-op batched tensorflow |\n|--------|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 1.301057 / 1.841788 (-0.540731) | 14.066786 / 8.074308 (5.992478) | 13.641455 / 10.191392 (3.450063) | 0.138838 / 0.680424 (-0.541586) | 0.016733 / 0.534201 (-0.517468) | 0.391823 / 0.579283 (-0.187460) | 0.390817 / 0.434364 (-0.043547) | 0.487682 / 0.540337 (-0.052656) | 0.581134 / 1.386936 (-0.805802) |\n\n</details>\n</details>\n\n![](https://cml.dev/watermark.png#b065547654efa0ec633cf373ac1512884c68b2e1 \"CML watermark\")\n" ]

[]

[]

[ "_The documentation is not available anymore as the PR was closed or merged._", "<details>\n<summary>Show benchmarks</summary>\n\nPyArrow==6.0.0\n\n<details>\n<summary>Show updated benchmarks!</summary>\n\n### Benchmark: benchmark_array_xd.json\n\n| metric | read_batch_formatted_as_numpy after write_array2d | read_batch_formatted_as_numpy after write_flattened_sequence | read_batch_formatted_as_numpy after write_nested_sequence | read_batch_unformated after write_array2d | read_batch_unformated after write_flattened_sequence | read_batch_unformated after write_nested_sequence | read_col_formatted_as_numpy after write_array2d | read_col_formatted_as_numpy after write_flattened_sequence | read_col_formatted_as_numpy after write_nested_sequence | read_col_unformated after write_array2d | read_col_unformated after write_flattened_sequence | read_col_unformated after write_nested_sequence | read_formatted_as_numpy after write_array2d | read_formatted_as_numpy after write_flattened_sequence | read_formatted_as_numpy after write_nested_sequence | read_unformated after write_array2d | read_unformated after write_flattened_sequence | read_unformated after write_nested_sequence | write_array2d | write_flattened_sequence | write_nested_sequence |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.008798 / 0.011353 (-0.002554) | 0.005313 / 0.011008 (-0.005695) | 0.099234 / 0.038508 (0.060726) | 0.033935 / 0.023109 (0.010826) | 0.306610 / 0.275898 (0.030712) | 0.373151 / 0.323480 (0.049671) | 0.008305 / 0.007986 (0.000320) | 0.004647 / 0.004328 (0.000319) | 0.079984 / 0.004250 (0.075733) | 0.042546 / 0.037052 (0.005493) | 0.355105 / 0.258489 (0.096616) | 0.332769 / 0.293841 (0.038928) | 0.037708 / 0.128546 (-0.090839) | 0.012141 / 0.075646 (-0.063505) | 0.365338 / 0.419271 (-0.053933) | 0.048875 / 0.043533 (0.005343) | 0.301771 / 0.255139 (0.046632) | 0.323301 / 0.283200 (0.040101) | 0.099116 / 0.141683 (-0.042566) | 1.463948 / 1.452155 (0.011793) | 1.563006 / 1.492716 (0.070290) |\n\n### Benchmark: benchmark_getitem\\_100B.json\n\n| metric | get_batch_of\\_1024\\_random_rows | get_batch_of\\_1024\\_rows | get_first_row | get_last_row |\n|--------|---|---|---|---|\n| new / old (diff) | 0.219799 / 0.018006 (0.201793) | 0.524126 / 0.000490 (0.523636) | 0.003899 / 0.000200 (0.003699) | 0.000092 / 0.000054 (0.000037) |\n\n### Benchmark: benchmark_indices_mapping.json\n\n| metric | select | shard | shuffle | sort | train_test_split |\n|--------|---|---|---|---|---|\n| new / old (diff) | 0.028361 / 0.037411 (-0.009050) | 0.111386 / 0.014526 (0.096860) | 0.125749 / 0.176557 (-0.050807) | 0.167026 / 0.737135 (-0.570109) | 0.132082 / 0.296338 (-0.164257) |\n\n### Benchmark: benchmark_iterating.json\n\n| metric | read 5000 | read 50000 | read_batch 50000 10 | read_batch 50000 100 | read_batch 50000 1000 | read_formatted numpy 5000 | read_formatted pandas 5000 | read_formatted tensorflow 5000 | read_formatted torch 5000 | read_formatted_batch numpy 5000 10 | read_formatted_batch numpy 5000 1000 | shuffled read 5000 | shuffled read 50000 | shuffled read_batch 50000 10 | shuffled read_batch 50000 100 | shuffled read_batch 50000 1000 | shuffled read_formatted numpy 5000 | shuffled read_formatted_batch numpy 5000 10 | shuffled read_formatted_batch numpy 5000 1000 |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.385046 / 0.215209 (0.169837) | 3.933129 / 2.077655 (1.855475) | 1.823395 / 1.504120 (0.319276) | 1.646468 / 1.541195 (0.105273) | 1.658835 / 1.468490 (0.190344) | 0.708300 / 4.584777 (-3.876477) | 4.001478 / 3.745712 (0.255766) | 2.221773 / 5.269862 (-3.048089) | 1.597925 / 4.565676 (-2.967751) | 0.088699 / 0.424275 (-0.335577) | 0.013575 / 0.007607 (0.005968) | 0.520577 / 0.226044 (0.294533) | 5.044313 / 2.268929 (2.775385) | 2.239862 / 55.444624 (-53.204763) | 2.060394 / 6.876477 (-4.816083) | 2.060684 / 2.142072 (-0.081389) | 0.844862 / 4.805227 (-3.960365) | 0.190321 / 6.500664 (-6.310343) | 0.071595 / 0.075469 (-0.003875) |\n\n### Benchmark: benchmark_map_filter.json\n\n| metric | filter | map fast-tokenizer batched | map identity | map identity batched | map no-op batched | map no-op batched numpy | map no-op batched pandas | map no-op batched pytorch | map no-op batched tensorflow |\n|--------|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 1.400048 / 1.841788 (-0.441740) | 15.684159 / 8.074308 (7.609851) | 14.369298 / 10.191392 (4.177906) | 0.164874 / 0.680424 (-0.515550) | 0.033219 / 0.534201 (-0.500982) | 0.449176 / 0.579283 (-0.130107) | 0.456560 / 0.434364 (0.022196) | 0.517978 / 0.540337 (-0.022359) | 0.635467 / 1.386936 (-0.751469) |\n\n</details>\nPyArrow==latest\n\n<details>\n<summary>Show updated benchmarks!</summary>\n\n### Benchmark: benchmark_array_xd.json\n\n| metric | read_batch_formatted_as_numpy after write_array2d | read_batch_formatted_as_numpy after write_flattened_sequence | read_batch_formatted_as_numpy after write_nested_sequence | read_batch_unformated after write_array2d | read_batch_unformated after write_flattened_sequence | read_batch_unformated after write_nested_sequence | read_col_formatted_as_numpy after write_array2d | read_col_formatted_as_numpy after write_flattened_sequence | read_col_formatted_as_numpy after write_nested_sequence | read_col_unformated after write_array2d | read_col_unformated after write_flattened_sequence | read_col_unformated after write_nested_sequence | read_formatted_as_numpy after write_array2d | read_formatted_as_numpy after write_flattened_sequence | read_formatted_as_numpy after write_nested_sequence | read_unformated after write_array2d | read_unformated after write_flattened_sequence | read_unformated after write_nested_sequence | write_array2d | write_flattened_sequence | write_nested_sequence |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.007263 / 0.011353 (-0.004089) | 0.005451 / 0.011008 (-0.005558) | 0.078785 / 0.038508 (0.040277) | 0.032656 / 0.023109 (0.009546) | 0.346384 / 0.275898 (0.070486) | 0.390778 / 0.323480 (0.067299) | 0.005848 / 0.007986 (-0.002137) | 0.004565 / 0.004328 (0.000236) | 0.077903 / 0.004250 (0.073652) | 0.048659 / 0.037052 (0.011606) | 0.368629 / 0.258489 (0.110140) | 0.401632 / 0.293841 (0.107791) | 0.038516 / 0.128546 (-0.090030) | 0.011895 / 0.075646 (-0.063752) | 0.089185 / 0.419271 (-0.330086) | 0.049875 / 0.043533 (0.006342) | 0.344771 / 0.255139 (0.089632) | 0.378237 / 0.283200 (0.095038) | 0.099184 / 0.141683 (-0.042498) | 1.505058 / 1.452155 (0.052903) | 1.555330 / 1.492716 (0.062614) |\n\n### Benchmark: benchmark_getitem\\_100B.json\n\n| metric | get_batch_of\\_1024\\_random_rows | get_batch_of\\_1024\\_rows | get_first_row | get_last_row |\n|--------|---|---|---|---|\n| new / old (diff) | 0.209132 / 0.018006 (0.191126) | 0.479928 / 0.000490 (0.479438) | 0.005923 / 0.000200 (0.005723) | 0.000113 / 0.000054 (0.000058) |\n\n### Benchmark: benchmark_indices_mapping.json\n\n| metric | select | shard | shuffle | sort | train_test_split |\n|--------|---|---|---|---|---|\n| new / old (diff) | 0.029187 / 0.037411 (-0.008224) | 0.117026 / 0.014526 (0.102500) | 0.131834 / 0.176557 (-0.044722) | 0.172797 / 0.737135 (-0.564339) | 0.129098 / 0.296338 (-0.167240) |\n\n### Benchmark: benchmark_iterating.json\n\n| metric | read 5000 | read 50000 | read_batch 50000 10 | read_batch 50000 100 | read_batch 50000 1000 | read_formatted numpy 5000 | read_formatted pandas 5000 | read_formatted tensorflow 5000 | read_formatted torch 5000 | read_formatted_batch numpy 5000 10 | read_formatted_batch numpy 5000 1000 | shuffled read 5000 | shuffled read 50000 | shuffled read_batch 50000 10 | shuffled read_batch 50000 100 | shuffled read_batch 50000 1000 | shuffled read_formatted numpy 5000 | shuffled read_formatted_batch numpy 5000 10 | shuffled read_formatted_batch numpy 5000 1000 |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.450214 / 0.215209 (0.235005) | 4.323950 / 2.077655 (2.246295) | 2.210100 / 1.504120 (0.705980) | 2.058733 / 1.541195 (0.517538) | 1.968191 / 1.468490 (0.499701) | 0.694918 / 4.584777 (-3.889859) | 4.176559 / 3.745712 (0.430846) | 2.118211 / 5.269862 (-3.151651) | 1.410652 / 4.565676 (-3.155024) | 0.093606 / 0.424275 (-0.330669) | 0.013729 / 0.007607 (0.006122) | 0.528463 / 0.226044 (0.302418) | 5.311766 / 2.268929 (3.042837) | 2.522981 / 55.444624 (-52.921644) | 2.177191 / 6.876477 (-4.699285) | 2.211448 / 2.142072 (0.069375) | 0.824334 / 4.805227 (-3.980893) | 0.166642 / 6.500664 (-6.334022) | 0.062774 / 0.075469 (-0.012695) |\n\n### Benchmark: benchmark_map_filter.json\n\n| metric | filter | map fast-tokenizer batched | map identity | map identity batched | map no-op batched | map no-op batched numpy | map no-op batched pandas | map no-op batched pytorch | map no-op batched tensorflow |\n|--------|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 1.367573 / 1.841788 (-0.474215) | 15.913637 / 8.074308 (7.839328) | 13.397411 / 10.191392 (3.206019) | 0.162599 / 0.680424 (-0.517825) | 0.020325 / 0.534201 (-0.513876) | 0.438745 / 0.579283 (-0.140538) | 0.449892 / 0.434364 (0.015528) | 0.556226 / 0.540337 (0.015888) | 0.672661 / 1.386936 (-0.714275) |\n\n</details>\n</details>\n\n![](https://cml.dev/watermark.png#5f810b7011a8a4ab077a1847c024d2d9e267b065 \"CML watermark\")\n" ]

[]

[]

[ "Hi @kradonneoh, thanks for reporting.\r\n\r\nPlease note that to work with audio datasets (and specifically with MP3 files) we have detailed installation instructions in our docs: https://huggingface.co/docs/datasets/installation#audio\r\n- one of the requirements is torchaudio<0.12.0\r\n\r\nLet us know if the problem persists after having followed them.", "I saw that and have followed it (hence the Expected Behavior section of the bug report). \r\n\r\nIs there no intention of updating to the latest version? It does limit the version of `torch` I can use, which isn’t ideal.", "@kradonneoh hey! actually with `ffmpeg4` loading of mp3 files should work, so this is a not expected behavior and we need to investigate it. It works on my side with `torchaudio==0.13` and `ffmpeg==4.2.7`. Which `torchaudio` version do you use?\r\n\r\n`datasets` should support decoding of mp3 files with `torchaudio` when its version is `>0.12` but as you noted it requires `ffmpeg>4`, we need to fix this in the documentation, thank you for pointing to this! \r\n\r\nBut according to your traceback it seems that it tries to use [`libsndfile`](https://github.com/libsndfile/libsndfile) backend for mp3 decoding. And `libsndfile` library supports mp3 decoding starting from version 1.1.0 which on Linux has to be compiled from source for now afaik. \r\n\r\nfyi - we are aiming at getting rid of `torchaudio` dependency at all by the next major library release in favor of `libsndfile` too.", "We now decode MP3 with `soundfile`, so I'm closing this issue" ]

[]

[]

[ "Hi! The correct path is still `dill._dill.XXXX` in the latest release. What do you get when you run `python -c \"import dill; print(dill.__version__)\"` in your environment?", "`0.3.6` I feel like that's bad news, because it's probably not the issue.\r\n\r\nMy mistake, about the wrong path guess. I think I didn't notice that the first `dill` in the path isn't supposed to be included in the path specification in python.\r\n<img width=\"146\" alt=\"Screen Shot 2023-01-31 at 12 58 32 PM\" src=\"https://user-images.githubusercontent.com/35349273/215844209-74af6a8f-9bff-4c75-9495-44c658c8e9f7.png\">\r\n", "Hi, @avivbrokman, this issue you report appeared only with old versions of dill. See:\r\n- #288\r\n\r\nAre you sure you are in the right Python environment?\r\n- Please note that Jupyter (where I guess you get the error) may have multiple execution backends (IPython kernels) that might be different from the Python environment your are using to get the dill version\r\n - Have you run `import dill; print(dill.__version__)` in the same Jupyter/IPython that you were using when you got the error while executing `import datasets`?", "I'm using spyder, and I am still getting `0.3.6` for `dill`, so unfortunately #288 isn't applicable, I think. However, I found something odd that I believe is a clue: \r\n\r\n```\r\nimport inspect\r\nimport dill\r\n\r\ninspect.getfile(dill)\r\n>>> '/Users/avivbrokman/opt/anaconda3/lib/python3.9/site-packages/dill/__init__.py'\r\n```\r\n\r\nI checked out the directory, and there is no `dill` subdirectory within '/Users/avivbrokman/opt/anaconda3/lib/python3.9/site-packages/dill`, as there should be. Rather, `_dill.py` is in '/Users/avivbrokman/opt/anaconda3/lib/python3.9/site-packages/dill` itself. \r\n\r\n If I run `pip install dill` or `pip install --upgrade dill`, I get the message `Requirement already satisfied: dill in ./opt/anaconda3/lib/python3.9/site-packages (0.3.6)`. If I run `conda upgrade dill`, I get the message `Solving environment: failed with repodata from current_repodata.json, will retry with next repodata source.` a couple of times, followed by\r\n\r\n```\r\nSolving environment: failed\r\nSolving environment: / \r\nFound conflicts! Looking for incompatible packages.\r\n```\r\n\r\nAnd then terminal proceeds to list conflicts between different packages I have.\r\n\r\nThis is all very strange to me because I recently uninstalled and reinstalled `anaconda`.\r\n", "As I said above, I guess this is not a problem with `datasets`. I think you have different Python environments: one with the new dill version (the one you get while using pip) and other with the old dill version (the one where you get the AttributeError).\r\n\r\nYou should update `dill` in the Python environment you are using within spyder.\r\n\r\nPlease note that the `_dill` module is present in the `dill` package since their 2.8.0 version." ]

[]

[]

[ "Hi @omar-araboghli, thanks for your proposal.\r\n\r\nHave you tried to use \"csv\" loader instead of \"text\"? That already has a `sep` argument.", "Hi @albertvillanova, thanks for the quick response!\r\n\r\nIndeed, I have been trying to use `csv` instead of `text`. However I am still not able to define range of rows as one sequence, that is achievable with passing `sample_by='paragraph'` to the `TextConfig`\r\n\r\nFor instance, the below code\r\n\r\n```python\r\nimport datasets\r\n\r\ndataset = datasets.load_dataset(\r\n path='csv',\r\n data_files={'train': TRAINING_SET_PATH},\r\n sep='\\t',\r\n header=None,\r\n column_names=['tokens', 'pos_tags', 'chunk_tags', 'ner_tags']\r\n)\r\n```\r\n\r\nleads to \r\n\r\n```python\r\ndataset\r\n>>> DatasetDict({\r\n train: Dataset({\r\n features: ['tokens', 'pos_tags', 'chunk_tags', 'ner_tags'],\r\n num_rows: 62543\r\n })\r\n})\r\n\r\ndataset['train'][0]\r\n>>> {'tokens': 'Distribution',\r\n 'pos_tags': 'NN',\r\n 'chunk_tags': 'O',\r\n 'ner_tags': 'O'\r\n}\r\n```\r\nIs there a way to deal with multiple csv rows as one dataset instance, where each column is a sequence of those rows ?" ]

[]

[]

[ "_The documentation is not available anymore as the PR was closed or merged._", "<details>\n<summary>Show benchmarks</summary>\n\nPyArrow==6.0.0\n\n<details>\n<summary>Show updated benchmarks!</summary>\n\n### Benchmark: benchmark_array_xd.json\n\n| metric | read_batch_formatted_as_numpy after write_array2d | read_batch_formatted_as_numpy after write_flattened_sequence | read_batch_formatted_as_numpy after write_nested_sequence | read_batch_unformated after write_array2d | read_batch_unformated after write_flattened_sequence | read_batch_unformated after write_nested_sequence | read_col_formatted_as_numpy after write_array2d | read_col_formatted_as_numpy after write_flattened_sequence | read_col_formatted_as_numpy after write_nested_sequence | read_col_unformated after write_array2d | read_col_unformated after write_flattened_sequence | read_col_unformated after write_nested_sequence | read_formatted_as_numpy after write_array2d | read_formatted_as_numpy after write_flattened_sequence | read_formatted_as_numpy after write_nested_sequence | read_unformated after write_array2d | read_unformated after write_flattened_sequence | read_unformated after write_nested_sequence | write_array2d | write_flattened_sequence | write_nested_sequence |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.008492 / 0.011353 (-0.002861) | 0.004717 / 0.011008 (-0.006292) | 0.101111 / 0.038508 (0.062602) | 0.029129 / 0.023109 (0.006019) | 0.307564 / 0.275898 (0.031666) | 0.367038 / 0.323480 (0.043558) | 0.007105 / 0.007986 (-0.000881) | 0.003622 / 0.004328 (-0.000706) | 0.078370 / 0.004250 (0.074120) | 0.036960 / 0.037052 (-0.000093) | 0.315612 / 0.258489 (0.057123) | 0.353601 / 0.293841 (0.059760) | 0.032900 / 0.128546 (-0.095647) | 0.011405 / 0.075646 (-0.064241) | 0.322331 / 0.419271 (-0.096940) | 0.040823 / 0.043533 (-0.002710) | 0.306734 / 0.255139 (0.051595) | 0.328155 / 0.283200 (0.044955) | 0.087169 / 0.141683 (-0.054514) | 1.460543 / 1.452155 (0.008389) | 1.498094 / 1.492716 (0.005378) |\n\n### Benchmark: benchmark_getitem\\_100B.json\n\n| metric | get_batch_of\\_1024\\_random_rows | get_batch_of\\_1024\\_rows | get_first_row | get_last_row |\n|--------|---|---|---|---|\n| new / old (diff) | 0.011863 / 0.018006 (-0.006143) | 0.416315 / 0.000490 (0.415826) | 0.003463 / 0.000200 (0.003263) | 0.000075 / 0.000054 (0.000020) |\n\n### Benchmark: benchmark_indices_mapping.json\n\n| metric | select | shard | shuffle | sort | train_test_split |\n|--------|---|---|---|---|---|\n| new / old (diff) | 0.023219 / 0.037411 (-0.014192) | 0.096469 / 0.014526 (0.081943) | 0.105960 / 0.176557 (-0.070596) | 0.148993 / 0.737135 (-0.588142) | 0.108112 / 0.296338 (-0.188226) |\n\n### Benchmark: benchmark_iterating.json\n\n| metric | read 5000 | read 50000 | read_batch 50000 10 | read_batch 50000 100 | read_batch 50000 1000 | read_formatted numpy 5000 | read_formatted pandas 5000 | read_formatted tensorflow 5000 | read_formatted torch 5000 | read_formatted_batch numpy 5000 10 | read_formatted_batch numpy 5000 1000 | shuffled read 5000 | shuffled read 50000 | shuffled read_batch 50000 10 | shuffled read_batch 50000 100 | shuffled read_batch 50000 1000 | shuffled read_formatted numpy 5000 | shuffled read_formatted_batch numpy 5000 10 | shuffled read_formatted_batch numpy 5000 1000 |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.415662 / 0.215209 (0.200453) | 4.155111 / 2.077655 (2.077456) | 1.834943 / 1.504120 (0.330823) | 1.622752 / 1.541195 (0.081557) | 1.701630 / 1.468490 (0.233140) | 0.690596 / 4.584777 (-3.894181) | 3.399385 / 3.745712 (-0.346327) | 3.140521 / 5.269862 (-2.129341) | 1.609152 / 4.565676 (-2.956524) | 0.082132 / 0.424275 (-0.342143) | 0.012343 / 0.007607 (0.004735) | 0.532715 / 0.226044 (0.306670) | 5.323032 / 2.268929 (3.054104) | 2.326625 / 55.444624 (-53.118000) | 1.944263 / 6.876477 (-4.932213) | 1.994015 / 2.142072 (-0.148058) | 0.813805 / 4.805227 (-3.991422) | 0.149233 / 6.500664 (-6.351431) | 0.065318 / 0.075469 (-0.010151) |\n\n### Benchmark: benchmark_map_filter.json\n\n| metric | filter | map fast-tokenizer batched | map identity | map identity batched | map no-op batched | map no-op batched numpy | map no-op batched pandas | map no-op batched pytorch | map no-op batched tensorflow |\n|--------|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 1.212441 / 1.841788 (-0.629347) | 13.979069 / 8.074308 (5.904761) | 14.003998 / 10.191392 (3.812606) | 0.146956 / 0.680424 (-0.533468) | 0.028564 / 0.534201 (-0.505637) | 0.392370 / 0.579283 (-0.186913) | 0.399695 / 0.434364 (-0.034669) | 0.473481 / 0.540337 (-0.066856) | 0.562625 / 1.386936 (-0.824311) |\n\n</details>\nPyArrow==latest\n\n<details>\n<summary>Show updated benchmarks!</summary>\n\n### Benchmark: benchmark_array_xd.json\n\n| metric | read_batch_formatted_as_numpy after write_array2d | read_batch_formatted_as_numpy after write_flattened_sequence | read_batch_formatted_as_numpy after write_nested_sequence | read_batch_unformated after write_array2d | read_batch_unformated after write_flattened_sequence | read_batch_unformated after write_nested_sequence | read_col_formatted_as_numpy after write_array2d | read_col_formatted_as_numpy after write_flattened_sequence | read_col_formatted_as_numpy after write_nested_sequence | read_col_unformated after write_array2d | read_col_unformated after write_flattened_sequence | read_col_unformated after write_nested_sequence | read_formatted_as_numpy after write_array2d | read_formatted_as_numpy after write_flattened_sequence | read_formatted_as_numpy after write_nested_sequence | read_unformated after write_array2d | read_unformated after write_flattened_sequence | read_unformated after write_nested_sequence | write_array2d | write_flattened_sequence | write_nested_sequence |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.006821 / 0.011353 (-0.004532) | 0.004570 / 0.011008 (-0.006438) | 0.076217 / 0.038508 (0.037709) | 0.028888 / 0.023109 (0.005779) | 0.345431 / 0.275898 (0.069533) | 0.389246 / 0.323480 (0.065766) | 0.005939 / 0.007986 (-0.002046) | 0.003356 / 0.004328 (-0.000973) | 0.075880 / 0.004250 (0.071629) | 0.041427 / 0.037052 (0.004374) | 0.344481 / 0.258489 (0.085992) | 0.398508 / 0.293841 (0.104667) | 0.031801 / 0.128546 (-0.096745) | 0.011763 / 0.075646 (-0.063884) | 0.085600 / 0.419271 (-0.333672) | 0.042656 / 0.043533 (-0.000876) | 0.345893 / 0.255139 (0.090754) | 0.376910 / 0.283200 (0.093711) | 0.092451 / 0.141683 (-0.049232) | 1.461222 / 1.452155 (0.009068) | 1.555822 / 1.492716 (0.063106) |\n\n### Benchmark: benchmark_getitem\\_100B.json\n\n| metric | get_batch_of\\_1024\\_random_rows | get_batch_of\\_1024\\_rows | get_first_row | get_last_row |\n|--------|---|---|---|---|\n| new / old (diff) | 0.235781 / 0.018006 (0.217774) | 0.418485 / 0.000490 (0.417995) | 0.005560 / 0.000200 (0.005360) | 0.000078 / 0.000054 (0.000023) |\n\n### Benchmark: benchmark_indices_mapping.json\n\n| metric | select | shard | shuffle | sort | train_test_split |\n|--------|---|---|---|---|---|\n| new / old (diff) | 0.025410 / 0.037411 (-0.012001) | 0.103780 / 0.014526 (0.089254) | 0.110183 / 0.176557 (-0.066374) | 0.151097 / 0.737135 (-0.586039) | 0.112539 / 0.296338 (-0.183799) |\n\n### Benchmark: benchmark_iterating.json\n\n| metric | read 5000 | read 50000 | read_batch 50000 10 | read_batch 50000 100 | read_batch 50000 1000 | read_formatted numpy 5000 | read_formatted pandas 5000 | read_formatted tensorflow 5000 | read_formatted torch 5000 | read_formatted_batch numpy 5000 10 | read_formatted_batch numpy 5000 1000 | shuffled read 5000 | shuffled read 50000 | shuffled read_batch 50000 10 | shuffled read_batch 50000 100 | shuffled read_batch 50000 1000 | shuffled read_formatted numpy 5000 | shuffled read_formatted_batch numpy 5000 10 | shuffled read_formatted_batch numpy 5000 1000 |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.436686 / 0.215209 (0.221477) | 4.341594 / 2.077655 (2.263940) | 2.062309 / 1.504120 (0.558190) | 1.857461 / 1.541195 (0.316267) | 1.947204 / 1.468490 (0.478713) | 0.699641 / 4.584777 (-3.885136) | 3.406983 / 3.745712 (-0.338729) | 3.294705 / 5.269862 (-1.975157) | 1.360582 / 4.565676 (-3.205095) | 0.083025 / 0.424275 (-0.341250) | 0.012461 / 0.007607 (0.004854) | 0.537767 / 0.226044 (0.311722) | 5.393316 / 2.268929 (3.124387) | 2.516692 / 55.444624 (-52.927932) | 2.163987 / 6.876477 (-4.712490) | 2.220480 / 2.142072 (0.078408) | 0.810648 / 4.805227 (-3.994579) | 0.151820 / 6.500664 (-6.348844) | 0.068080 / 0.075469 (-0.007389) |\n\n### Benchmark: benchmark_map_filter.json\n\n| metric | filter | map fast-tokenizer batched | map identity | map identity batched | map no-op batched | map no-op batched numpy | map no-op batched pandas | map no-op batched pytorch | map no-op batched tensorflow |\n|--------|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 1.279382 / 1.841788 (-0.562405) | 13.989947 / 8.074308 (5.915638) | 14.039229 / 10.191392 (3.847836) | 0.141071 / 0.680424 (-0.539352) | 0.017118 / 0.534201 (-0.517083) | 0.381558 / 0.579283 (-0.197725) | 0.390407 / 0.434364 (-0.043957) | 0.440920 / 0.540337 (-0.099418) | 0.525478 / 1.386936 (-0.861458) |\n\n</details>\n</details>\n\n![](https://cml.dev/watermark.png#eeedb5167d150888a640cd70ca63d6d72bbe1043 \"CML watermark\")\n" ]

[]

[]

[ "I think I need to create another PR on https://huggingface.co/datasets/huggingface/documentation-images/tree/main/datasets for hosting the images there?", "_The documentation is not available anymore as the PR was closed or merged._", "Thanks for the update @awsaf49 !", "> Thanks a lot for the updates!\r\n> \r\n> Just some minor things remain and the we should be good to ship this 🚀\r\n\r\n@sayakpaul I have updated the minor things. Please approve the workflows", "I think this PR is good to go..\r\n@sayakpaul @lhoestq ", "<details>\n<summary>Show benchmarks</summary>\n\nPyArrow==6.0.0\n\n<details>\n<summary>Show updated benchmarks!</summary>\n\n### Benchmark: benchmark_array_xd.json\n\n| metric | read_batch_formatted_as_numpy after write_array2d | read_batch_formatted_as_numpy after write_flattened_sequence | read_batch_formatted_as_numpy after write_nested_sequence | read_batch_unformated after write_array2d | read_batch_unformated after write_flattened_sequence | read_batch_unformated after write_nested_sequence | read_col_formatted_as_numpy after write_array2d | read_col_formatted_as_numpy after write_flattened_sequence | read_col_formatted_as_numpy after write_nested_sequence | read_col_unformated after write_array2d | read_col_unformated after write_flattened_sequence | read_col_unformated after write_nested_sequence | read_formatted_as_numpy after write_array2d | read_formatted_as_numpy after write_flattened_sequence | read_formatted_as_numpy after write_nested_sequence | read_unformated after write_array2d | read_unformated after write_flattened_sequence | read_unformated after write_nested_sequence | write_array2d | write_flattened_sequence | write_nested_sequence |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.009064 / 0.011353 (-0.002289) | 0.005262 / 0.011008 (-0.005746) | 0.099608 / 0.038508 (0.061100) | 0.035015 / 0.023109 (0.011906) | 0.296501 / 0.275898 (0.020602) | 0.353619 / 0.323480 (0.030139) | 0.007903 / 0.007986 (-0.000083) | 0.004093 / 0.004328 (-0.000235) | 0.075260 / 0.004250 (0.071009) | 0.043142 / 0.037052 (0.006089) | 0.307755 / 0.258489 (0.049266) | 0.336340 / 0.293841 (0.042499) | 0.038596 / 0.128546 (-0.089950) | 0.011861 / 0.075646 (-0.063786) | 0.334226 / 0.419271 (-0.085045) | 0.051472 / 0.043533 (0.007940) | 0.298539 / 0.255139 (0.043400) | 0.316856 / 0.283200 (0.033656) | 0.108620 / 0.141683 (-0.033063) | 1.434901 / 1.452155 (-0.017254) | 1.468368 / 1.492716 (-0.024348) |\n\n### Benchmark: benchmark_getitem\\_100B.json\n\n| metric | get_batch_of\\_1024\\_random_rows | get_batch_of\\_1024\\_rows | get_first_row | get_last_row |\n|--------|---|---|---|---|\n| new / old (diff) | 0.208402 / 0.018006 (0.190395) | 0.445799 / 0.000490 (0.445309) | 0.003704 / 0.000200 (0.003504) | 0.000084 / 0.000054 (0.000030) |\n\n### Benchmark: benchmark_indices_mapping.json\n\n| metric | select | shard | shuffle | sort | train_test_split |\n|--------|---|---|---|---|---|\n| new / old (diff) | 0.025435 / 0.037411 (-0.011976) | 0.105874 / 0.014526 (0.091348) | 0.115652 / 0.176557 (-0.060905) | 0.150872 / 0.737135 (-0.586263) | 0.121705 / 0.296338 (-0.174633) |\n\n### Benchmark: benchmark_iterating.json\n\n| metric | read 5000 | read 50000 | read_batch 50000 10 | read_batch 50000 100 | read_batch 50000 1000 | read_formatted numpy 5000 | read_formatted pandas 5000 | read_formatted tensorflow 5000 | read_formatted torch 5000 | read_formatted_batch numpy 5000 10 | read_formatted_batch numpy 5000 1000 | shuffled read 5000 | shuffled read 50000 | shuffled read_batch 50000 10 | shuffled read_batch 50000 100 | shuffled read_batch 50000 1000 | shuffled read_formatted numpy 5000 | shuffled read_formatted_batch numpy 5000 10 | shuffled read_formatted_batch numpy 5000 1000 |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.397816 / 0.215209 (0.182607) | 3.977766 / 2.077655 (1.900111) | 1.850848 / 1.504120 (0.346728) | 1.686062 / 1.541195 (0.144867) | 1.786277 / 1.468490 (0.317787) | 0.696250 / 4.584777 (-3.888527) | 3.785255 / 3.745712 (0.039543) | 3.355013 / 5.269862 (-1.914849) | 1.818232 / 4.565676 (-2.747444) | 0.085408 / 0.424275 (-0.338867) | 0.012567 / 0.007607 (0.004960) | 0.524185 / 0.226044 (0.298140) | 5.061975 / 2.268929 (2.793047) | 2.299866 / 55.444624 (-53.144758) | 1.966709 / 6.876477 (-4.909768) | 2.018760 / 2.142072 (-0.123313) | 0.841341 / 4.805227 (-3.963886) | 0.166374 / 6.500664 (-6.334290) | 0.061854 / 0.075469 (-0.013615) |\n\n### Benchmark: benchmark_map_filter.json\n\n| metric | filter | map fast-tokenizer batched | map identity | map identity batched | map no-op batched | map no-op batched numpy | map no-op batched pandas | map no-op batched pytorch | map no-op batched tensorflow |\n|--------|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 1.221666 / 1.841788 (-0.620122) | 14.373194 / 8.074308 (6.298886) | 14.253614 / 10.191392 (4.062222) | 0.172979 / 0.680424 (-0.507445) | 0.029176 / 0.534201 (-0.505025) | 0.447399 / 0.579283 (-0.131884) | 0.443663 / 0.434364 (0.009299) | 0.537071 / 0.540337 (-0.003267) | 0.640539 / 1.386936 (-0.746397) |\n\n</details>\nPyArrow==latest\n\n<details>\n<summary>Show updated benchmarks!</summary>\n\n### Benchmark: benchmark_array_xd.json\n\n| metric | read_batch_formatted_as_numpy after write_array2d | read_batch_formatted_as_numpy after write_flattened_sequence | read_batch_formatted_as_numpy after write_nested_sequence | read_batch_unformated after write_array2d | read_batch_unformated after write_flattened_sequence | read_batch_unformated after write_nested_sequence | read_col_formatted_as_numpy after write_array2d | read_col_formatted_as_numpy after write_flattened_sequence | read_col_formatted_as_numpy after write_nested_sequence | read_col_unformated after write_array2d | read_col_unformated after write_flattened_sequence | read_col_unformated after write_nested_sequence | read_formatted_as_numpy after write_array2d | read_formatted_as_numpy after write_flattened_sequence | read_formatted_as_numpy after write_nested_sequence | read_unformated after write_array2d | read_unformated after write_flattened_sequence | read_unformated after write_nested_sequence | write_array2d | write_flattened_sequence | write_nested_sequence |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.007019 / 0.011353 (-0.004334) | 0.005091 / 0.011008 (-0.005917) | 0.074588 / 0.038508 (0.036080) | 0.032391 / 0.023109 (0.009282) | 0.340548 / 0.275898 (0.064650) | 0.367159 / 0.323480 (0.043679) | 0.005594 / 0.007986 (-0.002392) | 0.004003 / 0.004328 (-0.000325) | 0.073946 / 0.004250 (0.069695) | 0.045921 / 0.037052 (0.008868) | 0.340245 / 0.258489 (0.081756) | 0.397958 / 0.293841 (0.104117) | 0.036539 / 0.128546 (-0.092007) | 0.012258 / 0.075646 (-0.063388) | 0.087406 / 0.419271 (-0.331865) | 0.049276 / 0.043533 (0.005743) | 0.345235 / 0.255139 (0.090096) | 0.361250 / 0.283200 (0.078050) | 0.100757 / 0.141683 (-0.040926) | 1.464644 / 1.452155 (0.012489) | 1.545852 / 1.492716 (0.053136) |\n\n### Benchmark: benchmark_getitem\\_100B.json\n\n| metric | get_batch_of\\_1024\\_random_rows | get_batch_of\\_1024\\_rows | get_first_row | get_last_row |\n|--------|---|---|---|---|\n| new / old (diff) | 0.222952 / 0.018006 (0.204945) | 0.434607 / 0.000490 (0.434117) | 0.000438 / 0.000200 (0.000238) | 0.000060 / 0.000054 (0.000006) |\n\n### Benchmark: benchmark_indices_mapping.json\n\n| metric | select | shard | shuffle | sort | train_test_split |\n|--------|---|---|---|---|---|\n| new / old (diff) | 0.028834 / 0.037411 (-0.008577) | 0.107523 / 0.014526 (0.092997) | 0.122077 / 0.176557 (-0.054479) | 0.156574 / 0.737135 (-0.580561) | 0.122917 / 0.296338 (-0.173421) |\n\n### Benchmark: benchmark_iterating.json\n\n| metric | read 5000 | read 50000 | read_batch 50000 10 | read_batch 50000 100 | read_batch 50000 1000 | read_formatted numpy 5000 | read_formatted pandas 5000 | read_formatted tensorflow 5000 | read_formatted torch 5000 | read_formatted_batch numpy 5000 10 | read_formatted_batch numpy 5000 1000 | shuffled read 5000 | shuffled read 50000 | shuffled read_batch 50000 10 | shuffled read_batch 50000 100 | shuffled read_batch 50000 1000 | shuffled read_formatted numpy 5000 | shuffled read_formatted_batch numpy 5000 10 | shuffled read_formatted_batch numpy 5000 1000 |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.417292 / 0.215209 (0.202083) | 4.165980 / 2.077655 (2.088325) | 1.996731 / 1.504120 (0.492611) | 1.802946 / 1.541195 (0.261751) | 1.878456 / 1.468490 (0.409966) | 0.711035 / 4.584777 (-3.873742) | 3.847357 / 3.745712 (0.101644) | 2.088354 / 5.269862 (-3.181508) | 1.344763 / 4.565676 (-3.220913) | 0.086356 / 0.424275 (-0.337919) | 0.012530 / 0.007607 (0.004923) | 0.511693 / 0.226044 (0.285648) | 5.126093 / 2.268929 (2.857165) | 2.490023 / 55.444624 (-52.954602) | 2.180274 / 6.876477 (-4.696202) | 2.221511 / 2.142072 (0.079438) | 0.836348 / 4.805227 (-3.968879) | 0.169554 / 6.500664 (-6.331110) | 0.064555 / 0.075469 (-0.010914) |\n\n### Benchmark: benchmark_map_filter.json\n\n| metric | filter | map fast-tokenizer batched | map identity | map identity batched | map no-op batched | map no-op batched numpy | map no-op batched pandas | map no-op batched pytorch | map no-op batched tensorflow |\n|--------|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 1.293466 / 1.841788 (-0.548321) | 14.785700 / 8.074308 (6.711392) | 13.858493 / 10.191392 (3.667101) | 0.161777 / 0.680424 (-0.518646) | 0.017794 / 0.534201 (-0.516407) | 0.426286 / 0.579283 (-0.152997) | 0.422517 / 0.434364 (-0.011847) | 0.530777 / 0.540337 (-0.009560) | 0.634822 / 1.386936 (-0.752114) |\n\n</details>\n</details>\n\n![](https://cml.dev/watermark.png#c6e08fcfc3a04e53430c26fa7c07da4cb18d977d \"CML watermark\")\n" ]

[]

[]

[ "Seems to work now, perhaps it was something internal with our university's network." ]

[ { "color": "a2eeef", "default": true, "description": "New feature or request", "id": 1935892871, "name": "enhancement", "node_id": "MDU6TGFiZWwxOTM1ODkyODcx", "url": "https://api.github.com/repos/huggingface/datasets/labels/enhancement" }, { "color": "BDE59C", "default": false, "description": "Issues a bit more difficult than \"Good First\" issues", "id": 3761482852, "name": "good second issue", "node_id": "LA_kwDODunzps7gM6xk", "url": "https://api.github.com/repos/huggingface/datasets/labels/good%20second%20issue" } ]

[ { "avatar_url": "https://avatars.githubusercontent.com/u/6368040?v=4", "events_url": "https://api.github.com/users/MFreidank/events{/privacy}", "followers_url": "https://api.github.com/users/MFreidank/followers", "following_url": "https://api.github.com/users/MFreidank/following{/other_user}", "gists_url": "https://api.github.com/users/MFreidank/gists{/gist_id}", "gravatar_id": "", "html_url": "https://github.com/MFreidank", "id": 6368040, "login": "MFreidank", "node_id": "MDQ6VXNlcjYzNjgwNDA=", "organizations_url": "https://api.github.com/users/MFreidank/orgs", "received_events_url": "https://api.github.com/users/MFreidank/received_events", "repos_url": "https://api.github.com/users/MFreidank/repos", "site_admin": false, "starred_url": "https://api.github.com/users/MFreidank/starred{/owner}{/repo}", "subscriptions_url": "https://api.github.com/users/MFreidank/subscriptions", "type": "User", "url": "https://api.github.com/users/MFreidank" } ]

[ "I'd be happy to have a look, if nobody else has started working on this yet @lhoestq. \r\n\r\nIt seems to me that for the `arrow` format features are currently attached as metadata [in `datasets.arrow_writer`](https://github.com/huggingface/datasets/blob/5f810b7011a8a4ab077a1847c024d2d9e267b065/src/datasets/arrow_writer.py#L412) and retrieved from the metadata at `load_dataset` time using [`datasets.features.features.from_arrow_schema`](https://github.com/huggingface/datasets/blob/5f810b7011a8a4ab077a1847c024d2d9e267b065/src/datasets/features/features.py#L1602). \r\n\r\nThis will need to be replicated for `parquet` via calls to [this api](https://arrow.apache.org/docs/python/generated/pyarrow.parquet.write_metadata.html) from `io.parquet.ParquetWriter` and `io.parquet.ParquetReader` [respectively](https://github.com/huggingface/datasets/blob/5f810b7011a8a4ab077a1847c024d2d9e267b065/src/datasets/io/parquet.py#L104).\r\n\r\nAny other important considerations?\r\n", "Thanks @MFreidank ! That's correct :)\r\n\r\nReading the metadata to infer the features can be ideally done in the `parquet.py` file in `packaged_builder` when a parquet file is read. You can cast the arrow table to the schema you get from the features.arrow_schema", "#self-assign" ]

[ { "color": "a2eeef", "default": true, "description": "New feature or request", "id": 1935892871, "name": "enhancement", "node_id": "MDU6TGFiZWwxOTM1ODkyODcx", "url": "https://api.github.com/repos/huggingface/datasets/labels/enhancement" }, { "color": "BDE59C", "default": false, "description": "Issues a bit more difficult than \"Good First\" issues", "id": 3761482852, "name": "good second issue", "node_id": "LA_kwDODunzps7gM6xk", "url": "https://api.github.com/repos/huggingface/datasets/labels/good%20second%20issue" } ]

[]

[ "Can I work on this issue? I am pretty new to this.", "Hi ! Sure :) you can comment `#self-assign` to assign yourself to this issue.\r\n\r\nI can give you some pointers to get started:\r\n\r\n`load_dataset` works roughly this way:\r\n1. it instantiate a dataset builder using `load_dataset_builder()`\r\n2. the builder download and prepare the dataset as Arrow files in the cache using `download_and_prepare()`\r\n3. the builder returns a Dataset object with `as_dataset()`\r\n\r\nOne way to approach this would be to implement `as_iterable_dataset()` in `builder.py`.\r\n\r\nAnd similarly to `as_dataset()`, you can use the `ArrowReader`. It has a `get_file_instructions()` method that can be helpful. It gives you the files to read as list of dictionaries with those keys: `filename`, `skip` and `take`.\r\n\r\nThe `skip` and `take` arguments are used in case the user wants to load a subset of the dataset, e.g.\r\n```python\r\nload_dataset(..., split=\"train[:10]\")\r\n```\r\n\r\nLet me know if you have questions or if I can help :)", "This use-case is a bit specific, and `load_dataset` already has enough parameters (plus, `streaming=True` also returns an iterable dataset, so we would have to explain the difference), so I think it would be better to add `IterableDataset.from_file` to the API (more flexible and aligned with the goal from https://github.com/huggingface/datasets/issues/3444) instead.", "> This use-case is a bit specific\r\n\r\nThis allows to use `datasets` for large scale training where map-style datasets are too slow and use too much memory in PyTorch. So I would still consider adding it.\r\n\r\nAlternatively we could add this feature one level bellow:\r\n```python\r\nbuilder = load_dataset_builder(...)\r\nbuilder.download_and_prepare()\r\nids = builder.as_iterable_dataset()\r\n```", "Yes, I see how this can be useful. Still, I think `Dataset.to_iterable` + `IterableDataset.from_file` would be much cleaner in terms of the API design (and more flexible since `load_dataset` can only access the \"initial\" (unprocessed) version of a dataset).\r\n\r\nAnd since it can be tricky to manually find the \"initial\" version of a dataset in the cache, maybe `load_dataset` could return an iterable dataset streamed from the cache if `streaming=True` and the cache is up-to-date. ", "> This allows to use datasets for large scale training where map-style datasets are too slow and use too much memory in PyTorch.\r\n\r\nI second that. e.g. In my last experiment Oscar-en uses 16GB RSS RAM per process and when using multiple processes the host quickly runs out cpu memory. ", ">And since it can be tricky to manually find the \"initial\" version of a dataset in the cache, maybe load_dataset could return an iterable dataset streamed from the cache if streaming=True and the cache is up-to-date.\r\n\r\nThis is exactly the need on JeanZay (HPC) - I have the dataset cache ready, but the compute node is offline, so making streaming work off a local cache would address that need.\r\n\r\nIf you will have a working POC I can be the tester. ", "> Yes, I see how this can be useful. Still, I think Dataset.to_iterable + IterableDataset.from_file would be much cleaner in terms of the API design (and more flexible since load_dataset can only access the \"initial\" (unprocessed) version of a dataset).\r\n\r\nI like `IterableDataset.from_file` as well. On the other hand `Dataset.to_iterable` first requires to load a Dataset object, which can take time depending on your hardware and your dataset size (sometimes 1h+).\r\n\r\n> And since it can be tricky to manually find the \"initial\" version of a dataset in the cache, maybe load_dataset could return an iterable dataset streamed from the cache if streaming=True and the cache is up-to-date.\r\n\r\nThat would definitely do the job. I was suggesting a different parameter just to make explicit the difference between\r\n- streaming from the raw data\r\n- streaming from the local cache\r\n\r\nBut I'd be fine with streaming from cache is the cache is up-to-date since it's always faster. We could log a message as usual to make it explicit that the cache is used", "> I was suggesting a different parameter just to make explicit the difference between\r\n\r\nMosaicML's `streaming` library does the same (tries to stream from the local cache if possible), so logging a message should be explicit enough :).", "Ok ! Sounds good then :)", "Hi Both! It has been a while since my first issue so I am gonna go for this one ! #self-assign", "#self-assign", "I like idea of `IterableDataset.from_file`. ", "https://github.com/huggingface/datasets/pull/5821 should be helpful to implement `IterableDataset.from_file`, since it defines a new ArrowExamplesIterable that takes an Arrow tables generator function (e.g. from a file) and can be used in an IterableDataset", "@lhoestq I have just started working on this issue. ", "@lhoestq Thank you for taking over." ]

[]

[]

[ "_The documentation is not available anymore as the PR was closed or merged._", "<details>\n<summary>Show benchmarks</summary>\n\nPyArrow==6.0.0\n\n<details>\n<summary>Show updated benchmarks!</summary>\n\n### Benchmark: benchmark_array_xd.json\n\n| metric | read_batch_formatted_as_numpy after write_array2d | read_batch_formatted_as_numpy after write_flattened_sequence | read_batch_formatted_as_numpy after write_nested_sequence | read_batch_unformated after write_array2d | read_batch_unformated after write_flattened_sequence | read_batch_unformated after write_nested_sequence | read_col_formatted_as_numpy after write_array2d | read_col_formatted_as_numpy after write_flattened_sequence | read_col_formatted_as_numpy after write_nested_sequence | read_col_unformated after write_array2d | read_col_unformated after write_flattened_sequence | read_col_unformated after write_nested_sequence | read_formatted_as_numpy after write_array2d | read_formatted_as_numpy after write_flattened_sequence | read_formatted_as_numpy after write_nested_sequence | read_unformated after write_array2d | read_unformated after write_flattened_sequence | read_unformated after write_nested_sequence | write_array2d | write_flattened_sequence | write_nested_sequence |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.009963 / 0.011353 (-0.001390) | 0.005512 / 0.011008 (-0.005496) | 0.100495 / 0.038508 (0.061987) | 0.039929 / 0.023109 (0.016820) | 0.299749 / 0.275898 (0.023850) | 0.372330 / 0.323480 (0.048850) | 0.008689 / 0.007986 (0.000703) | 0.004334 / 0.004328 (0.000006) | 0.076469 / 0.004250 (0.072218) | 0.048091 / 0.037052 (0.011039) | 0.303884 / 0.258489 (0.045395) | 0.352747 / 0.293841 (0.058906) | 0.038941 / 0.128546 (-0.089605) | 0.012541 / 0.075646 (-0.063105) | 0.334227 / 0.419271 (-0.085044) | 0.048802 / 0.043533 (0.005269) | 0.295800 / 0.255139 (0.040661) | 0.316222 / 0.283200 (0.033022) | 0.108246 / 0.141683 (-0.033437) | 1.452735 / 1.452155 (0.000580) | 1.466293 / 1.492716 (-0.026423) |\n\n### Benchmark: benchmark_getitem\\_100B.json\n\n| metric | get_batch_of\\_1024\\_random_rows | get_batch_of\\_1024\\_rows | get_first_row | get_last_row |\n|--------|---|---|---|---|\n| new / old (diff) | 0.010497 / 0.018006 (-0.007510) | 0.507427 / 0.000490 (0.506937) | 0.003054 / 0.000200 (0.002854) | 0.000084 / 0.000054 (0.000030) |\n\n### Benchmark: benchmark_indices_mapping.json\n\n| metric | select | shard | shuffle | sort | train_test_split |\n|--------|---|---|---|---|---|\n| new / old (diff) | 0.029529 / 0.037411 (-0.007883) | 0.114151 / 0.014526 (0.099625) | 0.120599 / 0.176557 (-0.055957) | 0.161881 / 0.737135 (-0.575255) | 0.127669 / 0.296338 (-0.168669) |\n\n### Benchmark: benchmark_iterating.json\n\n| metric | read 5000 | read 50000 | read_batch 50000 10 | read_batch 50000 100 | read_batch 50000 1000 | read_formatted numpy 5000 | read_formatted pandas 5000 | read_formatted tensorflow 5000 | read_formatted torch 5000 | read_formatted_batch numpy 5000 10 | read_formatted_batch numpy 5000 1000 | shuffled read 5000 | shuffled read 50000 | shuffled read_batch 50000 10 | shuffled read_batch 50000 100 | shuffled read_batch 50000 1000 | shuffled read_formatted numpy 5000 | shuffled read_formatted_batch numpy 5000 10 | shuffled read_formatted_batch numpy 5000 1000 |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.399631 / 0.215209 (0.184421) | 3.992997 / 2.077655 (1.915343) | 1.803770 / 1.504120 (0.299650) | 1.612301 / 1.541195 (0.071106) | 1.717846 / 1.468490 (0.249356) | 0.706753 / 4.584777 (-3.878024) | 3.798224 / 3.745712 (0.052512) | 2.169733 / 5.269862 (-3.100128) | 1.358264 / 4.565676 (-3.207413) | 0.086828 / 0.424275 (-0.337447) | 0.012606 / 0.007607 (0.004999) | 0.512085 / 0.226044 (0.286041) | 5.101491 / 2.268929 (2.832563) | 2.285688 / 55.444624 (-53.158936) | 1.955160 / 6.876477 (-4.921317) | 2.045887 / 2.142072 (-0.096186) | 0.878836 / 4.805227 (-3.926392) | 0.166483 / 6.500664 (-6.334181) | 0.062656 / 0.075469 (-0.012814) |\n\n### Benchmark: benchmark_map_filter.json\n\n| metric | filter | map fast-tokenizer batched | map identity | map identity batched | map no-op batched | map no-op batched numpy | map no-op batched pandas | map no-op batched pytorch | map no-op batched tensorflow |\n|--------|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 1.215152 / 1.841788 (-0.626636) | 15.436187 / 8.074308 (7.361879) | 14.489951 / 10.191392 (4.298559) | 0.199019 / 0.680424 (-0.481404) | 0.029148 / 0.534201 (-0.505053) | 0.440309 / 0.579283 (-0.138974) | 0.452041 / 0.434364 (0.017677) | 0.527102 / 0.540337 (-0.013236) | 0.634302 / 1.386936 (-0.752634) |\n\n</details>\nPyArrow==latest\n\n<details>\n<summary>Show updated benchmarks!</summary>\n\n### Benchmark: benchmark_array_xd.json\n\n| metric | read_batch_formatted_as_numpy after write_array2d | read_batch_formatted_as_numpy after write_flattened_sequence | read_batch_formatted_as_numpy after write_nested_sequence | read_batch_unformated after write_array2d | read_batch_unformated after write_flattened_sequence | read_batch_unformated after write_nested_sequence | read_col_formatted_as_numpy after write_array2d | read_col_formatted_as_numpy after write_flattened_sequence | read_col_formatted_as_numpy after write_nested_sequence | read_col_unformated after write_array2d | read_col_unformated after write_flattened_sequence | read_col_unformated after write_nested_sequence | read_formatted_as_numpy after write_array2d | read_formatted_as_numpy after write_flattened_sequence | read_formatted_as_numpy after write_nested_sequence | read_unformated after write_array2d | read_unformated after write_flattened_sequence | read_unformated after write_nested_sequence | write_array2d | write_flattened_sequence | write_nested_sequence |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.007814 / 0.011353 (-0.003539) | 0.005582 / 0.011008 (-0.005427) | 0.075466 / 0.038508 (0.036958) | 0.034421 / 0.023109 (0.011312) | 0.342345 / 0.275898 (0.066447) | 0.389943 / 0.323480 (0.066463) | 0.006346 / 0.007986 (-0.001639) | 0.004442 / 0.004328 (0.000113) | 0.074440 / 0.004250 (0.070190) | 0.056383 / 0.037052 (0.019331) | 0.340293 / 0.258489 (0.081804) | 0.394416 / 0.293841 (0.100575) | 0.037217 / 0.128546 (-0.091330) | 0.012597 / 0.075646 (-0.063050) | 0.087005 / 0.419271 (-0.332267) | 0.051626 / 0.043533 (0.008094) | 0.336690 / 0.255139 (0.081551) | 0.369143 / 0.283200 (0.085943) | 0.110764 / 0.141683 (-0.030919) | 1.459003 / 1.452155 (0.006849) | 1.557333 / 1.492716 (0.064617) |\n\n### Benchmark: benchmark_getitem\\_100B.json\n\n| metric | get_batch_of\\_1024\\_random_rows | get_batch_of\\_1024\\_rows | get_first_row | get_last_row |\n|--------|---|---|---|---|\n| new / old (diff) | 0.319596 / 0.018006 (0.301590) | 0.514697 / 0.000490 (0.514207) | 0.005286 / 0.000200 (0.005086) | 0.000086 / 0.000054 (0.000032) |\n\n### Benchmark: benchmark_indices_mapping.json\n\n| metric | select | shard | shuffle | sort | train_test_split |\n|--------|---|---|---|---|---|\n| new / old (diff) | 0.032579 / 0.037411 (-0.004832) | 0.111094 / 0.014526 (0.096568) | 0.127827 / 0.176557 (-0.048730) | 0.169967 / 0.737135 (-0.567168) | 0.133149 / 0.296338 (-0.163189) |\n\n### Benchmark: benchmark_iterating.json\n\n| metric | read 5000 | read 50000 | read_batch 50000 10 | read_batch 50000 100 | read_batch 50000 1000 | read_formatted numpy 5000 | read_formatted pandas 5000 | read_formatted tensorflow 5000 | read_formatted torch 5000 | read_formatted_batch numpy 5000 10 | read_formatted_batch numpy 5000 1000 | shuffled read 5000 | shuffled read 50000 | shuffled read_batch 50000 10 | shuffled read_batch 50000 100 | shuffled read_batch 50000 1000 | shuffled read_formatted numpy 5000 | shuffled read_formatted_batch numpy 5000 10 | shuffled read_formatted_batch numpy 5000 1000 |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.424637 / 0.215209 (0.209428) | 4.217889 / 2.077655 (2.140235) | 2.044844 / 1.504120 (0.540724) | 1.863513 / 1.541195 (0.322319) | 1.975674 / 1.468490 (0.507184) | 0.695493 / 4.584777 (-3.889284) | 3.815562 / 3.745712 (0.069850) | 3.534427 / 5.269862 (-1.735435) | 1.684874 / 4.565676 (-2.880802) | 0.085560 / 0.424275 (-0.338715) | 0.012439 / 0.007607 (0.004832) | 0.541231 / 0.226044 (0.315187) | 5.287166 / 2.268929 (3.018237) | 2.596622 / 55.444624 (-52.848002) | 2.315913 / 6.876477 (-4.560564) | 2.418454 / 2.142072 (0.276381) | 0.838947 / 4.805227 (-3.966281) | 0.168149 / 6.500664 (-6.332515) | 0.066439 / 0.075469 (-0.009030) |\n\n### Benchmark: benchmark_map_filter.json\n\n| metric | filter | map fast-tokenizer batched | map identity | map identity batched | map no-op batched | map no-op batched numpy | map no-op batched pandas | map no-op batched pytorch | map no-op batched tensorflow |\n|--------|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 1.264814 / 1.841788 (-0.576974) | 15.861324 / 8.074308 (7.787016) | 14.352515 / 10.191392 (4.161123) | 0.167032 / 0.680424 (-0.513391) | 0.017766 / 0.534201 (-0.516435) | 0.421821 / 0.579283 (-0.157462) | 0.426657 / 0.434364 (-0.007707) | 0.526742 / 0.540337 (-0.013595) | 0.623851 / 1.386936 (-0.763085) |\n\n</details>\n</details>\n\n![](https://cml.dev/watermark.png#69b19755e9e37b746ef56780a62d21ef20c574d5 \"CML watermark\")\n" ]

[]

[]

[]

[]

[]

[ "_The documentation is not available anymore as the PR was closed or merged._", "<details>\n<summary>Show benchmarks</summary>\n\nPyArrow==6.0.0\n\n<details>\n<summary>Show updated benchmarks!</summary>\n\n### Benchmark: benchmark_array_xd.json\n\n| metric | read_batch_formatted_as_numpy after write_array2d | read_batch_formatted_as_numpy after write_flattened_sequence | read_batch_formatted_as_numpy after write_nested_sequence | read_batch_unformated after write_array2d | read_batch_unformated after write_flattened_sequence | read_batch_unformated after write_nested_sequence | read_col_formatted_as_numpy after write_array2d | read_col_formatted_as_numpy after write_flattened_sequence | read_col_formatted_as_numpy after write_nested_sequence | read_col_unformated after write_array2d | read_col_unformated after write_flattened_sequence | read_col_unformated after write_nested_sequence | read_formatted_as_numpy after write_array2d | read_formatted_as_numpy after write_flattened_sequence | read_formatted_as_numpy after write_nested_sequence | read_unformated after write_array2d | read_unformated after write_flattened_sequence | read_unformated after write_nested_sequence | write_array2d | write_flattened_sequence | write_nested_sequence |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.008167 / 0.011353 (-0.003186) | 0.004404 / 0.011008 (-0.006605) | 0.100462 / 0.038508 (0.061954) | 0.028835 / 0.023109 (0.005726) | 0.326759 / 0.275898 (0.050861) | 0.355150 / 0.323480 (0.031670) | 0.007200 / 0.007986 (-0.000786) | 0.003293 / 0.004328 (-0.001035) | 0.078006 / 0.004250 (0.073756) | 0.033298 / 0.037052 (-0.003754) | 0.307119 / 0.258489 (0.048630) | 0.337689 / 0.293841 (0.043848) | 0.033016 / 0.128546 (-0.095530) | 0.011383 / 0.075646 (-0.064263) | 0.321989 / 0.419271 (-0.097283) | 0.039793 / 0.043533 (-0.003740) | 0.295388 / 0.255139 (0.040249) | 0.322694 / 0.283200 (0.039494) | 0.082989 / 0.141683 (-0.058694) | 1.496701 / 1.452155 (0.044546) | 1.548861 / 1.492716 (0.056145) |\n\n### Benchmark: benchmark_getitem\\_100B.json\n\n| metric | get_batch_of\\_1024\\_random_rows | get_batch_of\\_1024\\_rows | get_first_row | get_last_row |\n|--------|---|---|---|---|\n| new / old (diff) | 0.176587 / 0.018006 (0.158580) | 0.397660 / 0.000490 (0.397170) | 0.001063 / 0.000200 (0.000863) | 0.000070 / 0.000054 (0.000015) |\n\n### Benchmark: benchmark_indices_mapping.json\n\n| metric | select | shard | shuffle | sort | train_test_split |\n|--------|---|---|---|---|---|\n| new / old (diff) | 0.022386 / 0.037411 (-0.015025) | 0.096380 / 0.014526 (0.081854) | 0.103032 / 0.176557 (-0.073525) | 0.135050 / 0.737135 (-0.602086) | 0.105941 / 0.296338 (-0.190397) |\n\n### Benchmark: benchmark_iterating.json\n\n| metric | read 5000 | read 50000 | read_batch 50000 10 | read_batch 50000 100 | read_batch 50000 1000 | read_formatted numpy 5000 | read_formatted pandas 5000 | read_formatted tensorflow 5000 | read_formatted torch 5000 | read_formatted_batch numpy 5000 10 | read_formatted_batch numpy 5000 1000 | shuffled read 5000 | shuffled read 50000 | shuffled read_batch 50000 10 | shuffled read_batch 50000 100 | shuffled read_batch 50000 1000 | shuffled read_formatted numpy 5000 | shuffled read_formatted_batch numpy 5000 10 | shuffled read_formatted_batch numpy 5000 1000 |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.430989 / 0.215209 (0.215780) | 4.310309 / 2.077655 (2.232654) | 2.142596 / 1.504120 (0.638477) | 1.952043 / 1.541195 (0.410848) | 1.817803 / 1.468490 (0.349312) | 0.690026 / 4.584777 (-3.894751) | 3.315413 / 3.745712 (-0.430299) | 3.370336 / 5.269862 (-1.899525) | 1.668707 / 4.565676 (-2.896970) | 0.081860 / 0.424275 (-0.342415) | 0.012493 / 0.007607 (0.004886) | 0.527779 / 0.226044 (0.301735) | 5.318732 / 2.268929 (3.049804) | 2.467029 / 55.444624 (-52.977596) | 2.247171 / 6.876477 (-4.629306) | 2.270825 / 2.142072 (0.128752) | 0.802288 / 4.805227 (-4.002939) | 0.148895 / 6.500664 (-6.351770) | 0.064967 / 0.075469 (-0.010503) |\n\n### Benchmark: benchmark_map_filter.json\n\n| metric | filter | map fast-tokenizer batched | map identity | map identity batched | map no-op batched | map no-op batched numpy | map no-op batched pandas | map no-op batched pytorch | map no-op batched tensorflow |\n|--------|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 1.259304 / 1.841788 (-0.582484) | 13.662441 / 8.074308 (5.588133) | 14.074662 / 10.191392 (3.883270) | 0.152907 / 0.680424 (-0.527516) | 0.028340 / 0.534201 (-0.505861) | 0.397356 / 0.579283 (-0.181927) | 0.392600 / 0.434364 (-0.041764) | 0.467935 / 0.540337 (-0.072402) | 0.539890 / 1.386936 (-0.847046) |\n\n</details>\nPyArrow==latest\n\n<details>\n<summary>Show updated benchmarks!</summary>\n\n### Benchmark: benchmark_array_xd.json\n\n| metric | read_batch_formatted_as_numpy after write_array2d | read_batch_formatted_as_numpy after write_flattened_sequence | read_batch_formatted_as_numpy after write_nested_sequence | read_batch_unformated after write_array2d | read_batch_unformated after write_flattened_sequence | read_batch_unformated after write_nested_sequence | read_col_formatted_as_numpy after write_array2d | read_col_formatted_as_numpy after write_flattened_sequence | read_col_formatted_as_numpy after write_nested_sequence | read_col_unformated after write_array2d | read_col_unformated after write_flattened_sequence | read_col_unformated after write_nested_sequence | read_formatted_as_numpy after write_array2d | read_formatted_as_numpy after write_flattened_sequence | read_formatted_as_numpy after write_nested_sequence | read_unformated after write_array2d | read_unformated after write_flattened_sequence | read_unformated after write_nested_sequence | write_array2d | write_flattened_sequence | write_nested_sequence |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.006156 / 0.011353 (-0.005197) | 0.004371 / 0.011008 (-0.006637) | 0.076391 / 0.038508 (0.037883) | 0.026455 / 0.023109 (0.003346) | 0.339816 / 0.275898 (0.063917) | 0.370032 / 0.323480 (0.046552) | 0.004614 / 0.007986 (-0.003372) | 0.003200 / 0.004328 (-0.001129) | 0.075408 / 0.004250 (0.071157) | 0.034100 / 0.037052 (-0.002953) | 0.341232 / 0.258489 (0.082743) | 0.380290 / 0.293841 (0.086449) | 0.031021 / 0.128546 (-0.097525) | 0.011562 / 0.075646 (-0.064084) | 0.085564 / 0.419271 (-0.333708) | 0.041431 / 0.043533 (-0.002102) | 0.359570 / 0.255139 (0.104431) | 0.366919 / 0.283200 (0.083719) | 0.088242 / 0.141683 (-0.053441) | 1.460703 / 1.452155 (0.008548) | 1.534351 / 1.492716 (0.041635) |\n\n### Benchmark: benchmark_getitem\\_100B.json\n\n| metric | get_batch_of\\_1024\\_random_rows | get_batch_of\\_1024\\_rows | get_first_row | get_last_row |\n|--------|---|---|---|---|\n| new / old (diff) | 0.225703 / 0.018006 (0.207697) | 0.395014 / 0.000490 (0.394524) | 0.000385 / 0.000200 (0.000185) | 0.000060 / 0.000054 (0.000006) |\n\n### Benchmark: benchmark_indices_mapping.json\n\n| metric | select | shard | shuffle | sort | train_test_split |\n|--------|---|---|---|---|---|\n| new / old (diff) | 0.023975 / 0.037411 (-0.013436) | 0.098658 / 0.014526 (0.084132) | 0.105043 / 0.176557 (-0.071513) | 0.139988 / 0.737135 (-0.597148) | 0.106854 / 0.296338 (-0.189484) |\n\n### Benchmark: benchmark_iterating.json\n\n| metric | read 5000 | read 50000 | read_batch 50000 10 | read_batch 50000 100 | read_batch 50000 1000 | read_formatted numpy 5000 | read_formatted pandas 5000 | read_formatted tensorflow 5000 | read_formatted torch 5000 | read_formatted_batch numpy 5000 10 | read_formatted_batch numpy 5000 1000 | shuffled read 5000 | shuffled read 50000 | shuffled read_batch 50000 10 | shuffled read_batch 50000 100 | shuffled read_batch 50000 1000 | shuffled read_formatted numpy 5000 | shuffled read_formatted_batch numpy 5000 10 | shuffled read_formatted_batch numpy 5000 1000 |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.442454 / 0.215209 (0.227245) | 4.430860 / 2.077655 (2.353205) | 2.084823 / 1.504120 (0.580704) | 1.870421 / 1.541195 (0.329226) | 1.901618 / 1.468490 (0.433128) | 0.699214 / 4.584777 (-3.885563) | 3.336911 / 3.745712 (-0.408801) | 1.856479 / 5.269862 (-3.413383) | 1.166496 / 4.565676 (-3.399180) | 0.083189 / 0.424275 (-0.341086) | 0.012293 / 0.007607 (0.004686) | 0.543147 / 0.226044 (0.317102) | 5.452030 / 2.268929 (3.183101) | 2.506689 / 55.444624 (-52.937936) | 2.168186 / 6.876477 (-4.708291) | 2.172277 / 2.142072 (0.030205) | 0.813554 / 4.805227 (-3.991673) | 0.152074 / 6.500664 (-6.348590) | 0.066891 / 0.075469 (-0.008579) |\n\n### Benchmark: benchmark_map_filter.json\n\n| metric | filter | map fast-tokenizer batched | map identity | map identity batched | map no-op batched | map no-op batched numpy | map no-op batched pandas | map no-op batched pytorch | map no-op batched tensorflow |\n|--------|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 1.278635 / 1.841788 (-0.563153) | 13.690232 / 8.074308 (5.615924) | 13.403201 / 10.191392 (3.211809) | 0.128171 / 0.680424 (-0.552253) | 0.016687 / 0.534201 (-0.517514) | 0.378645 / 0.579283 (-0.200638) | 0.382922 / 0.434364 (-0.051442) | 0.467483 / 0.540337 (-0.072854) | 0.559026 / 1.386936 (-0.827910) |\n\n</details>\n</details>\n\n![](https://cml.dev/watermark.png#b262d411ec0e252615a140c4e3e60e7dbd38eef1 \"CML watermark\")\n" ]

[]

[]

[ "@albertvillanova It looks like that issue has been fixed so I made a PR to unpin sqlalchemy! ", "The source issue:\r\n- https://github.com/pandas-dev/pandas/issues/40686\r\n\r\nhas been fixed:\r\n- https://github.com/pandas-dev/pandas/pull/48576\r\n\r\nThe fix was released yesterday (2023-04-03) only in `pandas-2.0.0`:\r\n- https://github.com/pandas-dev/pandas/releases/tag/v2.0.0\r\n\r\nbut it will not be back-ported to `pandas-1`:\r\n- https://github.com/pandas-dev/pandas/pull/48576#issuecomment-1466467159\r\n\r\nAlso note that `pandas-2.0.0` dropped support for Python 3.7:\r\n- https://github.com/pandas-dev/pandas/issues/41678\r\n- https://github.com/pandas-dev/pandas/pull/41989\r\n\r\nTherefore, we cannot unpin `sqlalchemy` until we drop support for Python 3.7 (these Python users cannot use `pandas-2`)." ]

[]

[]

[ "_The documentation is not available anymore as the PR was closed or merged._", "<details>\n<summary>Show benchmarks</summary>\n\nPyArrow==6.0.0\n\n<details>\n<summary>Show updated benchmarks!</summary>\n\n### Benchmark: benchmark_array_xd.json\n\n| metric | read_batch_formatted_as_numpy after write_array2d | read_batch_formatted_as_numpy after write_flattened_sequence | read_batch_formatted_as_numpy after write_nested_sequence | read_batch_unformated after write_array2d | read_batch_unformated after write_flattened_sequence | read_batch_unformated after write_nested_sequence | read_col_formatted_as_numpy after write_array2d | read_col_formatted_as_numpy after write_flattened_sequence | read_col_formatted_as_numpy after write_nested_sequence | read_col_unformated after write_array2d | read_col_unformated after write_flattened_sequence | read_col_unformated after write_nested_sequence | read_formatted_as_numpy after write_array2d | read_formatted_as_numpy after write_flattened_sequence | read_formatted_as_numpy after write_nested_sequence | read_unformated after write_array2d | read_unformated after write_flattened_sequence | read_unformated after write_nested_sequence | write_array2d | write_flattened_sequence | write_nested_sequence |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.012442 / 0.011353 (0.001089) | 0.006274 / 0.011008 (-0.004734) | 0.128249 / 0.038508 (0.089741) | 0.040117 / 0.023109 (0.017008) | 0.383725 / 0.275898 (0.107827) | 0.510494 / 0.323480 (0.187014) | 0.009037 / 0.007986 (0.001051) | 0.008256 / 0.004328 (0.003927) | 0.105329 / 0.004250 (0.101079) | 0.046909 / 0.037052 (0.009857) | 0.401980 / 0.258489 (0.143491) | 0.461332 / 0.293841 (0.167491) | 0.065629 / 0.128546 (-0.062917) | 0.020043 / 0.075646 (-0.055604) | 0.453773 / 0.419271 (0.034501) | 0.063456 / 0.043533 (0.019923) | 0.384458 / 0.255139 (0.129319) | 0.449699 / 0.283200 (0.166499) | 0.118197 / 0.141683 (-0.023486) | 1.915080 / 1.452155 (0.462925) | 1.957132 / 1.492716 (0.464416) |\n\n### Benchmark: benchmark_getitem\\_100B.json\n\n| metric | get_batch_of\\_1024\\_random_rows | get_batch_of\\_1024\\_rows | get_first_row | get_last_row |\n|--------|---|---|---|---|\n| new / old (diff) | 0.209657 / 0.018006 (0.191651) | 0.592478 / 0.000490 (0.591988) | 0.004137 / 0.000200 (0.003937) | 0.000124 / 0.000054 (0.000069) |\n\n### Benchmark: benchmark_indices_mapping.json\n\n| metric | select | shard | shuffle | sort | train_test_split |\n|--------|---|---|---|---|---|\n| new / old (diff) | 0.029607 / 0.037411 (-0.007804) | 0.129559 / 0.014526 (0.115033) | 0.148326 / 0.176557 (-0.028231) | 0.190506 / 0.737135 (-0.546629) | 0.143177 / 0.296338 (-0.153162) |\n\n### Benchmark: benchmark_iterating.json\n\n| metric | read 5000 | read 50000 | read_batch 50000 10 | read_batch 50000 100 | read_batch 50000 1000 | read_formatted numpy 5000 | read_formatted pandas 5000 | read_formatted tensorflow 5000 | read_formatted torch 5000 | read_formatted_batch numpy 5000 10 | read_formatted_batch numpy 5000 1000 | shuffled read 5000 | shuffled read 50000 | shuffled read_batch 50000 10 | shuffled read_batch 50000 100 | shuffled read_batch 50000 1000 | shuffled read_formatted numpy 5000 | shuffled read_formatted_batch numpy 5000 10 | shuffled read_formatted_batch numpy 5000 1000 |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.626166 / 0.215209 (0.410957) | 6.612680 / 2.077655 (4.535026) | 2.432354 / 1.504120 (0.928234) | 2.051482 / 1.541195 (0.510287) | 2.055822 / 1.468490 (0.587332) | 1.210099 / 4.584777 (-3.374678) | 5.498117 / 3.745712 (1.752405) | 3.054838 / 5.269862 (-2.215024) | 2.182875 / 4.565676 (-2.382802) | 0.144518 / 0.424275 (-0.279757) | 0.014132 / 0.007607 (0.006525) | 0.801805 / 0.226044 (0.575761) | 7.911235 / 2.268929 (5.642307) | 3.372762 / 55.444624 (-52.071862) | 2.517266 / 6.876477 (-4.359210) | 2.515329 / 2.142072 (0.373256) | 1.501731 / 4.805227 (-3.303497) | 0.252569 / 6.500664 (-6.248096) | 0.080987 / 0.075469 (0.005518) |\n\n### Benchmark: benchmark_map_filter.json\n\n| metric | filter | map fast-tokenizer batched | map identity | map identity batched | map no-op batched | map no-op batched numpy | map no-op batched pandas | map no-op batched pytorch | map no-op batched tensorflow |\n|--------|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 1.709880 / 1.841788 (-0.131907) | 18.640340 / 8.074308 (10.566032) | 23.560908 / 10.191392 (13.369516) | 0.265680 / 0.680424 (-0.414744) | 0.046438 / 0.534201 (-0.487763) | 0.571973 / 0.579283 (-0.007310) | 0.642425 / 0.434364 (0.208061) | 0.698167 / 0.540337 (0.157830) | 0.842132 / 1.386936 (-0.544804) |\n\n</details>\nPyArrow==latest\n\n<details>\n<summary>Show updated benchmarks!</summary>\n\n### Benchmark: benchmark_array_xd.json\n\n| metric | read_batch_formatted_as_numpy after write_array2d | read_batch_formatted_as_numpy after write_flattened_sequence | read_batch_formatted_as_numpy after write_nested_sequence | read_batch_unformated after write_array2d | read_batch_unformated after write_flattened_sequence | read_batch_unformated after write_nested_sequence | read_col_formatted_as_numpy after write_array2d | read_col_formatted_as_numpy after write_flattened_sequence | read_col_formatted_as_numpy after write_nested_sequence | read_col_unformated after write_array2d | read_col_unformated after write_flattened_sequence | read_col_unformated after write_nested_sequence | read_formatted_as_numpy after write_array2d | read_formatted_as_numpy after write_flattened_sequence | read_formatted_as_numpy after write_nested_sequence | read_unformated after write_array2d | read_unformated after write_flattened_sequence | read_unformated after write_nested_sequence | write_array2d | write_flattened_sequence | write_nested_sequence |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.009268 / 0.011353 (-0.002085) | 0.006052 / 0.011008 (-0.004956) | 0.133448 / 0.038508 (0.094939) | 0.034417 / 0.023109 (0.011308) | 0.435573 / 0.275898 (0.159675) | 0.479642 / 0.323480 (0.156162) | 0.008016 / 0.007986 (0.000030) | 0.006616 / 0.004328 (0.002288) | 0.106256 / 0.004250 (0.102005) | 0.048995 / 0.037052 (0.011942) | 0.450056 / 0.258489 (0.191567) | 0.511027 / 0.293841 (0.217187) | 0.052928 / 0.128546 (-0.075618) | 0.020824 / 0.075646 (-0.054822) | 0.450105 / 0.419271 (0.030834) | 0.062729 / 0.043533 (0.019196) | 0.438887 / 0.255139 (0.183748) | 0.468732 / 0.283200 (0.185532) | 0.116101 / 0.141683 (-0.025582) | 1.909689 / 1.452155 (0.457534) | 2.042007 / 1.492716 (0.549291) |\n\n### Benchmark: benchmark_getitem\\_100B.json\n\n| metric | get_batch_of\\_1024\\_random_rows | get_batch_of\\_1024\\_rows | get_first_row | get_last_row |\n|--------|---|---|---|---|\n| new / old (diff) | 0.198265 / 0.018006 (0.180259) | 0.541799 / 0.000490 (0.541309) | 0.003938 / 0.000200 (0.003738) | 0.000116 / 0.000054 (0.000062) |\n\n### Benchmark: benchmark_indices_mapping.json\n\n| metric | select | shard | shuffle | sort | train_test_split |\n|--------|---|---|---|---|---|\n| new / old (diff) | 0.035933 / 0.037411 (-0.001478) | 0.130754 / 0.014526 (0.116229) | 0.146143 / 0.176557 (-0.030414) | 0.202042 / 0.737135 (-0.535094) | 0.155648 / 0.296338 (-0.140691) |\n\n### Benchmark: benchmark_iterating.json\n\n| metric | read 5000 | read 50000 | read_batch 50000 10 | read_batch 50000 100 | read_batch 50000 1000 | read_formatted numpy 5000 | read_formatted pandas 5000 | read_formatted tensorflow 5000 | read_formatted torch 5000 | read_formatted_batch numpy 5000 10 | read_formatted_batch numpy 5000 1000 | shuffled read 5000 | shuffled read 50000 | shuffled read_batch 50000 10 | shuffled read_batch 50000 100 | shuffled read_batch 50000 1000 | shuffled read_formatted numpy 5000 | shuffled read_formatted_batch numpy 5000 10 | shuffled read_formatted_batch numpy 5000 1000 |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.691123 / 0.215209 (0.475914) | 6.708370 / 2.077655 (4.630715) | 2.957120 / 1.504120 (1.453000) | 2.558350 / 1.541195 (1.017155) | 2.611271 / 1.468490 (1.142781) | 1.327355 / 4.584777 (-3.257422) | 5.755975 / 3.745712 (2.010263) | 3.295556 / 5.269862 (-1.974305) | 2.159831 / 4.565676 (-2.405845) | 0.161409 / 0.424275 (-0.262866) | 0.015470 / 0.007607 (0.007863) | 0.840611 / 0.226044 (0.614567) | 8.550064 / 2.268929 (6.281136) | 3.832013 / 55.444624 (-51.612612) | 3.032909 / 6.876477 (-3.843568) | 3.155651 / 2.142072 (1.013578) | 1.612486 / 4.805227 (-3.192741) | 0.273789 / 6.500664 (-6.226875) | 0.085618 / 0.075469 (0.010149) |\n\n### Benchmark: benchmark_map_filter.json\n\n| metric | filter | map fast-tokenizer batched | map identity | map identity batched | map no-op batched | map no-op batched numpy | map no-op batched pandas | map no-op batched pytorch | map no-op batched tensorflow |\n|--------|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 1.808376 / 1.841788 (-0.033412) | 18.267614 / 8.074308 (10.193306) | 21.047679 / 10.191392 (10.856286) | 0.259089 / 0.680424 (-0.421335) | 0.029211 / 0.534201 (-0.504990) | 0.556303 / 0.579283 (-0.022980) | 0.625264 / 0.434364 (0.190900) | 0.680814 / 0.540337 (0.140476) | 0.810146 / 1.386936 (-0.576790) |\n\n</details>\n</details>\n\n![](https://cml.dev/watermark.png#20ea76c80e07acad78cf67198a4046a982feda21 \"CML watermark\")\n", "<details>\n<summary>Show benchmarks</summary>\n\nPyArrow==6.0.0\n\n<details>\n<summary>Show updated benchmarks!</summary>\n\n### Benchmark: benchmark_array_xd.json\n\n| metric | read_batch_formatted_as_numpy after write_array2d | read_batch_formatted_as_numpy after write_flattened_sequence | read_batch_formatted_as_numpy after write_nested_sequence | read_batch_unformated after write_array2d | read_batch_unformated after write_flattened_sequence | read_batch_unformated after write_nested_sequence | read_col_formatted_as_numpy after write_array2d | read_col_formatted_as_numpy after write_flattened_sequence | read_col_formatted_as_numpy after write_nested_sequence | read_col_unformated after write_array2d | read_col_unformated after write_flattened_sequence | read_col_unformated after write_nested_sequence | read_formatted_as_numpy after write_array2d | read_formatted_as_numpy after write_flattened_sequence | read_formatted_as_numpy after write_nested_sequence | read_unformated after write_array2d | read_unformated after write_flattened_sequence | read_unformated after write_nested_sequence | write_array2d | write_flattened_sequence | write_nested_sequence |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.008779 / 0.011353 (-0.002574) | 0.004644 / 0.011008 (-0.006364) | 0.099814 / 0.038508 (0.061306) | 0.029830 / 0.023109 (0.006721) | 0.299159 / 0.275898 (0.023261) | 0.354815 / 0.323480 (0.031335) | 0.006968 / 0.007986 (-0.001018) | 0.003521 / 0.004328 (-0.000808) | 0.077687 / 0.004250 (0.073437) | 0.035019 / 0.037052 (-0.002034) | 0.309548 / 0.258489 (0.051059) | 0.345228 / 0.293841 (0.051387) | 0.033644 / 0.128546 (-0.094902) | 0.011564 / 0.075646 (-0.064083) | 0.321835 / 0.419271 (-0.097437) | 0.041798 / 0.043533 (-0.001735) | 0.298190 / 0.255139 (0.043051) | 0.328874 / 0.283200 (0.045674) | 0.088175 / 0.141683 (-0.053508) | 1.481755 / 1.452155 (0.029600) | 1.503085 / 1.492716 (0.010369) |\n\n### Benchmark: benchmark_getitem\\_100B.json\n\n| metric | get_batch_of\\_1024\\_random_rows | get_batch_of\\_1024\\_rows | get_first_row | get_last_row |\n|--------|---|---|---|---|\n| new / old (diff) | 0.170930 / 0.018006 (0.152924) | 0.422155 / 0.000490 (0.421666) | 0.001708 / 0.000200 (0.001509) | 0.000083 / 0.000054 (0.000028) |\n\n### Benchmark: benchmark_indices_mapping.json\n\n| metric | select | shard | shuffle | sort | train_test_split |\n|--------|---|---|---|---|---|\n| new / old (diff) | 0.022588 / 0.037411 (-0.014824) | 0.095775 / 0.014526 (0.081249) | 0.103939 / 0.176557 (-0.072618) | 0.138441 / 0.737135 (-0.598694) | 0.107896 / 0.296338 (-0.188442) |\n\n### Benchmark: benchmark_iterating.json\n\n| metric | read 5000 | read 50000 | read_batch 50000 10 | read_batch 50000 100 | read_batch 50000 1000 | read_formatted numpy 5000 | read_formatted pandas 5000 | read_formatted tensorflow 5000 | read_formatted torch 5000 | read_formatted_batch numpy 5000 10 | read_formatted_batch numpy 5000 1000 | shuffled read 5000 | shuffled read 50000 | shuffled read_batch 50000 10 | shuffled read_batch 50000 100 | shuffled read_batch 50000 1000 | shuffled read_formatted numpy 5000 | shuffled read_formatted_batch numpy 5000 10 | shuffled read_formatted_batch numpy 5000 1000 |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.418243 / 0.215209 (0.203034) | 4.171432 / 2.077655 (2.093777) | 1.906029 / 1.504120 (0.401909) | 1.698174 / 1.541195 (0.156979) | 1.748339 / 1.468490 (0.279849) | 0.691026 / 4.584777 (-3.893751) | 3.393354 / 3.745712 (-0.352358) | 2.722412 / 5.269862 (-2.547450) | 1.462439 / 4.565676 (-3.103238) | 0.084713 / 0.424275 (-0.339562) | 0.012131 / 0.007607 (0.004524) | 0.522153 / 0.226044 (0.296109) | 5.197916 / 2.268929 (2.928988) | 2.314270 / 55.444624 (-53.130354) | 1.986599 / 6.876477 (-4.889878) | 2.012757 / 2.142072 (-0.129315) | 0.802540 / 4.805227 (-4.002687) | 0.148673 / 6.500664 (-6.351991) | 0.065924 / 0.075469 (-0.009545) |\n\n### Benchmark: benchmark_map_filter.json\n\n| metric | filter | map fast-tokenizer batched | map identity | map identity batched | map no-op batched | map no-op batched numpy | map no-op batched pandas | map no-op batched pytorch | map no-op batched tensorflow |\n|--------|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 1.263790 / 1.841788 (-0.577998) | 13.874784 / 8.074308 (5.800476) | 13.842276 / 10.191392 (3.650884) | 0.149002 / 0.680424 (-0.531422) | 0.028550 / 0.534201 (-0.505651) | 0.396913 / 0.579283 (-0.182370) | 0.401543 / 0.434364 (-0.032821) | 0.473754 / 0.540337 (-0.066583) | 0.560455 / 1.386936 (-0.826481) |\n\n</details>\nPyArrow==latest\n\n<details>\n<summary>Show updated benchmarks!</summary>\n\n### Benchmark: benchmark_array_xd.json\n\n| metric | read_batch_formatted_as_numpy after write_array2d | read_batch_formatted_as_numpy after write_flattened_sequence | read_batch_formatted_as_numpy after write_nested_sequence | read_batch_unformated after write_array2d | read_batch_unformated after write_flattened_sequence | read_batch_unformated after write_nested_sequence | read_col_formatted_as_numpy after write_array2d | read_col_formatted_as_numpy after write_flattened_sequence | read_col_formatted_as_numpy after write_nested_sequence | read_col_unformated after write_array2d | read_col_unformated after write_flattened_sequence | read_col_unformated after write_nested_sequence | read_formatted_as_numpy after write_array2d | read_formatted_as_numpy after write_flattened_sequence | read_formatted_as_numpy after write_nested_sequence | read_unformated after write_array2d | read_unformated after write_flattened_sequence | read_unformated after write_nested_sequence | write_array2d | write_flattened_sequence | write_nested_sequence |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.006724 / 0.011353 (-0.004629) | 0.004507 / 0.011008 (-0.006502) | 0.098447 / 0.038508 (0.059939) | 0.027888 / 0.023109 (0.004779) | 0.428956 / 0.275898 (0.153058) | 0.451557 / 0.323480 (0.128077) | 0.005056 / 0.007986 (-0.002929) | 0.003363 / 0.004328 (-0.000965) | 0.075990 / 0.004250 (0.071740) | 0.038688 / 0.037052 (0.001635) | 0.421550 / 0.258489 (0.163061) | 0.459480 / 0.293841 (0.165639) | 0.031408 / 0.128546 (-0.097138) | 0.011559 / 0.075646 (-0.064088) | 0.320054 / 0.419271 (-0.099217) | 0.041917 / 0.043533 (-0.001616) | 0.420878 / 0.255139 (0.165739) | 0.444813 / 0.283200 (0.161613) | 0.090409 / 0.141683 (-0.051274) | 1.490058 / 1.452155 (0.037904) | 1.645206 / 1.492716 (0.152489) |\n\n### Benchmark: benchmark_getitem\\_100B.json\n\n| metric | get_batch_of\\_1024\\_random_rows | get_batch_of\\_1024\\_rows | get_first_row | get_last_row |\n|--------|---|---|---|---|\n| new / old (diff) | 0.221105 / 0.018006 (0.203099) | 0.407537 / 0.000490 (0.407047) | 0.000410 / 0.000200 (0.000210) | 0.000059 / 0.000054 (0.000004) |\n\n### Benchmark: benchmark_indices_mapping.json\n\n| metric | select | shard | shuffle | sort | train_test_split |\n|--------|---|---|---|---|---|\n| new / old (diff) | 0.024658 / 0.037411 (-0.012754) | 0.099230 / 0.014526 (0.084705) | 0.107788 / 0.176557 (-0.068769) | 0.143040 / 0.737135 (-0.594096) | 0.109440 / 0.296338 (-0.186899) |\n\n### Benchmark: benchmark_iterating.json\n\n| metric | read 5000 | read 50000 | read_batch 50000 10 | read_batch 50000 100 | read_batch 50000 1000 | read_formatted numpy 5000 | read_formatted pandas 5000 | read_formatted tensorflow 5000 | read_formatted torch 5000 | read_formatted_batch numpy 5000 10 | read_formatted_batch numpy 5000 1000 | shuffled read 5000 | shuffled read 50000 | shuffled read_batch 50000 10 | shuffled read_batch 50000 100 | shuffled read_batch 50000 1000 | shuffled read_formatted numpy 5000 | shuffled read_formatted_batch numpy 5000 10 | shuffled read_formatted_batch numpy 5000 1000 |\n|--------|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 0.453303 / 0.215209 (0.238094) | 4.520376 / 2.077655 (2.442722) | 2.133909 / 1.504120 (0.629789) | 1.926996 / 1.541195 (0.385801) | 2.019870 / 1.468490 (0.551380) | 0.707423 / 4.584777 (-3.877354) | 3.391903 / 3.745712 (-0.353809) | 1.860661 / 5.269862 (-3.409201) | 1.159940 / 4.565676 (-3.405736) | 0.083773 / 0.424275 (-0.340502) | 0.012228 / 0.007607 (0.004621) | 0.554666 / 0.226044 (0.328622) | 5.567564 / 2.268929 (3.298636) | 2.636718 / 55.444624 (-52.807907) | 2.240215 / 6.876477 (-4.636262) | 2.218951 / 2.142072 (0.076879) | 0.817167 / 4.805227 (-3.988060) | 0.151633 / 6.500664 (-6.349032) | 0.066515 / 0.075469 (-0.008954) |\n\n### Benchmark: benchmark_map_filter.json\n\n| metric | filter | map fast-tokenizer batched | map identity | map identity batched | map no-op batched | map no-op batched numpy | map no-op batched pandas | map no-op batched pytorch | map no-op batched tensorflow |\n|--------|---|---|---|---|---|---|---|---|---|\n| new / old (diff) | 1.296665 / 1.841788 (-0.545123) | 13.997898 / 8.074308 (5.923590) | 13.286607 / 10.191392 (3.095215) | 0.148906 / 0.680424 (-0.531518) | 0.016600 / 0.534201 (-0.517601) | 0.377459 / 0.579283 (-0.201824) | 0.379938 / 0.434364 (-0.054426) | 0.461628 / 0.540337 (-0.078709) | 0.550592 / 1.386936 (-0.836344) |\n\n</details>\n</details>\n\n![](https://cml.dev/watermark.png#053f51a3e2adb762236eb29dd02791307f45f02f \"CML watermark\")\n" ]