number
int64 2.59k
7.1k
| title
stringlengths 1
290
| body
stringlengths 0
36.2k
⌀ | labels
listlengths 0
4
| pull_request
dict | is_pull_request
bool 2
classes | comments
listlengths 0
30
|
|---|---|---|---|---|---|---|
6,791
|
`add_faiss_index` raises ValueError: not enough values to unpack (expected 2, got 1)
|
### Describe the bug
Calling `add_faiss_index` on a `Dataset` with a column argument raises a ValueError. The following is the trace
```python
214 def replacement_add(self, x):
215 """Adds vectors to the index.
216 The index must be trained before vectors can be added to it.
217 The vectors are implicitly numbered in sequence. When `n` vectors are
(...)
224 `dtype` must be float32.
225 """
--> 227 n, d = x.shape
228 assert d == self.d
229 x = np.ascontiguousarray(x, dtype='float32')
ValueError: not enough values to unpack (expected 2, got 1)
```
### Steps to reproduce the bug
1. Load any dataset like `ds = datasets.load_dataset("wikimedia/wikipedia", "20231101.en")["train"]`
2. Add an FAISS index on any column `ds.add_faiss_index('title')`
### Expected behavior
The index should be created
### Environment info
- `datasets` version: 2.18.0
- Platform: Linux-6.5.0-26-generic-x86_64-with-glibc2.35
- Python version: 3.9.19
- `huggingface_hub` version: 0.22.2
- PyArrow version: 15.0.2
- Pandas version: 2.2.1
- `fsspec` version: 2024.2.0
- `faiss-cpu` version: 1.8.0
|
[] |
{
"url": null,
"html_url": null,
"diff_url": null,
"patch_url": null,
"merged_at": null
}
| false
|
[
"I realized I was passing a string column to this instead of float. Is it possible to add a warning or error to prevent users from falsely believing there's a bug?",
"Hello!\r\n\r\nI agree that we could add some safeguards around the type of `ds[column]`. At least for FAISS, we need the column to be made of embeddings as FAISS doesn't perform the embeddings itself.\r\n\r\nI can propose a PR sometime this week.",
"@Dref360 thanks for the initiative!"
] |
6,790
|
PyArrow 'Memory mapping file failed: Cannot allocate memory' bug
|
### Describe the bug
Hello,
I've been struggling with a problem using Huggingface datasets caused by PyArrow memory allocation. I finally managed to solve it, and thought to document it since similar issues have been raised here before (https://github.com/huggingface/datasets/issues/5710, https://github.com/huggingface/datasets/issues/6176).
In my case, I was trying to load ~70k dataset files from disk using `datasets.load_from_disk(data_path)` (meaning 70k repeated calls to load_from_disk). This triggered an (uninformative) exception around 64k loaded files:
```
File "pyarrow/io.pxi", line 1053, in pyarrow.lib.memory_map
File "pyarrow/io.pxi", line 1000, in pyarrow.lib.MemoryMappedFile._open
File "pyarrow/error.pxi", line 154, in pyarrow.lib.pyarrow_internal_check_status
File "pyarrow/error.pxi", line 91, in pyarrow.lib.check_status
OSError: Memory mapping file failed: Cannot allocate memory
```
Despite system RAM usage being very low. After a lot of digging around, I discovered that my Ubuntu machine had a limit on the maximum number of memory mapped files in `/proc/sys/vm/max_map_count` set to 65530, which was causing my data loader to crash. Increasing the limit in the file (`echo <new_mmap_size> | sudo tee /proc/sys/vm/max_map_count`) made the issue go away.
While this isn't a bug as such in either Datasets or PyArrow, this behavior can be very confusing to users. Maybe this should be mentioned in documentation? I suspect the other issues raised here about memory mapping OOM errors could actually be consequence of system configuration.
Br,
Lauri
### Steps to reproduce the bug
```
import numpy as np
import pyarrow as pa
import tqdm
# Write some data to disk
arr = pa.array(np.arange(100))
schema = pa.schema([
pa.field('nums', arr.type)
])
with pa.OSFile('arraydata.arrow', 'wb') as sink:
with pa.ipc.new_file(sink, schema=schema) as writer:
batch = pa.record_batch([arr], schema=schema)
writer.write(batch)
# Number of times to open the memory map
nums = 70000
# Read the data back
arrays = [pa.memory_map('arraydata.arrow', 'r') for _ in tqdm.tqdm(range(nums))]
```
### Expected behavior
No errors.
### Environment info
datasets: 2.18.0
pyarrow: 15.0.0
|
[] |
{
"url": null,
"html_url": null,
"diff_url": null,
"patch_url": null,
"merged_at": null
}
| false
|
[] |
6,789
|
Issue with map
|
### Describe the bug
Map has been taking extremely long to preprocess my data.
It seems to process 1000 examples (which it does really fast in about 10 seconds), then it hangs for a good 1-2 minutes, before it moves on to the next batch of 1000 examples.
It also keeps eating up my hard drive space for some reason by creating a file named tmp1335llua that is over 300GB.
Trying to set num_proc to be >1 also gives me the following error: NameError: name 'processor' is not defined
Please advise on how I could optimise this?
### Steps to reproduce the bug
In general, I have been using map as per normal. Here is a snippet of my code:
````
########################### DATASET LOADING AND PREP #########################
def load_custom_dataset(split):
ds = []
if split == 'train':
for dset in args.train_datasets:
ds.append(load_from_disk(dset))
if split == 'test':
for dset in args.test_datasets:
ds.append(load_from_disk(dset))
ds_to_return = concatenate_datasets(ds)
ds_to_return = ds_to_return.shuffle(seed=22)
return ds_to_return
def prepare_dataset(batch):
# load and (possibly) resample audio data to 16kHz
audio = batch["audio"]
# compute log-Mel input features from input audio array
batch["input_features"] = processor.feature_extractor(audio["array"], sampling_rate=audio["sampling_rate"]).input_features[0]
# compute input length of audio sample in seconds
batch["input_length"] = len(audio["array"]) / audio["sampling_rate"]
# optional pre-processing steps
transcription = batch["sentence"]
if do_lower_case:
transcription = transcription.lower()
if do_remove_punctuation:
transcription = normalizer(transcription).strip()
# encode target text to label ids
batch["labels"] = processor.tokenizer(transcription).input_ids
return batch
print('DATASET PREPARATION IN PROGRESS...')
# case 3: combine_and_shuffle is true, only train provided
# load train datasets
train_set = load_custom_dataset('train')
# split dataset
raw_dataset = DatasetDict()
raw_dataset = train_set.train_test_split(test_size = args.test_size, shuffle=True, seed=42)
raw_dataset = raw_dataset.cast_column("audio", Audio(sampling_rate=args.sampling_rate))
print("Before Map:")
print(raw_dataset)
raw_dataset = raw_dataset.map(prepare_dataset, num_proc=1)
print("After Map:")
print(raw_dataset)
````
### Expected behavior
Based on the speed at which map is processing examples, I would expect a 5-6 hours completion for all mapping
However, because it hangs every 1000 examples, I instead roughly estimate it would take about 40 hours!
Moreover, i cant even finish the map because it keeps exponentially eating up my hard drive space
### Environment info
- `datasets` version: 2.18.0
- Platform: Windows-10-10.0.22631-SP0
- Python version: 3.10.14
- `huggingface_hub` version: 0.22.2
- PyArrow version: 15.0.2
- Pandas version: 2.2.1
- `fsspec` version: 2024.2.0
|
[] |
{
"url": null,
"html_url": null,
"diff_url": null,
"patch_url": null,
"merged_at": null
}
| false
|
[
"Default `writer_batch_size `is set to 1000 (see [map](https://huggingface.co/docs/datasets/v2.16.1/en/package_reference/main_classes#datasets.Dataset.map)).\r\nThe \"tmp1335llua\" is probably the temp file it creates while writing to disk.\r\nMaybe try lowering the `writer_batch_size`.\r\n\r\nFor multi-processing you should probably pass the `processor `as an argument (with e.g. partial) to the function or create it inside so that the sub-processes have access to it and maybe add `if __name__ == \"__main__\"` (not sure that's necessary?).\r\n",
"Hi @Modexus,\r\n\r\nThank you very much for the help! Yep after playing around with map, I managed to get the parallel processing to work by implementing it like you suggested.\r\n\r\nRegarding the temp files, it seems like the temp files just keep growing in size as the map continues. Eventually, once map finishes, the temp files are deleted, but they are instead saved as cache .arrow files. These cache files are absolutely gigantic (~ 30-50x the size of the initial dataset!).\r\n\r\nAfter playing around with the `prepare_dataset()` function above, it seems this issue is caused by the following line in the function, where the log-Mel spectrogram of the audio is calculated:\r\n\r\n`# compute log-Mel input features from input audio array\r\n batch[\"input_features\"] = processor.feature_extractor(audio[\"array\"], \r\n sampling_rate=audio[\"sampling_rate\"]).input_features[0]\r\n`\r\n\r\nWhen I remove this line, the final cache files are approximately the same size as the initial dataset.\r\n\r\nCan I check whether this is expected behavior with the whisper feature extractor? I cant imagine the spectrograms are that large!\r\n\r\nThank you so much for the help!",
"I'm having a similar issue with the spectrographs taking up an incredibly large amount of space. (i.e. 100GB for 3GB of audio). Is this really normal behavior?",
"Upon taking a look at the hex contents of the mapped dataset files I found that the overwhelming majority of the data contained within them was duplicated junk similar to this. I'm not very familiar with the inner workings of AI but I have to assume this is an inefficient way of storing data at best and a bug at worst.\r\n\r\n",
"Same problem, dataset.map takes long time to process 12GB raw audio data and create 200GB cache file. Is there any method can run process(map) during train, instead current run \r\nonce and save cache file ? ",
"Same issue here. Just trying to normalise image data for a 300MB dataset, ends up with an 11GB cache. The initial .map() call takes 80s over the 15000 images, but then simply iterating over the dataset takes almost 2 minutes. It should be doing no processing here! Something seems wrong.\r\nkeep_in_memory=True also offers no speedup.\r\nEDIT: Running the normalisation with set_transform (i.e. on the fly) iterates through the dataset in 18s. With no normalisation it takes around 14s. No reason for .map() to take 5 mins!",
"@eufrizz How you handle this using set_transform?\r\nI have a really big dataset of size 1.2TB and i am going to use it for fine-tunning whisper model. if i use map for dataset_preparing function it will take over 20 days!!!",
"> @eufrizz How you handle this using set_transform?\n> I have a really big dataset of size 1.2TB and i am going to use it for fine-tunning whisper model. if i use map for dataset_preparing function it will take over 20 days!!!\n\nJust give the preprocessing function you were using for map to set_transform. Just look at the set_transform documentation. If you're going to do lots of epochs you might be better off just saving the preprocessed data into a new dataset. "
] |
6,788
|
A Question About the Map Function
|
### Describe the bug
Hello,
I have a question regarding the map function in the Hugging Face datasets.
The situation is as follows: when I load a jsonl file using load_dataset(..., streaming=False), and then utilize the map function to process it, I specify that the returned example should be of type Torch.tensor. However, I noticed that after applying the map function, the datatype automatically changes to List, which leads to errors in my program.
I attempted to use load_dataset(..., streaming=True), and this issue no longer occurs. I'm not entirely clear on why this happens. Could you please provide some insights into this?
### Steps to reproduce the bug
1.dataset = load_dataset(xxx, streaming = False)
2. dataset.map(function), function will return torch.Tensor.
3. you will find the format of data in dataset is List.
### Expected behavior
I expected to receieve the format of data is torch.Tensor.
### Environment info
2.18.0
|
[] |
{
"url": null,
"html_url": null,
"diff_url": null,
"patch_url": null,
"merged_at": null
}
| false
|
[
"All data is saved in the arrow format on disk.\r\nIf you return a tensor it gets converted to arrow before saving to disk when using map.\r\n\r\nTo get a tensor when you access data elements you can use `dataset.set_format(\"pt\")`.\r\nNote that this just changes how the data is loaded, not how it is stored.",
"> All data is saved in the arrow format on disk. If you return a tensor it gets converted to arrow before saving to disk when using map.\r\n> \r\n> To get a tensor when you access data elements you can use `dataset.set_format(\"pt\")`. Note that this just changes how the data is loaded, not how it is stored.\r\n\r\nThank you very much for your explanation, I understand what you mean now. So you're saying that when streaming=True, there's no need to convert it to the arrow format and save it to disk. But if we directly load all formats and then convert them into the arrow format after passing through the map function, it will convert torch.Tensor into a List. I see."
] |
6,787
|
TimeoutError in map
|
### Describe the bug
```python
from datasets import Dataset
def worker(example):
while True:
continue
example['a'] = 100
return example
data = Dataset.from_list([{"a": 1}, {"a": 2}])
data = data.map(worker)
print(data[0])
```
I'm implementing a worker function whose runtime will depend on specific examples (e.g., while most examples take 0.01s in worker, several examples may take 50s).
Therefore, I would like to know how the current implementation will handle those subprocesses that require a long (e.g., >= 5min) or even infinite time.
I notice that the current implementation set a timeout of 0.05 second
https://github.com/huggingface/datasets/blob/c3ddb1ef00334a6f973679a51e783905fbc9ef0b/src/datasets/utils/py_utils.py#L674
However, this example code still gets stuck.
### Steps to reproduce the bug
run the example above
### Expected behavior
I want to set a default worker to handle these timeout cases, instead of getting stuck
### Environment info
main branch version
|
[] |
{
"url": null,
"html_url": null,
"diff_url": null,
"patch_url": null,
"merged_at": null
}
| false
|
[
"From my current understanding, this timeout is only used when we need to get the results.\r\n\r\nOne of:\r\n1. All tasks are done\r\n2. One worker died\r\n\r\nYour function should work fine and it's definitely a bug if it doesn't.",
"When one of the `map`'s worker processes crashes, the linked code re-raises an error from the crash and returns it to the caller.\r\n\r\nIf your question is how to limit the time of long-running tasks/worker processes, such functionality doesn't exist in `datasets` (yet), which means you need to implement it yourself.\r\n\r\nE.g., you can implement it using the built-in `signal` module like this:\r\n```python\r\nimport time\r\nimport signal\r\nfrom contextlib import contextmanager\r\n\r\nfrom datasets import Dataset\r\n\r\n\r\n@contextmanager\r\ndef max_exec_time(t):\r\n def raise_timeout_handler(signum, frame):\r\n raise TimeoutError\r\n \r\n orig_handler = signal.getsignal(signal.SIGALRM)\r\n signal.signal(signal.SIGALRM, raise_timeout_handler)\r\n try:\r\n signal.alarm(t)\r\n yield\r\n finally:\r\n signal.alarm(0)\r\n signal.signal(signal.SIGALRM, orig_handler)\r\n\r\n\r\ndef worker(example, rank):\r\n try:\r\n with max_exec_time(20): # 20 sec execution limit\r\n if rank % 2 == 0:\r\n time.sleep(50) # simulate a long-running task\r\n example[\"a\"] = 100\r\n except TimeoutError:\r\n example[\"a\"] = None # Or return empty batches here in the \"batched\" mode\r\n return example\r\n\r\ndata = Dataset.from_list([{\"a\": 1}, {\"a\": 2}])\r\ndata = data.map(worker, num_proc=2, with_rank=True)\r\nprint(data[0])\r\n```",
"> From my current understanding, this timeout is only used when we need to get the results.\r\n> \r\n> One of:\r\n> \r\n> 1. All tasks are done\r\n> 2. One worker died\r\n> \r\n> Your function should work fine and it's definitely a bug if it doesn't.\r\n\r\nthanks for responding! can you reproduce the stuck with the above example code?",
"> When one of the `map`'s worker processes crashes, the linked code re-raises an error from the crash and returns it to the caller.\r\n> \r\n> If your question is how to limit the time of long-running tasks/worker processes, such functionality doesn't exist in `datasets` (yet), which means you need to implement it yourself.\r\n> \r\n> E.g., you can implement it using the built-in `signal` module like this:\r\n> \r\n> ```python\r\n> import time\r\n> import signal\r\n> from contextlib import contextmanager\r\n> \r\n> from datasets import Dataset\r\n> \r\n> \r\n> @contextmanager\r\n> def max_exec_time(t):\r\n> def raise_timeout_handler(signum, frame):\r\n> raise TimeoutError\r\n> \r\n> orig_handler = signal.getsignal(signal.SIGALRM)\r\n> signal.signal(signal.SIGALRM, raise_timeout_handler)\r\n> try:\r\n> signal.alarm(t)\r\n> yield\r\n> finally:\r\n> signal.alarm(0)\r\n> signal.signal(signal.SIGALRM, orig_handler)\r\n> \r\n> \r\n> def worker(example, rank):\r\n> try:\r\n> with max_exec_time(20): # 20 sec execution limit\r\n> if rank % 2 == 0:\r\n> time.sleep(50) # simulate a long-running task\r\n> example[\"a\"] = 100\r\n> except TimeoutError:\r\n> example[\"a\"] = None # Or return empty batches here in the \"batched\" mode\r\n> return example\r\n> \r\n> data = Dataset.from_list([{\"a\": 1}, {\"a\": 2}])\r\n> data = data.map(worker, num_proc=2, with_rank=True)\r\n> print(data[0])\r\n> ```\r\n\r\nthanks for responding! However, I don't think we should use `signal` in the context of multiprocessing since sometimes it will crash one process and raise the following error\r\nhttps://github.com/huggingface/datasets/blob/c3ddb1ef00334a6f973679a51e783905fbc9ef0b/src/datasets/utils/py_utils.py#L664",
"> thanks for responding! However, I don't think we should use signal in the context of multiprocessing since sometimes it will crash one process and raise the following error\r\n\r\nThe above code has `try/except` to catch the error from the handler. Or do you get an error other than `TimeoutError`?",
"> > thanks for responding! However, I don't think we should use signal in the context of multiprocessing since sometimes it will crash one process and raise the following error\r\n> \r\n> The above code has `try/except` to catch the error from the handler. Or do you get an error other than `TimeoutError`?\r\n\r\nyup, it will raise the RuntimeError: https://github.com/huggingface/datasets/blob/c3ddb1ef00334a6f973679a51e783905fbc9ef0b/src/datasets/utils/py_utils.py#L667C19-L670C22\r\n\r\n```\r\n raise RuntimeError(\r\n \"One of the subprocesses has abruptly died during map operation.\"\r\n \"To debug the error, disable multiprocessing.\"\r\n )\r\n```"
] |
6,786
|
Make Image cast storage faster
|
PR for issue #6782.
Makes `cast_storage` of the `Image` class faster by removing the slow call to `.pylist`.
Instead directly convert each `ListArray` item to either `Array2DExtensionType` or `Array3DExtensionType`.
This also preserves the `dtype` removing the warning if the array is already `uint8`.
|
[] |
{
"url": "https://api.github.com/repos/huggingface/datasets/pulls/6786",
"html_url": "https://github.com/huggingface/datasets/pull/6786",
"diff_url": "https://github.com/huggingface/datasets/pull/6786.diff",
"patch_url": "https://github.com/huggingface/datasets/pull/6786.patch",
"merged_at": null
}
| true
|
[
"The docs for this PR live [here](https://moon-ci-docs.huggingface.co/docs/datasets/pr_6786). All of your documentation changes will be reflected on that endpoint. The docs are available until 30 days after the last update.",
"Hi ! Thanks for diving into this, this conversion to python lists is indeed quite slow.\r\n\r\nArray2DExtensionType and Array3DExtensionType currently rely on pyarrow lists, but we will soon modify them to use FixedShapeTensorArray instead which is more efficient (e.g. doesn't need to store an offset for each value). So ideally it would be cool to speed this code up without using those extension types or it will be blocking to improve Array2DExtensionType and Array3DExtensionType.\r\n\r\nIf I understand correctly you just need the logic from ArrayExtensionArray.to_numpy ? If so feel free to make a separate function and ArrayExtensionArray.to_numpy can call it"
] |
6,785
|
rename datasets-server to dataset-viewer
|
See https://github.com/huggingface/dataset-viewer/issues/2650
Tell me if it's OK, or if it's a breaking change that must be handled differently.
Also note that the docs page is still https://huggingface.co/docs/datasets-server/, so I didn't change it.
And the API URL is still https://datasets-server.huggingface.co/ (and [might always be](https://github.com/huggingface/dataset-viewer/issues/2666)), so I let it too.
|
[] |
{
"url": "https://api.github.com/repos/huggingface/datasets/pulls/6785",
"html_url": "https://github.com/huggingface/datasets/pull/6785",
"diff_url": "https://github.com/huggingface/datasets/pull/6785.diff",
"patch_url": "https://github.com/huggingface/datasets/pull/6785.patch",
"merged_at": "2024-04-08T12:35:02"
}
| true
|
[
"The docs for this PR live [here](https://moon-ci-docs.huggingface.co/docs/datasets/pr_6785). All of your documentation changes will be reflected on that endpoint. The docs are available until 30 days after the last update.",
"<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) | 0.005224 / 0.011353 (-0.006129) | 0.003938 / 0.011008 (-0.007070) | 0.063829 / 0.038508 (0.025321) | 0.030975 / 0.023109 (0.007865) | 0.265090 / 0.275898 (-0.010808) | 0.290994 / 0.323480 (-0.032486) | 0.003083 / 0.007986 (-0.004902) | 0.002810 / 0.004328 (-0.001518) | 0.048860 / 0.004250 (0.044609) | 0.044663 / 0.037052 (0.007611) | 0.272161 / 0.258489 (0.013672) | 0.306966 / 0.293841 (0.013125) | 0.028028 / 0.128546 (-0.100518) | 0.010616 / 0.075646 (-0.065031) | 0.211649 / 0.419271 (-0.207623) | 0.035906 / 0.043533 (-0.007626) | 0.251779 / 0.255139 (-0.003360) | 0.275543 / 0.283200 (-0.007657) | 0.017710 / 0.141683 (-0.123973) | 1.127015 / 1.452155 (-0.325139) | 1.173319 / 1.492716 (-0.319397) |\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) | 0.090625 / 0.018006 (0.072619) | 0.301973 / 0.000490 (0.301483) | 0.000217 / 0.000200 (0.000017) | 0.000053 / 0.000054 (-0.000002) |\n\n### Benchmark: benchmark_indices_mapping.json\n\n| metric | select | shard | shuffle | sort | train_test_split |\n|--------|---|---|---|---|---|\n| new / old (diff) | 0.018868 / 0.037411 (-0.018543) | 0.062402 / 0.014526 (0.047876) | 0.074053 / 0.176557 (-0.102504) | 0.121484 / 0.737135 (-0.615652) | 0.078674 / 0.296338 (-0.217664) |\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) | 0.277821 / 0.215209 (0.062612) | 2.761642 / 2.077655 (0.683987) | 1.452735 / 1.504120 (-0.051385) | 1.336303 / 1.541195 (-0.204891) | 1.343045 / 1.468490 (-0.125445) | 0.560917 / 4.584777 (-4.023860) | 2.353427 / 3.745712 (-1.392286) | 2.699067 / 5.269862 (-2.570795) | 1.704752 / 4.565676 (-2.860925) | 0.062668 / 0.424275 (-0.361607) | 0.005120 / 0.007607 (-0.002487) | 0.330455 / 0.226044 (0.104410) | 3.264604 / 2.268929 (0.995675) | 1.791940 / 55.444624 (-53.652685) | 1.526083 / 6.876477 (-5.350394) | 1.541429 / 2.142072 (-0.600643) | 0.630343 / 4.805227 (-4.174884) | 0.115189 / 6.500664 (-6.385475) | 0.041716 / 0.075469 (-0.033753) |\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) | 0.975008 / 1.841788 (-0.866779) | 11.326924 / 8.074308 (3.252616) | 9.810300 / 10.191392 (-0.381092) | 0.141068 / 0.680424 (-0.539356) | 0.013950 / 0.534201 (-0.520251) | 0.285691 / 0.579283 (-0.293592) | 0.257968 / 0.434364 (-0.176396) | 0.322976 / 0.540337 (-0.217361) | 0.411114 / 1.386936 (-0.975822) |\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) | 0.005176 / 0.011353 (-0.006177) | 0.003631 / 0.011008 (-0.007377) | 0.050006 / 0.038508 (0.011498) | 0.030622 / 0.023109 (0.007513) | 0.277364 / 0.275898 (0.001466) | 0.299752 / 0.323480 (-0.023728) | 0.004110 / 0.007986 (-0.003876) | 0.002694 / 0.004328 (-0.001634) | 0.048966 / 0.004250 (0.044715) | 0.039634 / 0.037052 (0.002582) | 0.289959 / 0.258489 (0.031470) | 0.320689 / 0.293841 (0.026848) | 0.029285 / 0.128546 (-0.099261) | 0.010435 / 0.075646 (-0.065211) | 0.057432 / 0.419271 (-0.361840) | 0.032554 / 0.043533 (-0.010979) | 0.277354 / 0.255139 (0.022215) | 0.296872 / 0.283200 (0.013673) | 0.017338 / 0.141683 (-0.124344) | 1.134174 / 1.452155 (-0.317981) | 1.184695 / 1.492716 (-0.308021) |\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) | 0.089953 / 0.018006 (0.071947) | 0.299372 / 0.000490 (0.298882) | 0.000212 / 0.000200 (0.000012) | 0.000043 / 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) | 0.021349 / 0.037411 (-0.016062) | 0.075167 / 0.014526 (0.060641) | 0.085910 / 0.176557 (-0.090647) | 0.124729 / 0.737135 (-0.612406) | 0.088313 / 0.296338 (-0.208025) |\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) | 0.291939 / 0.215209 (0.076730) | 2.851077 / 2.077655 (0.773423) | 1.609382 / 1.504120 (0.105262) | 1.469656 / 1.541195 (-0.071539) | 1.490469 / 1.468490 (0.021979) | 0.570421 / 4.584777 (-4.014356) | 2.441438 / 3.745712 (-1.304274) | 2.756514 / 5.269862 (-2.513347) | 1.714202 / 4.565676 (-2.851474) | 0.063656 / 0.424275 (-0.360619) | 0.005640 / 0.007607 (-0.001967) | 0.336240 / 0.226044 (0.110196) | 3.355434 / 2.268929 (1.086505) | 1.947553 / 55.444624 (-53.497072) | 1.672776 / 6.876477 (-5.203700) | 1.685316 / 2.142072 (-0.456757) | 0.638849 / 4.805227 (-4.166378) | 0.116304 / 6.500664 (-6.384360) | 0.041588 / 0.075469 (-0.033881) |\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) | 1.026700 / 1.841788 (-0.815088) | 12.044628 / 8.074308 (3.970319) | 10.464007 / 10.191392 (0.272615) | 0.156169 / 0.680424 (-0.524255) | 0.015624 / 0.534201 (-0.518577) | 0.287233 / 0.579283 (-0.292050) | 0.270374 / 0.434364 (-0.163990) | 0.325255 / 0.540337 (-0.215083) | 0.412021 / 1.386936 (-0.974915) |\n\n</details>\n</details>\n\n![](https://cml.dev/watermark.png#6f7f1718e3db54d7923ebe4383301fdd380c18b9 \"CML watermark\")\n"
] |
6,784
|
Extract data on the fly in packaged builders
|
Instead of waiting for data files to be extracted in the packaged builders, we can prepend the compression prefix and extract them as they are being read (using `fsspec`). This saves disk space (deleting extracted archives is not set by default) and slightly speeds up dataset generation (less disk reads)
|
[] |
{
"url": "https://api.github.com/repos/huggingface/datasets/pulls/6784",
"html_url": "https://github.com/huggingface/datasets/pull/6784",
"diff_url": "https://github.com/huggingface/datasets/pull/6784.diff",
"patch_url": "https://github.com/huggingface/datasets/pull/6784.patch",
"merged_at": "2024-04-16T16:31:29"
}
| true
|
[
"The docs for this PR live [here](https://moon-ci-docs.huggingface.co/docs/datasets/pr_6784). All of your documentation changes will be reflected on that endpoint. The docs are available until 30 days after the last update.",
"CI failures are unrelated, so this is ready for the review",
"<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) | 0.005130 / 0.011353 (-0.006223) | 0.003784 / 0.011008 (-0.007224) | 0.064899 / 0.038508 (0.026391) | 0.029456 / 0.023109 (0.006347) | 0.253384 / 0.275898 (-0.022514) | 0.273509 / 0.323480 (-0.049971) | 0.004116 / 0.007986 (-0.003870) | 0.002713 / 0.004328 (-0.001615) | 0.053984 / 0.004250 (0.049733) | 0.043538 / 0.037052 (0.006485) | 0.264696 / 0.258489 (0.006207) | 0.298321 / 0.293841 (0.004480) | 0.027916 / 0.128546 (-0.100630) | 0.010734 / 0.075646 (-0.064912) | 0.208284 / 0.419271 (-0.210988) | 0.035873 / 0.043533 (-0.007659) | 0.251028 / 0.255139 (-0.004111) | 0.270835 / 0.283200 (-0.012364) | 0.017475 / 0.141683 (-0.124208) | 1.130728 / 1.452155 (-0.321426) | 1.188672 / 1.492716 (-0.304044) |\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) | 0.094191 / 0.018006 (0.076185) | 0.304064 / 0.000490 (0.303575) | 0.000251 / 0.000200 (0.000051) | 0.000058 / 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) | 0.018414 / 0.037411 (-0.018998) | 0.061550 / 0.014526 (0.047024) | 0.074200 / 0.176557 (-0.102357) | 0.120250 / 0.737135 (-0.616885) | 0.076018 / 0.296338 (-0.220321) |\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) | 0.302517 / 0.215209 (0.087308) | 2.943936 / 2.077655 (0.866282) | 1.584847 / 1.504120 (0.080727) | 1.464501 / 1.541195 (-0.076694) | 1.472402 / 1.468490 (0.003912) | 0.570971 / 4.584777 (-4.013806) | 2.383207 / 3.745712 (-1.362505) | 2.811520 / 5.269862 (-2.458342) | 1.746997 / 4.565676 (-2.818680) | 0.063391 / 0.424275 (-0.360884) | 0.005296 / 0.007607 (-0.002311) | 0.358948 / 0.226044 (0.132903) | 3.604704 / 2.268929 (1.335776) | 1.935813 / 55.444624 (-53.508812) | 1.659944 / 6.876477 (-5.216533) | 1.687151 / 2.142072 (-0.454922) | 0.658044 / 4.805227 (-4.147183) | 0.120425 / 6.500664 (-6.380240) | 0.042694 / 0.075469 (-0.032775) |\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) | 0.986308 / 1.841788 (-0.855479) | 11.727945 / 8.074308 (3.653637) | 9.532785 / 10.191392 (-0.658607) | 0.140071 / 0.680424 (-0.540352) | 0.013472 / 0.534201 (-0.520729) | 0.285828 / 0.579283 (-0.293455) | 0.261571 / 0.434364 (-0.172793) | 0.323114 / 0.540337 (-0.217223) | 0.418132 / 1.386936 (-0.968804) |\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) | 0.005428 / 0.011353 (-0.005925) | 0.003954 / 0.011008 (-0.007054) | 0.050336 / 0.038508 (0.011828) | 0.029941 / 0.023109 (0.006831) | 0.281483 / 0.275898 (0.005585) | 0.304822 / 0.323480 (-0.018658) | 0.004151 / 0.007986 (-0.003835) | 0.002862 / 0.004328 (-0.001466) | 0.049196 / 0.004250 (0.044945) | 0.040266 / 0.037052 (0.003213) | 0.293515 / 0.258489 (0.035026) | 0.319165 / 0.293841 (0.025324) | 0.029186 / 0.128546 (-0.099360) | 0.010838 / 0.075646 (-0.064809) | 0.058789 / 0.419271 (-0.360483) | 0.032847 / 0.043533 (-0.010686) | 0.280164 / 0.255139 (0.025025) | 0.299609 / 0.283200 (0.016410) | 0.018291 / 0.141683 (-0.123392) | 1.153858 / 1.452155 (-0.298297) | 1.219108 / 1.492716 (-0.273608) |\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) | 0.093783 / 0.018006 (0.075777) | 0.301526 / 0.000490 (0.301037) | 0.000211 / 0.000200 (0.000011) | 0.000055 / 0.000054 (0.000001) |\n\n### Benchmark: benchmark_indices_mapping.json\n\n| metric | select | shard | shuffle | sort | train_test_split |\n|--------|---|---|---|---|---|\n| new / old (diff) | 0.022105 / 0.037411 (-0.015306) | 0.074844 / 0.014526 (0.060318) | 0.087147 / 0.176557 (-0.089409) | 0.127678 / 0.737135 (-0.609457) | 0.088630 / 0.296338 (-0.207709) |\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) | 0.286805 / 0.215209 (0.071596) | 2.828664 / 2.077655 (0.751009) | 1.579771 / 1.504120 (0.075651) | 1.463137 / 1.541195 (-0.078058) | 1.509238 / 1.468490 (0.040748) | 0.583425 / 4.584777 (-4.001352) | 2.424905 / 3.745712 (-1.320807) | 2.819354 / 5.269862 (-2.450508) | 1.784695 / 4.565676 (-2.780981) | 0.063374 / 0.424275 (-0.360901) | 0.005337 / 0.007607 (-0.002270) | 0.342291 / 0.226044 (0.116247) | 3.404319 / 2.268929 (1.135390) | 1.956909 / 55.444624 (-53.487716) | 1.694317 / 6.876477 (-5.182160) | 1.696256 / 2.142072 (-0.445817) | 0.655748 / 4.805227 (-4.149480) | 0.116785 / 6.500664 (-6.383879) | 0.040930 / 0.075469 (-0.034539) |\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) | 1.034463 / 1.841788 (-0.807325) | 12.252041 / 8.074308 (4.177733) | 10.593960 / 10.191392 (0.402568) | 0.139311 / 0.680424 (-0.541112) | 0.016177 / 0.534201 (-0.518023) | 0.288910 / 0.579283 (-0.290373) | 0.281588 / 0.434364 (-0.152776) | 0.323066 / 0.540337 (-0.217272) | 0.427604 / 1.386936 (-0.959332) |\n\n</details>\n</details>\n\n![](https://cml.dev/watermark.png#a188022dc43a76a119d90c03832d51d6e4a94d91 \"CML watermark\")\n"
] |
6,783
|
AttributeError: module 'numpy' has no attribute 'object'. in Kaggle Notebook
|
### Describe the bug
# problem
I can't resample audio dataset in Kaggle Notebook. It looks like some code in `datasets` library use aliases that were deprecated in NumPy 1.20.
## code for resampling
```
from datasets import load_dataset, Audio
from transformers import AutoFeatureExtractor
from transformers import AutoModelForAudioClassification, TrainingArguments, Trainer
minds = load_dataset("PolyAI/minds14", name="en-US", split="train")
feature_extractor = AutoFeatureExtractor.from_pretrained("facebook/wav2vec2-base")
def preprocess_function(examples):
audio_arrays = [x["array"] for x in examples["audio"]]
inputs = feature_extractor(
audio_arrays, sampling_rate=feature_extractor.sampling_rate, max_length=16000, truncation=True
)
return inputs
dataset = dataset.map(preprocess_function, remove_columns="audio", batched=True, batch_size=100)
```
## the error I got
<details>
<summary>Click to expand</summary>
```
---------------------------------------------------------------------------
AttributeError Traceback (most recent call last)
Cell In[20], line 1
----> 1 dataset = dataset.map(preprocess_function, remove_columns="audio", batched=True, batch_size=100)
2 dataset
File /opt/conda/lib/python3.10/site-packages/datasets/arrow_dataset.py:1955, in Dataset.map(self, function, with_indices, with_rank, input_columns, batched, batch_size, drop_last_batch, remove_columns, keep_in_memory, load_from_cache_file, cache_file_name, writer_batch_size, features, disable_nullable, fn_kwargs, num_proc, suffix_template, new_fingerprint, desc)
1952 disable_tqdm = not logging.is_progress_bar_enabled()
1954 if num_proc is None or num_proc == 1:
-> 1955 return self._map_single(
1956 function=function,
1957 with_indices=with_indices,
1958 with_rank=with_rank,
1959 input_columns=input_columns,
1960 batched=batched,
1961 batch_size=batch_size,
1962 drop_last_batch=drop_last_batch,
1963 remove_columns=remove_columns,
1964 keep_in_memory=keep_in_memory,
1965 load_from_cache_file=load_from_cache_file,
1966 cache_file_name=cache_file_name,
1967 writer_batch_size=writer_batch_size,
1968 features=features,
1969 disable_nullable=disable_nullable,
1970 fn_kwargs=fn_kwargs,
1971 new_fingerprint=new_fingerprint,
1972 disable_tqdm=disable_tqdm,
1973 desc=desc,
1974 )
1975 else:
1977 def format_cache_file_name(cache_file_name, rank):
File /opt/conda/lib/python3.10/site-packages/datasets/arrow_dataset.py:520, in transmit_tasks.<locals>.wrapper(*args, **kwargs)
518 self: "Dataset" = kwargs.pop("self")
519 # apply actual function
--> 520 out: Union["Dataset", "DatasetDict"] = func(self, *args, **kwargs)
521 datasets: List["Dataset"] = list(out.values()) if isinstance(out, dict) else [out]
522 for dataset in datasets:
523 # Remove task templates if a column mapping of the template is no longer valid
File /opt/conda/lib/python3.10/site-packages/datasets/arrow_dataset.py:487, in transmit_format.<locals>.wrapper(*args, **kwargs)
480 self_format = {
481 "type": self._format_type,
482 "format_kwargs": self._format_kwargs,
483 "columns": self._format_columns,
484 "output_all_columns": self._output_all_columns,
485 }
486 # apply actual function
--> 487 out: Union["Dataset", "DatasetDict"] = func(self, *args, **kwargs)
488 datasets: List["Dataset"] = list(out.values()) if isinstance(out, dict) else [out]
489 # re-apply format to the output
File /opt/conda/lib/python3.10/site-packages/datasets/fingerprint.py:458, in fingerprint_transform.<locals>._fingerprint.<locals>.wrapper(*args, **kwargs)
452 kwargs[fingerprint_name] = update_fingerprint(
453 self._fingerprint, transform, kwargs_for_fingerprint
454 )
456 # Call actual function
--> 458 out = func(self, *args, **kwargs)
460 # Update fingerprint of in-place transforms + update in-place history of transforms
462 if inplace: # update after calling func so that the fingerprint doesn't change if the function fails
File /opt/conda/lib/python3.10/site-packages/datasets/arrow_dataset.py:2356, in Dataset._map_single(self, function, with_indices, with_rank, input_columns, batched, batch_size, drop_last_batch, remove_columns, keep_in_memory, load_from_cache_file, cache_file_name, writer_batch_size, features, disable_nullable, fn_kwargs, new_fingerprint, rank, offset, disable_tqdm, desc, cache_only)
2354 writer.write_table(batch)
2355 else:
-> 2356 writer.write_batch(batch)
2357 if update_data and writer is not None:
2358 writer.finalize() # close_stream=bool(buf_writer is None)) # We only close if we are writing in a file
File /opt/conda/lib/python3.10/site-packages/datasets/arrow_writer.py:507, in ArrowWriter.write_batch(self, batch_examples, writer_batch_size)
505 col_try_type = try_features[col] if try_features is not None and col in try_features else None
506 typed_sequence = OptimizedTypedSequence(batch_examples[col], type=col_type, try_type=col_try_type, col=col)
--> 507 arrays.append(pa.array(typed_sequence))
508 inferred_features[col] = typed_sequence.get_inferred_type()
509 schema = inferred_features.arrow_schema if self.pa_writer is None else self.schema
File /opt/conda/lib/python3.10/site-packages/pyarrow/array.pxi:236, in pyarrow.lib.array()
File /opt/conda/lib/python3.10/site-packages/pyarrow/array.pxi:110, in pyarrow.lib._handle_arrow_array_protocol()
File /opt/conda/lib/python3.10/site-packages/datasets/arrow_writer.py:184, in TypedSequence.__arrow_array__(self, type)
182 out = numpy_to_pyarrow_listarray(data)
183 elif isinstance(data, list) and data and isinstance(first_non_null_value(data)[1], np.ndarray):
--> 184 out = list_of_np_array_to_pyarrow_listarray(data)
185 else:
186 trying_cast_to_python_objects = True
File /opt/conda/lib/python3.10/site-packages/datasets/features/features.py:1174, in list_of_np_array_to_pyarrow_listarray(l_arr, type)
1172 """Build a PyArrow ListArray from a possibly nested list of NumPy arrays"""
1173 if len(l_arr) > 0:
-> 1174 return list_of_pa_arrays_to_pyarrow_listarray(
1175 [numpy_to_pyarrow_listarray(arr, type=type) if arr is not None else None for arr in l_arr]
1176 )
1177 else:
1178 return pa.array([], type=type)
File /opt/conda/lib/python3.10/site-packages/datasets/features/features.py:1163, in list_of_pa_arrays_to_pyarrow_listarray(l_arr)
1160 null_indices = [i for i, arr in enumerate(l_arr) if arr is None]
1161 l_arr = [arr for arr in l_arr if arr is not None]
1162 offsets = np.cumsum(
-> 1163 [0] + [len(arr) for arr in l_arr], dtype=np.object
1164 ) # convert to dtype object to allow None insertion
1165 offsets = np.insert(offsets, null_indices, None)
1166 offsets = pa.array(offsets, type=pa.int32())
File /opt/conda/lib/python3.10/site-packages/numpy/__init__.py:324, in __getattr__(attr)
319 warnings.warn(
320 f"In the future `np.{attr}` will be defined as the "
321 "corresponding NumPy scalar.", FutureWarning, stacklevel=2)
323 if attr in __former_attrs__:
--> 324 raise AttributeError(__former_attrs__[attr])
326 if attr == 'testing':
327 import numpy.testing as testing
AttributeError: module 'numpy' has no attribute 'object'.
`np.object` was a deprecated alias for the builtin `object`. To avoid this error in existing code, use `object` by itself. Doing this will not modify any behavior and is safe.
The aliases was originally deprecated in NumPy 1.20; for more details and guidance see the original release note at:
https://numpy.org/devdocs/release/1.20.0-notes.html#deprecations
```
</details>
### Steps to reproduce the bug
Run above code in Kaggle Notebook.
### Expected behavior
I can resample audio data without fail.
### Environment info
- `datasets` version: 2.1.0
- Platform: Linux-5.15.133+-x86_64-with-glibc2.31
- Python version: 3.10.13
- PyArrow version: 11.0.0
- Pandas version: 2.2.1
|
[] |
{
"url": null,
"html_url": null,
"diff_url": null,
"patch_url": null,
"merged_at": null
}
| false
|
[
"Hi! You can fix this by updating the `datasets` package with `pip install -U datasets` and restarting the notebook.\r\n",
"Kaggle removed the problematic `datasets==2.1.0` pin last week, so I'm closing this issue (now it pre-installs the latest version)."
] |
6,782
|
Image cast_storage very slow for arrays (e.g. numpy, tensors)
|
Update: see comments below
### Describe the bug
Operations that save an image from a path are very slow.
I believe the reason for this is that the image data (`numpy`) is converted into `pyarrow` format but then back to python using `.pylist()` before being converted to a numpy array again.
`pylist` is already slow but used on a multi-dimensional numpy array such as an image it takes a very long time.
From the trace below we can see that `__arrow_array__` takes a long time.
It is currently also called in `get_inferred_type`, this should be removable #6781 but doesn't change the underyling issue.
The conversion to `pyarrow` and back also leads to the `numpy` array having type `int64` which causes a warning message because the image type excepts `uint8`.
However, originally the `numpy` image array was in `uint8`.
### Steps to reproduce the bug
```python
from PIL import Image
import numpy as np
import datasets
import cProfile
image = Image.fromarray(np.random.randint(0, 255, (2048, 2048, 3), dtype=np.uint8))
image.save("test_image.jpg")
ds = datasets.Dataset.from_dict(
{"image": ["test_image.jpg"]},
features=datasets.Features({"image": datasets.Image(decode=True)}),
)
# load as numpy array, e.g. for further processing with map
# same result as map returning numpy arrays
ds.set_format("numpy")
cProfile.run("ds.map(writer_batch_size=1, load_from_cache_file=False)", "restats")
```
```bash
Fri Apr 5 14:56:17 2024 restats
66817 function calls (64992 primitive calls) in 33.382 seconds
Ordered by: cumulative time
List reduced from 1073 to 20 due to restriction <20>
ncalls tottime percall cumtime percall filename:lineno(function)
46/1 0.000 0.000 33.382 33.382 {built-in method builtins.exec}
1 0.000 0.000 33.382 33.382 <string>:1(<module>)
1 0.000 0.000 33.382 33.382 arrow_dataset.py:594(wrapper)
1 0.000 0.000 33.382 33.382 arrow_dataset.py:551(wrapper)
1 0.000 0.000 33.379 33.379 arrow_dataset.py:2916(map)
4 0.000 0.000 33.327 8.332 arrow_dataset.py:3277(_map_single)
1 0.000 0.000 33.311 33.311 arrow_writer.py:465(write)
2 0.000 0.000 33.311 16.656 arrow_writer.py:423(write_examples_on_file)
1 0.000 0.000 33.311 33.311 arrow_writer.py:527(write_batch)
2 14.484 7.242 33.260 16.630 arrow_writer.py:161(__arrow_array__)
1 0.001 0.001 16.438 16.438 arrow_writer.py:121(get_inferred_type)
1 0.000 0.000 14.398 14.398 threading.py:637(wait)
1 0.000 0.000 14.398 14.398 threading.py:323(wait)
8 14.398 1.800 14.398 1.800 {method 'acquire' of '_thread.lock' objects}
4/2 0.000 0.000 4.337 2.169 table.py:1800(wrapper)
2 0.000 0.000 4.337 2.169 table.py:1950(cast_array_to_feature)
2 0.475 0.238 4.337 2.169 image.py:209(cast_storage)
9 2.583 0.287 2.583 0.287 {built-in method numpy.array}
2 0.000 0.000 1.284 0.642 image.py:319(encode_np_array)
2 0.000 0.000 1.246 0.623 image.py:301(image_to_bytes)
```
### Expected behavior
The `numpy` image data should be passed through as it will be directly consumed by `pillow` to convert it to bytes.
As an example one can replace `list_of_np_array_to_pyarrow_listarray(data)` in `__arrow_array__` with just `out = data` as a test.
We have to change `cast_storage` of the `Image` feature so it handles the passed through data (& if to handle type before)
```python
bytes_array = pa.array(
[encode_np_array(arr)["bytes"] if arr is not None else None for arr in storage],
type=pa.binary(),
)
```
Leading to the following:
```bash
Fri Apr 5 15:44:27 2024 restats
66419 function calls (64595 primitive calls) in 0.937 seconds
Ordered by: cumulative time
List reduced from 1023 to 20 due to restriction <20>
ncalls tottime percall cumtime percall filename:lineno(function)
47/1 0.000 0.000 0.935 0.935 {built-in method builtins.exec}
2/1 0.000 0.000 0.935 0.935 <string>:1(<module>)
2/1 0.000 0.000 0.934 0.934 arrow_dataset.py:594(wrapper)
2/1 0.000 0.000 0.934 0.934 arrow_dataset.py:551(wrapper)
2/1 0.000 0.000 0.934 0.934 arrow_dataset.py:2916(map)
4 0.000 0.000 0.933 0.233 arrow_dataset.py:3277(_map_single)
1 0.000 0.000 0.883 0.883 arrow_writer.py:466(write)
2 0.000 0.000 0.883 0.441 arrow_writer.py:424(write_examples_on_file)
1 0.000 0.000 0.882 0.882 arrow_writer.py:528(write_batch)
2 0.000 0.000 0.877 0.439 arrow_writer.py:161(__arrow_array__)
4/2 0.000 0.000 0.877 0.439 table.py:1800(wrapper)
2 0.000 0.000 0.877 0.439 table.py:1950(cast_array_to_feature)
2 0.009 0.005 0.877 0.439 image.py:209(cast_storage)
2 0.000 0.000 0.868 0.434 image.py:335(encode_np_array)
2 0.000 0.000 0.856 0.428 image.py:317(image_to_bytes)
2 0.000 0.000 0.822 0.411 Image.py:2376(save)
2 0.000 0.000 0.822 0.411 PngImagePlugin.py:1233(_save)
2 0.000 0.000 0.822 0.411 ImageFile.py:517(_save)
2 0.000 0.000 0.821 0.411 ImageFile.py:545(_encode_tile)
589 0.803 0.001 0.803 0.001 {method 'encode' of 'ImagingEncoder' objects}
```
This is of course only a test as it passes through all `numpy` arrays irrespective of if they should be an image.
Also I guess `cast_storage` is meant for casting `pyarrow` storage exclusively.
Converting to `pyarrow` array seems like a good solution as it also handles `pytorch` tensors etc., maybe there is a more efficient way to create a PIL image from a `pyarrow` array?
Not sure how this should be handled but I would be happy to help if there is a good solution.
### Environment info
- `datasets` version: 2.18.1.dev0
- Platform: Linux-6.7.11-200.fc39.x86_64-x86_64-with-glibc2.38
- Python version: 3.12.2
- `huggingface_hub` version: 0.22.2
- PyArrow version: 15.0.2
- Pandas version: 2.2.1
- `fsspec` version: 2024.3.1
|
[] |
{
"url": null,
"html_url": null,
"diff_url": null,
"patch_url": null,
"merged_at": null
}
| false
|
[
"This may be a solution that only changes `cast_storage` of `Image`.\r\nHowever, I'm not totally sure that the assumptions hold that are made about the `ListArray`.\r\n\r\n```python\r\nelif pa.types.is_list(storage.type):\r\n from .features import Array3DExtensionType\r\n\r\n def get_shapes(arr):\r\n shape = ()\r\n while isinstance(arr, pa.ListArray):\r\n len_curr = len(arr)\r\n arr = arr.flatten()\r\n len_new = len(arr)\r\n shape = shape + (len_new // len_curr,)\r\n return shape\r\n\r\n def get_dtypes(arr):\r\n dtype = storage.type\r\n while hasattr(dtype, \"value_type\"):\r\n dtype = dtype.value_type\r\n return dtype\r\n\r\n arrays = []\r\n for i, is_null in enumerate(storage.is_null()):\r\n if not is_null.as_py():\r\n storage_part = storage.take([i])\r\n shape = get_shapes(storage_part)\r\n dtype = get_dtypes(storage_part)\r\n\r\n extension_type = Array3DExtensionType(shape=shape, dtype=str(dtype))\r\n array = pa.ExtensionArray.from_storage(extension_type, storage_part)\r\n arrays.append(array.to_numpy().squeeze(0))\r\n else:\r\n arrays.append(None)\r\n\r\n bytes_array = pa.array(\r\n [encode_np_array(arr)[\"bytes\"] if arr is not None else None for arr in arrays],\r\n type=pa.binary(),\r\n )\r\n path_array = pa.array([None] * len(storage), type=pa.string())\r\n storage = pa.StructArray.from_arrays(\r\n [bytes_array, path_array], [\"bytes\", \"path\"], mask=bytes_array.is_null()\r\n )\r\n```\r\n(Edited): to handle nulls\r\n\r\nNotably this doesn't change anything about the passing through of data or other things, just in the `Image` class.\r\nSeems quite fast:\r\n```bash\r\nFri Apr 5 17:55:51 2024 restats\r\n\r\n 63818 function calls (61995 primitive calls) in 0.812 seconds\r\n\r\n Ordered by: cumulative time\r\n List reduced from 1051 to 20 due to restriction <20>\r\n\r\n ncalls tottime percall cumtime percall filename:lineno(function)\r\n 47/1 0.000 0.000 0.810 0.810 {built-in method builtins.exec}\r\n 2/1 0.000 0.000 0.810 0.810 <string>:1(<module>)\r\n 2/1 0.000 0.000 0.809 0.809 arrow_dataset.py:594(wrapper)\r\n 2/1 0.000 0.000 0.809 0.809 arrow_dataset.py:551(wrapper)\r\n 2/1 0.000 0.000 0.809 0.809 arrow_dataset.py:2916(map)\r\n 3 0.000 0.000 0.807 0.269 arrow_dataset.py:3277(_map_single)\r\n 1 0.000 0.000 0.760 0.760 arrow_writer.py:589(finalize)\r\n 1 0.000 0.000 0.760 0.760 arrow_writer.py:423(write_examples_on_file)\r\n 1 0.000 0.000 0.759 0.759 arrow_writer.py:527(write_batch)\r\n 1 0.001 0.001 0.754 0.754 arrow_writer.py:161(__arrow_array__)\r\n 2/1 0.000 0.000 0.719 0.719 table.py:1800(wrapper)\r\n 1 0.000 0.000 0.719 0.719 table.py:1950(cast_array_to_feature)\r\n 1 0.006 0.006 0.718 0.718 image.py:209(cast_storage)\r\n 1 0.000 0.000 0.451 0.451 image.py:361(encode_np_array)\r\n 1 0.000 0.000 0.444 0.444 image.py:343(image_to_bytes)\r\n 1 0.000 0.000 0.413 0.413 Image.py:2376(save)\r\n 1 0.000 0.000 0.413 0.413 PngImagePlugin.py:1233(_save)\r\n 1 0.000 0.000 0.413 0.413 ImageFile.py:517(_save)\r\n 1 0.000 0.000 0.413 0.413 ImageFile.py:545(_encode_tile)\r\n 397 0.409 0.001 0.409 0.001 {method 'encode' of 'ImagingEncoder' objects}\r\n```",
"Also encounter this problem. Has been strugging with it for a long time...",
"This actually applies to all arrays (numpy or tensors like in torch), not only from external files.\r\n```python\r\nimport numpy as np\r\nimport datasets\r\n\r\nds = datasets.Dataset.from_dict(\r\n {\"image\": [np.random.randint(0, 255, (2048, 2048, 3), dtype=np.uint8)]},\r\n features=datasets.Features({\"image\": datasets.Image(decode=True)}),\r\n)\r\nds.set_format(\"numpy\")\r\n\r\nds = ds.map(load_from_cache_file=False)\r\n```"
] |
6,781
|
Remove get_inferred_type from ArrowWriter write_batch
|
Inferring the type seems to be unnecessary given that the pyarrow array has already been created.
Because pyarrow array creation is sometimes extremely slow this doubles the time write_batch takes.
|
[] |
{
"url": "https://api.github.com/repos/huggingface/datasets/pulls/6781",
"html_url": "https://github.com/huggingface/datasets/pull/6781",
"diff_url": "https://github.com/huggingface/datasets/pull/6781.diff",
"patch_url": "https://github.com/huggingface/datasets/pull/6781.patch",
"merged_at": null
}
| true
|
[
"The docs for this PR live [here](https://moon-ci-docs.huggingface.co/docs/datasets/pr_6781). All of your documentation changes will be reflected on that endpoint. The docs are available until 30 days after the last update.",
"Close in favor of #6786."
] |
6,780
|
Fix CI
|
Updates the `wmt_t2t` test to pin the `revision` to the version with a loading script (cc @albertvillanova).
Additionally, it replaces the occurrences of the `lhoestq/test` repo id with `hf-internal-testing/dataset_with_script` and re-enables logging checks in the `Dataset.from_sql` tests.
|
[] |
{
"url": "https://api.github.com/repos/huggingface/datasets/pulls/6780",
"html_url": "https://github.com/huggingface/datasets/pull/6780",
"diff_url": "https://github.com/huggingface/datasets/pull/6780.diff",
"patch_url": "https://github.com/huggingface/datasets/pull/6780.patch",
"merged_at": "2024-04-04T18:23:34"
}
| true
|
[
"The docs for this PR live [here](https://moon-ci-docs.huggingface.co/docs/datasets/pr_6780). All of your documentation changes will be reflected on that endpoint. The docs are available until 30 days after the last update.",
"<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) | 0.005074 / 0.011353 (-0.006279) | 0.003395 / 0.011008 (-0.007614) | 0.062358 / 0.038508 (0.023849) | 0.031041 / 0.023109 (0.007932) | 0.244039 / 0.275898 (-0.031859) | 0.266361 / 0.323480 (-0.057119) | 0.003201 / 0.007986 (-0.004785) | 0.002609 / 0.004328 (-0.001719) | 0.049269 / 0.004250 (0.045018) | 0.045713 / 0.037052 (0.008661) | 0.264075 / 0.258489 (0.005586) | 0.295428 / 0.293841 (0.001587) | 0.027882 / 0.128546 (-0.100664) | 0.010424 / 0.075646 (-0.065222) | 0.208417 / 0.419271 (-0.210854) | 0.035728 / 0.043533 (-0.007805) | 0.246803 / 0.255139 (-0.008336) | 0.267169 / 0.283200 (-0.016031) | 0.019797 / 0.141683 (-0.121885) | 1.163299 / 1.452155 (-0.288856) | 1.196118 / 1.492716 (-0.296599) |\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) | 0.106091 / 0.018006 (0.088085) | 0.303970 / 0.000490 (0.303480) | 0.000219 / 0.000200 (0.000019) | 0.000042 / 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) | 0.017955 / 0.037411 (-0.019456) | 0.060539 / 0.014526 (0.046013) | 0.072884 / 0.176557 (-0.103673) | 0.119205 / 0.737135 (-0.617931) | 0.074072 / 0.296338 (-0.222266) |\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) | 0.272676 / 0.215209 (0.057467) | 2.715169 / 2.077655 (0.637514) | 1.419090 / 1.504120 (-0.085030) | 1.303903 / 1.541195 (-0.237292) | 1.311903 / 1.468490 (-0.156587) | 0.562005 / 4.584777 (-4.022772) | 2.432817 / 3.745712 (-1.312896) | 2.770599 / 5.269862 (-2.499263) | 1.723043 / 4.565676 (-2.842633) | 0.064341 / 0.424275 (-0.359934) | 0.004923 / 0.007607 (-0.002684) | 0.330507 / 0.226044 (0.104463) | 3.240829 / 2.268929 (0.971901) | 1.787638 / 55.444624 (-53.656986) | 1.522971 / 6.876477 (-5.353506) | 1.529496 / 2.142072 (-0.612576) | 0.645768 / 4.805227 (-4.159459) | 0.116405 / 6.500664 (-6.384259) | 0.041524 / 0.075469 (-0.033945) |\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) | 0.968515 / 1.841788 (-0.873272) | 11.628911 / 8.074308 (3.554603) | 9.495023 / 10.191392 (-0.696369) | 0.142219 / 0.680424 (-0.538204) | 0.013859 / 0.534201 (-0.520342) | 0.285727 / 0.579283 (-0.293556) | 0.276842 / 0.434364 (-0.157522) | 0.321247 / 0.540337 (-0.219090) | 0.409958 / 1.386936 (-0.976978) |\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) | 0.005102 / 0.011353 (-0.006251) | 0.003213 / 0.011008 (-0.007796) | 0.049250 / 0.038508 (0.010742) | 0.030649 / 0.023109 (0.007540) | 0.276629 / 0.275898 (0.000731) | 0.297315 / 0.323480 (-0.026165) | 0.004198 / 0.007986 (-0.003787) | 0.002744 / 0.004328 (-0.001585) | 0.047899 / 0.004250 (0.043649) | 0.040596 / 0.037052 (0.003544) | 0.287248 / 0.258489 (0.028759) | 0.313573 / 0.293841 (0.019732) | 0.029067 / 0.128546 (-0.099480) | 0.010122 / 0.075646 (-0.065524) | 0.058869 / 0.419271 (-0.360402) | 0.033012 / 0.043533 (-0.010521) | 0.272995 / 0.255139 (0.017856) | 0.297102 / 0.283200 (0.013903) | 0.018209 / 0.141683 (-0.123474) | 1.157785 / 1.452155 (-0.294369) | 1.184999 / 1.492716 (-0.307717) |\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) | 0.094228 / 0.018006 (0.076221) | 0.302055 / 0.000490 (0.301565) | 0.000221 / 0.000200 (0.000021) | 0.000044 / 0.000054 (-0.000010) |\n\n### Benchmark: benchmark_indices_mapping.json\n\n| metric | select | shard | shuffle | sort | train_test_split |\n|--------|---|---|---|---|---|\n| new / old (diff) | 0.022020 / 0.037411 (-0.015391) | 0.074970 / 0.014526 (0.060444) | 0.087682 / 0.176557 (-0.088875) | 0.126506 / 0.737135 (-0.610629) | 0.092046 / 0.296338 (-0.204293) |\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) | 0.295634 / 0.215209 (0.080425) | 2.891554 / 2.077655 (0.813899) | 1.579963 / 1.504120 (0.075843) | 1.462924 / 1.541195 (-0.078271) | 1.463806 / 1.468490 (-0.004684) | 0.558371 / 4.584777 (-4.026406) | 2.513500 / 3.745712 (-1.232212) | 2.754146 / 5.269862 (-2.515716) | 1.762317 / 4.565676 (-2.803360) | 0.063965 / 0.424275 (-0.360310) | 0.005538 / 0.007607 (-0.002069) | 0.348114 / 0.226044 (0.122070) | 3.484558 / 2.268929 (1.215630) | 1.940002 / 55.444624 (-53.504623) | 1.658469 / 6.876477 (-5.218008) | 1.645777 / 2.142072 (-0.496295) | 0.639367 / 4.805227 (-4.165861) | 0.115605 / 6.500664 (-6.385059) | 0.040647 / 0.075469 (-0.034822) |\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) | 1.036002 / 1.841788 (-0.805786) | 12.286895 / 8.074308 (4.212587) | 10.146719 / 10.191392 (-0.044673) | 0.140867 / 0.680424 (-0.539557) | 0.015517 / 0.534201 (-0.518684) | 0.290126 / 0.579283 (-0.289157) | 0.298702 / 0.434364 (-0.135662) | 0.325518 / 0.540337 (-0.214819) | 0.412597 / 1.386936 (-0.974339) |\n\n</details>\n</details>\n\n![](https://cml.dev/watermark.png#c3ddb1ef00334a6f973679a51e783905fbc9ef0b \"CML watermark\")\n"
] |
6,779
|
Install dependencies with `uv` in CI
|
`diffusers` (https://github.com/huggingface/diffusers/pull/7116) and `huggingface_hub` (https://github.com/huggingface/huggingface_hub/pull/2072) also use `uv` to install their dependencies, so we can do the same here.
It seems to make the "Install dependencies" step in the `ubuntu` jobs 5-8x faster and 1.5-2x in the `windows` one.
Besides introducing `uv` in CI, this PR bumps the `tensorflow` minimal version requirement to align with Transformers and simplifies the SpaCy hashing tests (use blank language models instead of the pre-trained ones)
|
[] |
{
"url": "https://api.github.com/repos/huggingface/datasets/pulls/6779",
"html_url": "https://github.com/huggingface/datasets/pull/6779",
"diff_url": "https://github.com/huggingface/datasets/pull/6779.diff",
"patch_url": "https://github.com/huggingface/datasets/pull/6779.patch",
"merged_at": "2024-04-08T13:27:43"
}
| true
|
[
"The docs for this PR live [here](https://moon-ci-docs.huggingface.co/docs/datasets/pr_6779). All of your documentation changes will be reflected on that endpoint. The docs are available until 30 days after the last update.",
"<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) | 0.005336 / 0.011353 (-0.006017) | 0.004052 / 0.011008 (-0.006956) | 0.063475 / 0.038508 (0.024967) | 0.032963 / 0.023109 (0.009854) | 0.243906 / 0.275898 (-0.031992) | 0.269048 / 0.323480 (-0.054432) | 0.003363 / 0.007986 (-0.004622) | 0.002802 / 0.004328 (-0.001527) | 0.049487 / 0.004250 (0.045236) | 0.046990 / 0.037052 (0.009938) | 0.260169 / 0.258489 (0.001680) | 0.289145 / 0.293841 (-0.004696) | 0.028030 / 0.128546 (-0.100517) | 0.010706 / 0.075646 (-0.064940) | 0.213640 / 0.419271 (-0.205632) | 0.035866 / 0.043533 (-0.007667) | 0.245106 / 0.255139 (-0.010033) | 0.269588 / 0.283200 (-0.013612) | 0.019791 / 0.141683 (-0.121892) | 1.117684 / 1.452155 (-0.334470) | 1.183389 / 1.492716 (-0.309327) |\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) | 0.095736 / 0.018006 (0.077730) | 0.302586 / 0.000490 (0.302097) | 0.000220 / 0.000200 (0.000020) | 0.000051 / 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) | 0.018985 / 0.037411 (-0.018426) | 0.062097 / 0.014526 (0.047571) | 0.075617 / 0.176557 (-0.100939) | 0.120570 / 0.737135 (-0.616566) | 0.075949 / 0.296338 (-0.220390) |\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) | 0.279597 / 0.215209 (0.064388) | 2.754319 / 2.077655 (0.676665) | 1.444147 / 1.504120 (-0.059973) | 1.328414 / 1.541195 (-0.212781) | 1.371073 / 1.468490 (-0.097417) | 0.553851 / 4.584777 (-4.030926) | 2.351694 / 3.745712 (-1.394018) | 2.860771 / 5.269862 (-2.409091) | 1.749664 / 4.565676 (-2.816013) | 0.061736 / 0.424275 (-0.362539) | 0.005073 / 0.007607 (-0.002534) | 0.329974 / 0.226044 (0.103930) | 3.300487 / 2.268929 (1.031558) | 1.812809 / 55.444624 (-53.631815) | 1.559018 / 6.876477 (-5.317458) | 1.628664 / 2.142072 (-0.513408) | 0.635757 / 4.805227 (-4.169471) | 0.116468 / 6.500664 (-6.384196) | 0.042641 / 0.075469 (-0.032828) |\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) | 0.972048 / 1.841788 (-0.869740) | 11.952721 / 8.074308 (3.878412) | 9.754274 / 10.191392 (-0.437118) | 0.132026 / 0.680424 (-0.548398) | 0.015352 / 0.534201 (-0.518849) | 0.290574 / 0.579283 (-0.288709) | 0.275384 / 0.434364 (-0.158980) | 0.330688 / 0.540337 (-0.209650) | 0.414868 / 1.386936 (-0.972068) |\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) | 0.005412 / 0.011353 (-0.005941) | 0.003814 / 0.011008 (-0.007194) | 0.049988 / 0.038508 (0.011480) | 0.031617 / 0.023109 (0.008507) | 0.278975 / 0.275898 (0.003077) | 0.303540 / 0.323480 (-0.019940) | 0.004265 / 0.007986 (-0.003721) | 0.002804 / 0.004328 (-0.001525) | 0.049518 / 0.004250 (0.045268) | 0.041176 / 0.037052 (0.004123) | 0.291248 / 0.258489 (0.032759) | 0.317401 / 0.293841 (0.023560) | 0.029501 / 0.128546 (-0.099045) | 0.010392 / 0.075646 (-0.065255) | 0.057906 / 0.419271 (-0.361365) | 0.033056 / 0.043533 (-0.010477) | 0.280202 / 0.255139 (0.025063) | 0.298684 / 0.283200 (0.015484) | 0.018071 / 0.141683 (-0.123612) | 1.167691 / 1.452155 (-0.284464) | 1.211322 / 1.492716 (-0.281394) |\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) | 0.092325 / 0.018006 (0.074318) | 0.301209 / 0.000490 (0.300719) | 0.000221 / 0.000200 (0.000021) | 0.000043 / 0.000054 (-0.000011) |\n\n### Benchmark: benchmark_indices_mapping.json\n\n| metric | select | shard | shuffle | sort | train_test_split |\n|--------|---|---|---|---|---|\n| new / old (diff) | 0.021432 / 0.037411 (-0.015980) | 0.074556 / 0.014526 (0.060031) | 0.086049 / 0.176557 (-0.090508) | 0.125151 / 0.737135 (-0.611984) | 0.088279 / 0.296338 (-0.208059) |\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) | 0.296755 / 0.215209 (0.081546) | 2.922650 / 2.077655 (0.844995) | 1.606031 / 1.504120 (0.101911) | 1.489692 / 1.541195 (-0.051502) | 1.530206 / 1.468490 (0.061716) | 0.577827 / 4.584777 (-4.006950) | 2.459716 / 3.745712 (-1.285997) | 2.825192 / 5.269862 (-2.444669) | 1.788110 / 4.565676 (-2.777566) | 0.064011 / 0.424275 (-0.360264) | 0.005616 / 0.007607 (-0.001991) | 0.341612 / 0.226044 (0.115568) | 3.455123 / 2.268929 (1.186194) | 1.961635 / 55.444624 (-53.482990) | 1.688107 / 6.876477 (-5.188370) | 1.725490 / 2.142072 (-0.416583) | 0.656011 / 4.805227 (-4.149216) | 0.117633 / 6.500664 (-6.383031) | 0.041386 / 0.075469 (-0.034083) |\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) | 1.025786 / 1.841788 (-0.816002) | 12.294598 / 8.074308 (4.220290) | 10.241136 / 10.191392 (0.049744) | 0.130577 / 0.680424 (-0.549847) | 0.016094 / 0.534201 (-0.518107) | 0.291193 / 0.579283 (-0.288090) | 0.273016 / 0.434364 (-0.161348) | 0.327553 / 0.540337 (-0.212784) | 0.418556 / 1.386936 (-0.968380) |\n\n</details>\n</details>\n\n![](https://cml.dev/watermark.png#3575036af2fd5cccff7fa60de30e2e444cf8a54e \"CML watermark\")\n"
] |
6,778
|
Dataset.to_csv() missing commas in columns with lists
|
### Describe the bug
The `to_csv()` method does not output commas in lists. So when the Dataset is loaded back in the data structure of the column with a list is not correct.
Here's an example:
Obviously, it's not as trivial as inserting commas in the list, since its a comma-separated file. But hopefully there's a way to export the list in a way that it'll be imported by `load_dataset()` correctly.
### Steps to reproduce the bug
Here's some code to reproduce the bug:
```python
from datasets import Dataset
ds = Dataset.from_dict(
{
"pokemon": ["bulbasaur", "squirtle"],
"type": ["grass", "water"]
}
)
def ascii_to_hex(text):
return [ord(c) for c in text]
ds = ds.map(lambda x: {"int": ascii_to_hex(x['pokemon'])})
ds.to_csv('../output/temp.csv')
```
temp.csv then contains:
```
### Expected behavior
ACTUAL OUTPUT:
```
pokemon,type,int
bulbasaur,grass,[ 98 117 108 98 97 115 97 117 114]
squirtle,water,[115 113 117 105 114 116 108 101]
```
EXPECTED OUTPUT:
```
pokemon,type,int
bulbasaur,grass,[98, 117, 108, 98, 97, 115, 97, 117, 114]
squirtle,water,[115, 113, 117, 105, 114, 116, 108, 101]
```
or probably something more like this since it's a CSV file:
```
pokemon,type,int
bulbasaur,grass,"[98, 117, 108, 98, 97, 115, 97, 117, 114]"
squirtle,water,"[115, 113, 117, 105, 114, 116, 108, 101]"
```
### Environment info
### Package Version
Name: datasets
Version: 2.16.1
### Python
version: 3.10.12
### OS Info
PRETTY_NAME="Ubuntu 22.04.4 LTS"
NAME="Ubuntu"
VERSION_ID="22.04"
VERSION="22.04.4 LTS (Jammy Jellyfish)"
VERSION_CODENAME=jammy
ID=ubuntu
ID_LIKE=debian
...
UBUNTU_CODENAME=jammy
|
[] |
{
"url": null,
"html_url": null,
"diff_url": null,
"patch_url": null,
"merged_at": null
}
| false
|
[
"Hello!\r\n\r\nThis is due to how pandas write numpy arrays to csv. [Source](https://stackoverflow.com/questions/54753179/to-csv-saves-np-array-as-string-instead-of-as-a-list)\r\nTo fix this, you can convert them to list yourselves.\r\n\r\n```python\r\ndf = ds.to_pandas()\r\ndf['int'] = df['int'].apply(lambda arr: list(arr))\r\ndf.to_csv(index=False, '../output/temp.csv')\r\n```\r\n\r\nI think it would be good if `datasets` would do the conversion itself, but it's a breaking change and I would wait for the greenlight from someone from HF."
] |
6,777
|
.Jsonl metadata not detected
|
### Describe the bug
Hi I have the following directory structure:
|--dataset
| |-- images
| |-- metadata1000.csv
| |-- metadata1000.jsonl
| |-- padded_images
Example of metadata1000.jsonl file
{"caption": "a drawing depicts a full shot of a black t-shirt with a triangular pattern on the front there is a white label on the left side of the triangle", "image": "images/212734.png", "gaussian_padded_image": "padded_images/p_212734.png"}
{"caption": "an eye-level full shot of a large elephant and a baby elephant standing in a watering hole on the left side is a small elephant with its head turned to the right of dry land, trees, and bushes", "image": "images/212735.png", "gaussian_padded_image": "padded_images/p_212735.png"}
.
.
.
I'm trying to use dataset = load_dataset("imagefolder", data_dir='/dataset/', split='train') to load the the dataset, however it is not able to load according to the fields in the metadata1000.jsonl .
please assist to load the data properly
also getting
```
File "/workspace/train_trans_vae.py", line 1089, in <module>
print(get_metadata_patterns('/dataset/'))
File "/opt/conda/lib/python3.10/site-packages/datasets/data_files.py", line 499, in get_metadata_patterns
raise FileNotFoundError(f"The directory at {base_path} doesn't contain any metadata file") from None
FileNotFoundError: The directory at /dataset/ doesn't contain any metadata file
```
when trying
```
from datasets.data_files import get_metadata_patterns
print(get_metadata_patterns('/dataset/'))
```
### Steps to reproduce the bug
dataset Version: 2.18.0
make a similar jsonl and similar directory format
### Expected behavior
creates a dataset object with the column names, caption,image,gaussian_padded_image
### Environment info
dataset Version: 2.18.0
|
[] |
{
"url": null,
"html_url": null,
"diff_url": null,
"patch_url": null,
"merged_at": null
}
| false
|
[
"Hi! `metadata.jsonl` (or `metadata.csv`) is the only allowed name for the `imagefolder`'s metadata files.",
"@mariosasko hey i tried with metadata.jsonl also and it still doesn't get the right columns",
"@mariosasko it says metadata.csv not found\r\n<img width=\"1150\" alt=\"image\" src=\"https://github.com/huggingface/datasets/assets/81643693/3754980c-6185-4413-88fa-b499bcdd4195\">\r\n\r\ndataset = load_dataset('/dataset',metadata.csv) \r\n\r\n| workspace\r\n|| source code\r\n| dataset\r\n| |-- images\r\n| |-- metadata.csv\r\n| |-- metadata.jsonl\r\n| |-- padded_images\r\n\r\nExample of metadata.jsonl file\r\n{\"caption\": \"a drawing depicts a full shot of a black t-shirt with a triangular pattern on the front there is a white label on the left side of the triangle\", \"image\": \"images/212734.png\", \"gaussian_padded_image\": \"padded_images/p_212734.png\"}\r\n{\"caption\": \"an eye-level full shot of a large elephant and a baby elephant standing in a watering hole on the left side is a small elephant with its head turned to the right of dry land, trees, and bushes\", \"image\": \"images/212735.png\", \"gaussian_padded_image\": \"padded_images/p_212735.png\"}\r\n",
"Loading more than one image per row with `imagefolder` is not supported currently. You can subscribe to https://github.com/huggingface/datasets/issues/5760 to see when it will be.\r\n\r\nInstead, you can load the dataset with `Dataset.from_generator`:\r\n```python\r\nimport json\r\nfrom datasets import Dataset, Value, Image, Features\r\n\r\ndef gen():\r\n with open(\"./dataset/metadata.jsonl\") as f:\r\n for line in f:\r\n line = json.loads(line)\r\n yield {\"caption\": line[\"caption\"], \"image\": os.path.join(\"./dataset\", line[\"image\"], \"gaussian_padded_image\": os.path.join(\"./dataset\", line[\"gaussian_padded_image\"]))}\r\n\r\nfeatures = Features({\"caption\": Value(\"string\"), \"image\": Image(), \"gaussian_padded_image\": Image()})\r\ndataset = Dataset.from_generator(gen, features=features)\r\n```\r\n(E.g., if you want to share this dataset on the Hub, you can call `dataset.push_to_hub(...)` afterward)",
"hi Thanks for sharing this, Actually I was trying with a webdataset format of the data as well and it did'nt work. Could you share how i can create Dataset object from webdataset format of this data?"
] |
6,775
|
IndexError: Invalid key: 0 is out of bounds for size 0
|
### Describe the bug
I am trying to fine-tune llama2-7b model in GCP. The notebook I am using for this can be found [here](https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/community/model_garden/model_garden_pytorch_llama2_peft_finetuning.ipynb).
When I use the dataset given in the example, the training gets successfully completed (example dataset can be found [here](https://huggingface.co/datasets/timdettmers/openassistant-guanaco)).
However when I use my own dataset which is in the same format as the example dataset, I get the below error (my dataset can be found [here](https://huggingface.co/datasets/kk2491/finetune_dataset_002)).
I see the files are being read correctly from the logs:
### Steps to reproduce the bug
1. Clone the [vertex-ai-samples](https://github.com/GoogleCloudPlatform/vertex-ai-samples) repository.
2. Run the [llama2-7b peft fine-tuning](https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/community/model_garden/model_garden_pytorch_llama2_peft_finetuning.ipynb).
3. Change the dataset `kk2491/finetune_dataset_002`
### Expected behavior
The training should complete successfully, and model gets deployed to an endpoint.
### Environment info
Python version : Python 3.10.12
Dataset : https://huggingface.co/datasets/kk2491/finetune_dataset_002
|
[] |
{
"url": null,
"html_url": null,
"diff_url": null,
"patch_url": null,
"merged_at": null
}
| false
|
[
"Same problem.",
"Hi! You should be able to fix this by passing `remove_unused_columns=False` to the `transformers` `TrainingArguments` as explained in https://github.com/huggingface/peft/issues/1299.\r\n\r\n(I'm not familiar with Vertex AI, but I'd assume `remove_unused_columns` can be passed as a flag to the docker container) ",
"I had the same problem, but I spent a whole day trying different combination with my own dataset with the example data set and found the reason: the example data is multi-turn conversation between human and assistant, so # Humman or # Assistant appear at least twice. If your own custom data only has single turn conversation, it might end up with the same error. What you can do is repeat your single turn conversation twice in your training data (keep the key 'text' the same) and maybe it works. I guess the reason is the specific way processing the data requires and counts multi-turn only (single turn will be discarded so it ends up with no training data), but since I am using Google Vertex AI, I don't have direct access to the underlying code so that was just my guess. ",
"> Hi! You should be able to fix this by passing `remove_unused_columns=False` to the `transformers` `TrainingArguments` as explained in [huggingface/peft#1299](https://github.com/huggingface/peft/issues/1299).\r\n> \r\n> (I'm not familiar with Vertex AI, but I'd assume `remove_unused_columns` can be passed as a flag to the docker container)\r\n\r\n@mariosasko Thanks for the response and suggestion. \r\nWhen I set `remove_unused_columns` as `False` , I end up getting different error (will post the error soon). \r\nEither the Vertex-AI does not support `remove_unused_columns` or my dataset is completely wrong. \r\n\r\nThank you, \r\nKK",
"> I had the same problem, but I spent a whole day trying different combination with my own dataset with the example data set and found the reason: the example data is multi-turn conversation between human and assistant, so # Humman or # Assistant appear at least twice. If your own custom data only has single turn conversation, it might end up with the same error. What you can do is repeat your single turn conversation twice in your training data (keep the key 'text' the same) and maybe it works. I guess the reason is the specific way processing the data requires and counts multi-turn only (single turn will be discarded so it ends up with no training data), but since I am using Google Vertex AI, I don't have direct access to the underlying code so that was just my guess.\r\n\r\n@cyberyu Thanks for your suggestions. \r\nI have tried the approach you suggested, copied the same conversation in each jsonl element so every jsonl item has 2 `HUMAN` and `ASSISTANT`. \r\nHowever in my case, the issue persists. I am gonna give few more tries, and post the results here. \r\nYou can find my dataset [here](https://huggingface.co/datasets/kk2491/test/tree/main) \r\n\r\nThank you, \r\nKK ",
"> > I had the same problem, but I spent a whole day trying different combination with my own dataset with the example data set and found the reason: the example data is multi-turn conversation between human and assistant, so # Humman or # Assistant appear at least twice. If your own custom data only has single turn conversation, it might end up with the same error. What you can do is repeat your single turn conversation twice in your training data (keep the key 'text' the same) and maybe it works. I guess the reason is the specific way processing the data requires and counts multi-turn only (single turn will be discarded so it ends up with no training data), but since I am using Google Vertex AI, I don't have direct access to the underlying code so that was just my guess.\r\n> \r\n> @cyberyu Thanks for your suggestions. I have tried the approach you suggested, copied the same conversation in each jsonl element so every jsonl item has 2 `HUMAN` and `ASSISTANT`. However in my case, the issue persists. I am gonna give few more tries, and post the results here. You can find my dataset [here](https://huggingface.co/datasets/kk2491/test/tree/main)\r\n> \r\n> Thank you, KK\r\n\r\nI think another reason is your training sample length is too short. I saw a relevant report (https://discuss.huggingface.co/t/indexerror-invalid-key-16-is-out-of-bounds-for-size-0/14298/16) stating that the processing code might have a bug discarding sequence length short than max_seq_length, which is 512. Not sure the Vertex AI backend code has fixed that bug or not. So I tried to add some garbage content in your data, and extended the length longer than 512 for a single turn, and repeated twice. You can copy the following line as 5 repeated lines as your training data jsonl file of five samples (no eval or test needed, for speed up, set evaluation step to 5 and training step to 10,), and it will pass.\r\n\r\n{\"text\":\"### Human: You are a helpful AI Assistant familiar with customer service. You are a helpful AI Assistant familiar with customer service. You are a helpful AI Assistant familiar with customer service. You are a helpful AI Assistant familiar with customer service. You are a helpful AI Assistant familiar with customer service. You are a helpful AI Assistant familiar with customer service. You are a helpful AI Assistant familiar with customer service. You are a helpful AI Assistant familiar with customer service. You are a helpful AI Assistant familiar with customer service. You are a helpful AI Assistant familiar with customer service. You are a helpful AI Assistant familiar with customer service. You are a helpful AI Assistant familiar with customer service. You are a helpful AI Assistant familiar with customer service. You will handle customers queries and provide effective help message. Please provide response to 'Can Interplai software optimize routes for minimizing package handling and transfer times in distribution centers'? ### Assistant: Yes, Interplai software can optimize routes for distribution centers by streamlining package handling processes, minimizing transfer times between loading docks and storage areas, and optimizing warehouse layouts for efficient order fulfillment. ### Human: You are a helpful AI Assistant familiar with customer service. You are a helpful AI Assistant familiar with customer service. You are a helpful AI Assistant familiar with customer service. You are a helpful AI Assistant familiar with customer service. You are a helpful AI Assistant familiar with customer service. You are a helpful AI Assistant familiar with customer service. You are a helpful AI Assistant familiar with customer service. You are a helpful AI Assistant familiar with customer service. You are a helpful AI Assistant familiar with customer service. You are a helpful AI Assistant familiar with customer service. You are a helpful AI Assistant familiar with customer service. You are a helpful AI Assistant familiar with customer service. You are a helpful AI Assistant familiar with customer service. You will handle customers queries and provide effective help message. Please provide response to 'Can Interplai software optimize routes for minimizing package handling and transfer times in distribution centers'? ### Assistant: Yes, Interplai software can optimize routes for distribution centers by streamlining package handling processes, minimizing transfer times between loading docks and storage areas, and optimizing warehouse layouts for efficient order fulfillment.\"}\r\n",
"@cyberyu **Thank you so much, You saved my day (+ so many days)**. \r\nI tried the example you provided above, and the training is successfully completed in Vertex-AI (through GUI). \r\nI never thought there would be constraints on the length of the samples and also on the number of turns. \r\nI will update my complete dataset and see update here once the training is completed. \r\n\r\nThank you, \r\nKK "
] |
6,774
|
Generating split is very slow when Image format is PNG
|
### Describe the bug
When I create a dataset, it gets stuck while generating cached data.
The image format is PNG, and it will not get stuck when the image format is jpeg.
After debugging, I know that it is because of the `pa.array` operation in [arrow_writer](https://github.com/huggingface/datasets/blob/2.13.0/src/datasets/arrow_writer.py#L553), but i don't why.
### Steps to reproduce the bug
```
from datasets import Dataset
def generator(lines):
for line in lines:
img = Image.open(open(line["url"], "rb"))
# print(img.format) # "PNG"
yield {
"image": img,
}
lines = open(dataset_path, "r")
dataset = Dataset.from_generator(
generator,
gen_kwargs={"lines": lines}
)
```
### Expected behavior
Generating split done.
### Environment info
datasets 2.13.0
|
[] |
{
"url": null,
"html_url": null,
"diff_url": null,
"patch_url": null,
"merged_at": null
}
| false
|
[
"I think this is due to the speed of reading a `png` image using pillow compared to a `jpg` image.\r\nNotably the same is true with `tiff`, it is even faster than `jpg` in my case."
] |
6,773
|
Dataset on Hub re-downloads every time?
|
### Describe the bug
Hi, I have a dataset on the hub [here](https://huggingface.co/datasets/manestay/borderlines). It has 1k+ downloads, which I sure is mostly just me and my colleagues working with it. It should have far fewer, since I'm using the same machine with a properly set up HF_HOME variable. However, whenever I run the below function `load_borderlines_hf`, it downloads the entire dataset from the hub and then does the other logic:
https://github.com/manestay/borderlines/blob/4e161f444661e2ebfe643f3fe149d9258d63a57d/run_gpt/lib.py#L80
Let me know what I'm doing wrong here, or if it's a bug with the `datasets` library itself. On the hub I have my data stored in CSVs, but several columns are lists, so that's why I have the code to map splitting on `;`. I looked into dataset loading scripts, but it seemed difficult to set up. I have verified that other `datasets` and `models` on my system are using the cache properly (e.g. I have a 13B parameter model and large datasets, but those are cached and don't redownload).
__EDIT: __ as pointed out in the discussion below, it may be the `map()` calls that aren't being cached properly. Supposing the `load_dataset()` retrieve from the cache, then it should be the case that the `map()` calls also retrieve from the cached output. But the `map()` commands re-execute sometimes.
### Steps to reproduce the bug
1. Copy and paste the function from [here](https://github.com/manestay/borderlines/blob/4e161f444661e2ebfe643f3fe149d9258d63a57d/run_gpt/lib.py#L80) (lines 80-100)
2. Run it in Python `load_borderlines_hf(None)`
3. It completes successfully, downloading from HF hub, then doing the mapping logic etc.
4. If you run it again after some time, it will re-download, ignoring the cache
### Expected behavior
Re-running the code, which calls `datasets.load_dataset('manestay/borderlines', 'territories')`, should use the cached version
### Environment info
- `datasets` version: 2.16.1
- Platform: Linux-5.14.21-150500.55.7-default-x86_64-with-glibc2.31
- Python version: 3.10.13
- `huggingface_hub` version: 0.20.3
- PyArrow version: 15.0.0
- Pandas version: 1.5.3
- `fsspec` version: 2023.10.0
|
[] |
{
"url": null,
"html_url": null,
"diff_url": null,
"patch_url": null,
"merged_at": null
}
| false
|
[
"The caching works as expected when I try to reproduce this locally or on Colab...",
"hi @mariosasko , Thank you for checking. I also tried running this again just now, and it seems like the `load_dataset()` caches properly (though I'll double check later).\r\n\r\nI think the issue might be in the caching of the function output for `territories.map(lambda row: {'Claimants': row['Claimants'].split(';')})`. My current run re-ran this, even though I have run this many times before, and as demonstrated by loading from cache, the loaded dataset is the same.\r\n\r\nI wonder if the issue stems from using CSV output. Do you recommend changing to Parquet, and if so, is there an easy way to take the already uploaded data on the Hub and reformat?",
"This issue seems similar to https://github.com/huggingface/datasets/issues/6184 (`dill` serializes objects defined outside the `__main__` module by reference). You should be able to work around this limitation by defining the lambdas outside of `load_borderlines_hf` (as module variables) and then setting their `__module__` attribute's value to `None` to force serializing them by value, e.g., like this: \r\n```python\r\nsplit_Claimants_row = lambda row: {'Claimants': row['Claimants'].split(';')}\r\nsplit_Claimants_row.__module__ = None\r\n```",
"Thank you, I'll give this a try. Your fix makes sense to me, so this issue can be closed for now.\r\n\r\nUnrelated comment -- for \"Downloads last month\" on the hub page, I'm assuming for this project that each downloaded CSV is 1 download? The dataset consists of 51 CSVs, so I'm trying to see why it's incrementing so quickly (1125 2 days ago, 1246 right now).",
"This doc explains how we count \"Downloads last month\": https://huggingface.co/docs/hub/datasets-download-stats"
] |
6,772
|
`remove_columns`/`rename_columns` doc fixes
|
Use more consistent wording in `remove_columns` to explain why it's faster than `map` and update `remove_columns`/`rename_columns` docstrings to fix in-place calls.
Reported in https://github.com/huggingface/datasets/issues/6700
|
[] |
{
"url": "https://api.github.com/repos/huggingface/datasets/pulls/6772",
"html_url": "https://github.com/huggingface/datasets/pull/6772",
"diff_url": "https://github.com/huggingface/datasets/pull/6772.diff",
"patch_url": "https://github.com/huggingface/datasets/pull/6772.patch",
"merged_at": "2024-04-02T16:17:46"
}
| true
|
[
"The docs for this PR live [here](https://moon-ci-docs.huggingface.co/docs/datasets/pr_6772). All of your documentation changes will be reflected on that endpoint. The docs are available until 30 days after the last update.",
"<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) | 0.005728 / 0.011353 (-0.005624) | 0.003809 / 0.011008 (-0.007199) | 0.062930 / 0.038508 (0.024422) | 0.032320 / 0.023109 (0.009211) | 0.251072 / 0.275898 (-0.024826) | 0.275397 / 0.323480 (-0.048083) | 0.003314 / 0.007986 (-0.004671) | 0.002869 / 0.004328 (-0.001460) | 0.049070 / 0.004250 (0.044819) | 0.049282 / 0.037052 (0.012229) | 0.263546 / 0.258489 (0.005057) | 0.291471 / 0.293841 (-0.002370) | 0.028462 / 0.128546 (-0.100084) | 0.010528 / 0.075646 (-0.065119) | 0.211249 / 0.419271 (-0.208023) | 0.036840 / 0.043533 (-0.006693) | 0.250038 / 0.255139 (-0.005101) | 0.268883 / 0.283200 (-0.014317) | 0.021417 / 0.141683 (-0.120266) | 1.139754 / 1.452155 (-0.312400) | 1.197319 / 1.492716 (-0.295397) |\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) | 0.094191 / 0.018006 (0.076185) | 0.302413 / 0.000490 (0.301923) | 0.000220 / 0.000200 (0.000020) | 0.000048 / 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) | 0.018490 / 0.037411 (-0.018922) | 0.063361 / 0.014526 (0.048835) | 0.075854 / 0.176557 (-0.100702) | 0.121499 / 0.737135 (-0.615637) | 0.075982 / 0.296338 (-0.220356) |\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) | 0.286030 / 0.215209 (0.070821) | 2.778487 / 2.077655 (0.700832) | 1.440963 / 1.504120 (-0.063157) | 1.326217 / 1.541195 (-0.214977) | 1.359228 / 1.468490 (-0.109262) | 0.566999 / 4.584777 (-4.017778) | 2.453344 / 3.745712 (-1.292368) | 2.841448 / 5.269862 (-2.428413) | 1.825197 / 4.565676 (-2.740479) | 0.062301 / 0.424275 (-0.361974) | 0.004948 / 0.007607 (-0.002659) | 0.334578 / 0.226044 (0.108534) | 3.302327 / 2.268929 (1.033399) | 1.799808 / 55.444624 (-53.644817) | 1.529693 / 6.876477 (-5.346783) | 1.564684 / 2.142072 (-0.577389) | 0.632891 / 4.805227 (-4.172336) | 0.116594 / 6.500664 (-6.384070) | 0.042695 / 0.075469 (-0.032774) |\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) | 0.999994 / 1.841788 (-0.841794) | 12.767365 / 8.074308 (4.693057) | 10.550439 / 10.191392 (0.359047) | 0.133437 / 0.680424 (-0.546986) | 0.015252 / 0.534201 (-0.518949) | 0.293285 / 0.579283 (-0.285998) | 0.274773 / 0.434364 (-0.159590) | 0.328718 / 0.540337 (-0.211619) | 0.428021 / 1.386936 (-0.958915) |\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) | 0.005538 / 0.011353 (-0.005815) | 0.003738 / 0.011008 (-0.007271) | 0.050179 / 0.038508 (0.011671) | 0.032441 / 0.023109 (0.009332) | 0.294721 / 0.275898 (0.018823) | 0.322616 / 0.323480 (-0.000864) | 0.004255 / 0.007986 (-0.003731) | 0.002913 / 0.004328 (-0.001416) | 0.049044 / 0.004250 (0.044794) | 0.042361 / 0.037052 (0.005309) | 0.304162 / 0.258489 (0.045673) | 0.332757 / 0.293841 (0.038916) | 0.029355 / 0.128546 (-0.099191) | 0.010546 / 0.075646 (-0.065100) | 0.058213 / 0.419271 (-0.361058) | 0.032648 / 0.043533 (-0.010885) | 0.298241 / 0.255139 (0.043102) | 0.313710 / 0.283200 (0.030510) | 0.017836 / 0.141683 (-0.123847) | 1.135050 / 1.452155 (-0.317104) | 1.178277 / 1.492716 (-0.314439) |\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) | 0.094387 / 0.018006 (0.076381) | 0.301955 / 0.000490 (0.301466) | 0.000220 / 0.000200 (0.000020) | 0.000052 / 0.000054 (-0.000002) |\n\n### Benchmark: benchmark_indices_mapping.json\n\n| metric | select | shard | shuffle | sort | train_test_split |\n|--------|---|---|---|---|---|\n| new / old (diff) | 0.023135 / 0.037411 (-0.014276) | 0.078109 / 0.014526 (0.063583) | 0.087519 / 0.176557 (-0.089037) | 0.127815 / 0.737135 (-0.609320) | 0.090107 / 0.296338 (-0.206231) |\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) | 0.289149 / 0.215209 (0.073940) | 2.832354 / 2.077655 (0.754699) | 1.574003 / 1.504120 (0.069883) | 1.449190 / 1.541195 (-0.092005) | 1.465798 / 1.468490 (-0.002692) | 0.561953 / 4.584777 (-4.022824) | 2.445788 / 3.745712 (-1.299924) | 2.882453 / 5.269862 (-2.387409) | 1.813267 / 4.565676 (-2.752409) | 0.063163 / 0.424275 (-0.361112) | 0.005785 / 0.007607 (-0.001822) | 0.340125 / 0.226044 (0.114081) | 3.355370 / 2.268929 (1.086442) | 1.924226 / 55.444624 (-53.520398) | 1.643242 / 6.876477 (-5.233234) | 1.650149 / 2.142072 (-0.491924) | 0.654818 / 4.805227 (-4.150409) | 0.114968 / 6.500664 (-6.385696) | 0.042044 / 0.075469 (-0.033425) |\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) | 1.024867 / 1.841788 (-0.816921) | 12.656140 / 8.074308 (4.581832) | 10.927014 / 10.191392 (0.735622) | 0.155929 / 0.680424 (-0.524495) | 0.015356 / 0.534201 (-0.518845) | 0.289834 / 0.579283 (-0.289449) | 0.280889 / 0.434364 (-0.153475) | 0.331490 / 0.540337 (-0.208847) | 0.418037 / 1.386936 (-0.968899) |\n\n</details>\n</details>\n\n![](https://cml.dev/watermark.png#ad3467e9b138d1a9b87b661828a71139f4e46ece \"CML watermark\")\n"
] |
6,771
|
Datasets FileNotFoundError when trying to generate examples.
|
### Discussed in https://github.com/huggingface/datasets/discussions/6768
<div type='discussions-op-text'>
<sup>Originally posted by **RitchieP** April 1, 2024</sup>
Currently, I have a dataset hosted on Huggingface with a custom script [here](https://huggingface.co/datasets/RitchieP/VerbaLex_voice).
I'm loading my dataset as below.
```py
from datasets import load_dataset, IterableDatasetDict
dataset = IterableDatasetDict()
dataset["train"] = load_dataset("RitchieP/VerbaLex_voice", "ar", split="train", use_auth_token=True, streaming=True)
dataset["test"] = load_dataset("RitchieP/VerbaLex_voice", "ar", split="test", use_auth_token=True, streaming=True)
```
And when I try to see the data I have loaded with
```py
list(dataset["train"].take(1))
```
And it gives me this stack trace
```
---------------------------------------------------------------------------
FileNotFoundError Traceback (most recent call last)
Cell In[2], line 1
----> 1 list(dataset["train"].take(1))
File /opt/conda/lib/python3.10/site-packages/datasets/iterable_dataset.py:1388, in IterableDataset.__iter__(self)
1385 yield formatter.format_row(pa_table)
1386 return
-> 1388 for key, example in ex_iterable:
1389 if self.features:
1390 # `IterableDataset` automatically fills missing columns with None.
1391 # This is done with `_apply_feature_types_on_example`.
1392 example = _apply_feature_types_on_example(
1393 example, self.features, token_per_repo_id=self._token_per_repo_id
1394 )
File /opt/conda/lib/python3.10/site-packages/datasets/iterable_dataset.py:1044, in TakeExamplesIterable.__iter__(self)
1043 def __iter__(self):
-> 1044 yield from islice(self.ex_iterable, self.n)
File /opt/conda/lib/python3.10/site-packages/datasets/iterable_dataset.py:234, in ExamplesIterable.__iter__(self)
233 def __iter__(self):
--> 234 yield from self.generate_examples_fn(**self.kwargs)
File ~/.cache/huggingface/modules/datasets_modules/datasets/RitchieP--VerbaLex_voice/9465eaee58383cf9d7c3e14111d7abaea56398185a641b646897d6df4e4732f7/VerbaLex_voice.py:127, in VerbaLexVoiceDataset._generate_examples(self, local_extracted_archive_paths, archives, meta_path)
125 for i, audio_archive in enumerate(archives):
126 print(audio_archive)
--> 127 for path, file in audio_archive:
128 _, filename = os.path.split(path)
129 if filename in metadata:
File /opt/conda/lib/python3.10/site-packages/datasets/download/streaming_download_manager.py:869, in _IterableFromGenerator.__iter__(self)
868 def __iter__(self):
--> 869 yield from self.generator(*self.args, **self.kwargs)
File /opt/conda/lib/python3.10/site-packages/datasets/download/streaming_download_manager.py:919, in ArchiveIterable._iter_from_urlpath(cls, urlpath, download_config)
915 @classmethod
916 def _iter_from_urlpath(
917 cls, urlpath: str, download_config: Optional[DownloadConfig] = None
918 ) -> Generator[Tuple, None, None]:
--> 919 compression = _get_extraction_protocol(urlpath, download_config=download_config)
920 # Set block_size=0 to get faster streaming
921 # (e.g. for hf:// and https:// it uses streaming Requests file-like instances)
922 with xopen(urlpath, "rb", download_config=download_config, block_size=0) as f:
File /opt/conda/lib/python3.10/site-packages/datasets/download/streaming_download_manager.py:400, in _get_extraction_protocol(urlpath, download_config)
398 urlpath, storage_options = _prepare_path_and_storage_options(urlpath, download_config=download_config)
399 try:
--> 400 with fsspec.open(urlpath, **(storage_options or {})) as f:
401 return _get_extraction_protocol_with_magic_number(f)
402 except FileNotFoundError:
File /opt/conda/lib/python3.10/site-packages/fsspec/core.py:100, in OpenFile.__enter__(self)
97 def __enter__(self):
98 mode = self.mode.replace("t", "").replace("b", "") + "b"
--> 100 f = self.fs.open(self.path, mode=mode)
102 self.fobjects = [f]
104 if self.compression is not None:
File /opt/conda/lib/python3.10/site-packages/fsspec/spec.py:1307, in AbstractFileSystem.open(self, path, mode, block_size, cache_options, compression, **kwargs)
1305 else:
1306 ac = kwargs.pop("autocommit", not self._intrans)
-> 1307 f = self._open(
1308 path,
1309 mode=mode,
1310 block_size=block_size,
1311 autocommit=ac,
1312 cache_options=cache_options,
1313 **kwargs,
1314 )
1315 if compression is not None:
1316 from fsspec.compression import compr
File /opt/conda/lib/python3.10/site-packages/fsspec/implementations/local.py:180, in LocalFileSystem._open(self, path, mode, block_size, **kwargs)
178 if self.auto_mkdir and "w" in mode:
179 self.makedirs(self._parent(path), exist_ok=True)
--> 180 return LocalFileOpener(path, mode, fs=self, **kwargs)
File /opt/conda/lib/python3.10/site-packages/fsspec/implementations/local.py:302, in LocalFileOpener.__init__(self, path, mode, autocommit, fs, compression, **kwargs)
300 self.compression = get_compression(path, compression)
301 self.blocksize = io.DEFAULT_BUFFER_SIZE
--> 302 self._open()
File /opt/conda/lib/python3.10/site-packages/fsspec/implementations/local.py:307, in LocalFileOpener._open(self)
305 if self.f is None or self.f.closed:
306 if self.autocommit or "w" not in self.mode:
--> 307 self.f = open(self.path, mode=self.mode)
308 if self.compression:
309 compress = compr[self.compression]
FileNotFoundError: [Errno 2] No such file or directory: '/kaggle/working/h'
```
After looking into the stack trace, and referring to the source codes, it looks like its trying to access a directory in the notebook's environment and I don't understand why.
Not sure if its a bug in Datasets library, so I'm opening a discussions first. Feel free to ask for more information if needed. Appreciate any help in advance!</div>
Hi, referring to the discussion title above, after further digging, I think it's an issue within the datasets library. But not quite sure where it is.
If you require any more info or actions from me, please let me know. Appreciate any help in advance!
|
[] |
{
"url": null,
"html_url": null,
"diff_url": null,
"patch_url": null,
"merged_at": null
}
| false
|
[
"Hi! I've opened a PR in the repo to fix this issue: https://huggingface.co/datasets/RitchieP/VerbaLex_voice/discussions/6",
"@mariosasko Thanks for the PR and help! Guess I could close the issue for now. Appreciate the help!"
] |
6,770
|
[Bug Report] `datasets==2.18.0` is not compatible with `fsspec==2023.12.2`
|
### Describe the bug
`Datasets==2.18.0` is not compatible with `fsspec==2023.12.2`.
I have to downgrade fsspec to `fsspec==2023.10.0` to make `Datasets==2.18.0` work properly.
### Steps to reproduce the bug
To reproduce the bug:
1. Make sure that `Datasets==2.18.0` and `fsspec==2023.12.2`.
2. Run the following code:
```
from datasets import load_dataset
dataset = load_dataset("trec")
```
3. Then one will get the following error message:
```
Traceback (most recent call last):
File "<stdin>", line 1, in <module>
File "/opt/conda/lib/python3.10/site-packages/datasets/load.py", line 2556, in load_dataset
builder_instance = load_dataset_builder(
File "/opt/conda/lib/python3.10/site-packages/datasets/load.py", line 2265, in load_dataset_builder
builder_instance: DatasetBuilder = builder_cls(
File "/opt/conda/lib/python3.10/site-packages/datasets/builder.py", line 371, in __init__
self.config, self.config_id = self._create_builder_config(
File "/opt/conda/lib/python3.10/site-packages/datasets/builder.py", line 620, in _create_builder_config
builder_config._resolve_data_files(
File "/opt/conda/lib/python3.10/site-packages/datasets/builder.py", line 211, in _resolve_data_files
self.data_files = self.data_files.resolve(base_path, download_config)
File "/opt/conda/lib/python3.10/site-packages/datasets/data_files.py", line 799, in resolve
out[key] = data_files_patterns_list.resolve(base_path, download_config)
File "/opt/conda/lib/python3.10/site-packages/datasets/data_files.py", line 752, in resolve
resolve_pattern(
File "/opt/conda/lib/python3.10/site-packages/datasets/data_files.py", line 393, in resolve_pattern
raise FileNotFoundError(error_msg)
FileNotFoundError: Unable to find 'hf://datasets/trec@65752bf53af25bc935a0dce92fb5b6c930728450/default/train/0000.parquet' with any supported extension ['.csv', '.tsv', '.json', '.jsonl', '.parquet', '.geoparquet', '.gpq', '.arrow', '.txt', '.tar', '.blp', '.bmp', '.dib', '.bufr', '.cur', '.pcx', '.dcx', '.dds', '.ps', '.eps', '.fit', '.fits', '.fli', '.flc', '.ftc', '.ftu', '.gbr', '.gif', '.grib', '.h5', '.hdf', '.png', '.apng', '.jp2', '.j2k', '.jpc', '.jpf', '.jpx', '.j2c', '.icns', '.ico', '.im', '.iim', '.tif', '.tiff', '.jfif', '.jpe', '.jpg', '.jpeg', '.mpg', '.mpeg', '.msp', '.pcd', '.pxr', '.pbm', '.pgm', '.ppm', '.pnm', '.psd', '.bw', '.rgb', '.rgba', '.sgi', '.ras', '.tga', '.icb', '.vda', '.vst', '.webp', '.wmf', '.emf', '.xbm', '.xpm', '.BLP', '.BMP', '.DIB', '.BUFR', '.CUR', '.PCX', '.DCX', '.DDS', '.PS', '.EPS', '.FIT', '.FITS', '.FLI', '.FLC', '.FTC', '.FTU', '.GBR', '.GIF', '.GRIB', '.H5', '.HDF', '.PNG', '.APNG', '.JP2', '.J2K', '.JPC', '.JPF', '.JPX', '.J2C', '.ICNS', '.ICO', '.IM', '.IIM', '.TIF', '.TIFF', '.JFIF', '.JPE', '.JPG', '.JPEG', '.MPG', '.MPEG', '.MSP', '.PCD', '.PXR', '.PBM', '.PGM', '.PPM', '.PNM', '.PSD', '.BW', '.RGB', '.RGBA', '.SGI', '.RAS', '.TGA', '.ICB', '.VDA', '.VST', '.WEBP', '.WMF', '.EMF', '.XBM', '.XPM', '.aiff', '.au', '.avr', '.caf', '.flac', '.htk', '.svx', '.mat4', '.mat5', '.mpc2k', '.ogg', '.paf', '.pvf', '.raw', '.rf64', '.sd2', '.sds', '.ircam', '.voc', '.w64', '.wav', '.nist', '.wavex', '.wve', '.xi', '.mp3', '.opus', '.AIFF', '.AU', '.AVR', '.CAF', '.FLAC', '.HTK', '.SVX', '.MAT4', '.MAT5', '.MPC2K', '.OGG', '.PAF', '.PVF', '.RAW', '.RF64', '.SD2', '.SDS', '.IRCAM', '.VOC', '.W64', '.WAV', '.NIST', '.WAVEX', '.WVE', '.XI', '.MP3', '.OPUS', '.zip']
```
4. Similar issue also found for the following code:
```
dataset = load_dataset("sst", "default")
```
### Expected behavior
If the dataset is loaded correctly, one will have:
```
>>> print(dataset)
DatasetDict({
train: Dataset({
features: ['text', 'coarse_label', 'fine_label'],
num_rows: 5452
})
test: Dataset({
features: ['text', 'coarse_label', 'fine_label'],
num_rows: 500
})
})
>>>
```
### Environment info
- `datasets` version: 2.18.0
- Platform: Linux-6.2.0-35-generic-x86_64-with-glibc2.31
- Python version: 3.10.13
- `huggingface_hub` version: 0.20.3
- PyArrow version: 15.0.1
- Pandas version: 2.2.1
- `fsspec` version: 2023.12.2
|
[] |
{
"url": null,
"html_url": null,
"diff_url": null,
"patch_url": null,
"merged_at": null
}
| false
|
[
"You should be able to fix this by updating `huggingface_hub` with `pip install -U huggingface_hub`. We use this package under the hood to resolve the Hub's files."
] |
6,769
|
(Willing to PR) Datasets with custom python objects
|
### Feature request
Hi thanks for the library! I would like to have a huggingface Dataset, and one of its column is custom (non-serializable) Python objects. For example, a minimal code:
```
class MyClass:
pass
dataset = datasets.Dataset.from_list([
dict(a=MyClass(), b='hello'),
])
```
It gives error:
```
ArrowInvalid: Could not convert <__main__.MyClass object at 0x7a852830d050> with type MyClass: did not recognize Python value type when inferring an Arrow data type
```
I guess it is because Dataset forces to convert everything into arrow format. However, is there any ways to make the scenario work? Thanks!
### Motivation
(see above)
### Your contribution
Yes, I am happy to PR!
Cross-posted: https://discuss.huggingface.co/t/datasets-with-custom-python-objects/79050?u=fzyzcjy
EDIT: possibly related https://github.com/huggingface/datasets/issues/5766
|
[
{
"id": 1935892871,
"node_id": "MDU6TGFiZWwxOTM1ODkyODcx",
"url": "https://api.github.com/repos/huggingface/datasets/labels/enhancement",
"name": "enhancement",
"color": "a2eeef",
"default": true,
"description": "New feature or request"
}
] |
{
"url": null,
"html_url": null,
"diff_url": null,
"patch_url": null,
"merged_at": null
}
| false
|
[] |
6,767
|
fixing the issue 6755(small typo)
|
Fixed the issue #6755 on the typo mistake
|
[] |
{
"url": "https://api.github.com/repos/huggingface/datasets/pulls/6767",
"html_url": "https://github.com/huggingface/datasets/pull/6767",
"diff_url": "https://github.com/huggingface/datasets/pull/6767.diff",
"patch_url": "https://github.com/huggingface/datasets/pull/6767.patch",
"merged_at": "2024-04-02T14:01:18"
}
| true
|
[
"The docs for this PR live [here](https://moon-ci-docs.huggingface.co/docs/datasets/pr_6767). All of your documentation changes will be reflected on that endpoint. The docs are available until 30 days after the last update.",
"<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) | 0.005526 / 0.011353 (-0.005827) | 0.003839 / 0.011008 (-0.007169) | 0.064027 / 0.038508 (0.025519) | 0.032316 / 0.023109 (0.009206) | 0.250707 / 0.275898 (-0.025191) | 0.269222 / 0.323480 (-0.054258) | 0.004335 / 0.007986 (-0.003651) | 0.002703 / 0.004328 (-0.001626) | 0.049621 / 0.004250 (0.045370) | 0.047499 / 0.037052 (0.010446) | 0.262362 / 0.258489 (0.003873) | 0.292765 / 0.293841 (-0.001076) | 0.028661 / 0.128546 (-0.099885) | 0.010835 / 0.075646 (-0.064811) | 0.208910 / 0.419271 (-0.210362) | 0.036624 / 0.043533 (-0.006909) | 0.247448 / 0.255139 (-0.007691) | 0.270593 / 0.283200 (-0.012607) | 0.018988 / 0.141683 (-0.122695) | 1.141224 / 1.452155 (-0.310931) | 1.204944 / 1.492716 (-0.287772) |\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) | 0.096324 / 0.018006 (0.078318) | 0.292495 / 0.000490 (0.292006) | 0.000232 / 0.000200 (0.000032) | 0.000043 / 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) | 0.018379 / 0.037411 (-0.019032) | 0.065216 / 0.014526 (0.050690) | 0.074071 / 0.176557 (-0.102486) | 0.120793 / 0.737135 (-0.616343) | 0.075882 / 0.296338 (-0.220456) |\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) | 0.286354 / 0.215209 (0.071145) | 2.800766 / 2.077655 (0.723111) | 1.474126 / 1.504120 (-0.029994) | 1.358232 / 1.541195 (-0.182963) | 1.400639 / 1.468490 (-0.067851) | 0.578354 / 4.584777 (-4.006423) | 2.454441 / 3.745712 (-1.291271) | 2.927003 / 5.269862 (-2.342859) | 1.826127 / 4.565676 (-2.739550) | 0.063049 / 0.424275 (-0.361226) | 0.005010 / 0.007607 (-0.002597) | 0.342174 / 0.226044 (0.116129) | 3.415900 / 2.268929 (1.146971) | 1.854096 / 55.444624 (-53.590528) | 1.568626 / 6.876477 (-5.307851) | 1.660138 / 2.142072 (-0.481934) | 0.664059 / 4.805227 (-4.141168) | 0.120496 / 6.500664 (-6.380168) | 0.044664 / 0.075469 (-0.030805) |\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) | 0.988434 / 1.841788 (-0.853353) | 12.525563 / 8.074308 (4.451255) | 10.016862 / 10.191392 (-0.174530) | 0.134043 / 0.680424 (-0.546381) | 0.014349 / 0.534201 (-0.519852) | 0.287173 / 0.579283 (-0.292110) | 0.266499 / 0.434364 (-0.167865) | 0.325425 / 0.540337 (-0.214912) | 0.418772 / 1.386936 (-0.968164) |\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) | 0.005675 / 0.011353 (-0.005678) | 0.004238 / 0.011008 (-0.006770) | 0.051048 / 0.038508 (0.012540) | 0.033428 / 0.023109 (0.010319) | 0.283406 / 0.275898 (0.007508) | 0.309321 / 0.323480 (-0.014159) | 0.004354 / 0.007986 (-0.003631) | 0.003101 / 0.004328 (-0.001228) | 0.049369 / 0.004250 (0.045119) | 0.043252 / 0.037052 (0.006200) | 0.293097 / 0.258489 (0.034608) | 0.324392 / 0.293841 (0.030551) | 0.030524 / 0.128546 (-0.098022) | 0.010977 / 0.075646 (-0.064669) | 0.058546 / 0.419271 (-0.360726) | 0.033295 / 0.043533 (-0.010238) | 0.284929 / 0.255139 (0.029790) | 0.302925 / 0.283200 (0.019726) | 0.018586 / 0.141683 (-0.123097) | 1.156552 / 1.452155 (-0.295602) | 1.208856 / 1.492716 (-0.283860) |\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) | 0.096938 / 0.018006 (0.078932) | 0.305375 / 0.000490 (0.304886) | 0.000227 / 0.000200 (0.000027) | 0.000044 / 0.000054 (-0.000011) |\n\n### Benchmark: benchmark_indices_mapping.json\n\n| metric | select | shard | shuffle | sort | train_test_split |\n|--------|---|---|---|---|---|\n| new / old (diff) | 0.022658 / 0.037411 (-0.014754) | 0.078125 / 0.014526 (0.063599) | 0.087892 / 0.176557 (-0.088665) | 0.127745 / 0.737135 (-0.609390) | 0.089806 / 0.296338 (-0.206533) |\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) | 0.292434 / 0.215209 (0.077225) | 2.862329 / 2.077655 (0.784674) | 1.607948 / 1.504120 (0.103828) | 1.487179 / 1.541195 (-0.054016) | 1.542234 / 1.468490 (0.073744) | 0.579446 / 4.584777 (-4.005331) | 2.478549 / 3.745712 (-1.267163) | 2.923493 / 5.269862 (-2.346369) | 1.833161 / 4.565676 (-2.732515) | 0.064289 / 0.424275 (-0.359986) | 0.005638 / 0.007607 (-0.001969) | 0.350111 / 0.226044 (0.124067) | 3.436035 / 2.268929 (1.167107) | 1.970592 / 55.444624 (-53.474032) | 1.717474 / 6.876477 (-5.159002) | 1.753150 / 2.142072 (-0.388922) | 0.660495 / 4.805227 (-4.144732) | 0.119302 / 6.500664 (-6.381362) | 0.042633 / 0.075469 (-0.032836) |\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) | 1.018761 / 1.841788 (-0.823027) | 12.859834 / 8.074308 (4.785525) | 10.547789 / 10.191392 (0.356397) | 0.131986 / 0.680424 (-0.548438) | 0.016469 / 0.534201 (-0.517732) | 0.288585 / 0.579283 (-0.290698) | 0.270499 / 0.434364 (-0.163865) | 0.325801 / 0.540337 (-0.214537) | 0.416551 / 1.386936 (-0.970385) |\n\n</details>\n</details>\n\n![](https://cml.dev/watermark.png#7599f15537b094bfd18de5af7bb2a482c06d7a0e \"CML watermark\")\n"
] |
6,765
|
Compatibility issue between s3fs, fsspec, and datasets
|
### Describe the bug
Here is the full error stack when installing:
```
ERROR: pip's dependency resolver does not currently take into account all the packages that are installed. This behaviour is the source of the following dependency conflicts.
datasets 2.18.0 requires fsspec[http]<=2024.2.0,>=2023.1.0, but you have fsspec 2024.3.1 which is incompatible.
Successfully installed aiobotocore-2.12.1 aioitertools-0.11.0 botocore-1.34.51 fsspec-2024.3.1 jmespath-1.0.1 s3fs-2024.3.1 urllib3-2.0.7 wrapt-1.16.0
```
When I install with pip, pip allows this error to exist while still installing s3fs, but this error breaks poetry, since poetry will refuse to install s3fs because of the dependency conflict.
Maybe I'm missing something so maybe it's not a bug but some mistake on my end? Any input would be helpful. Thanks!
### Steps to reproduce the bug
1. conda create -n tmp python=3.10 -y
2. conda activate tmp
3. pip install datasets
4. pip install s3fs
### Expected behavior
I would expect there to be no error.
### Environment info
MacOS (ARM), Python3.10, conda 23.11.0.
|
[] |
{
"url": null,
"html_url": null,
"diff_url": null,
"patch_url": null,
"merged_at": null
}
| false
|
[
"Hi! Instead of running `pip install` separately for each package, you should pass all the packages to a single `pip install` call (in this case, `pip install datasets s3fs`) to let `pip` properly resolve their versions.",
"> Hi! Instead of running `pip install` separately for each package, you should pass all the packages to a single `pip install` call (in this case, `pip install datasets s3fs`) to let `pip` properly resolve their versions.\r\n\r\nThanks so much! My inexperience with pip is showing 😆 🙈 ",
"> Hi! Instead of running `pip install` separately for each package, you should pass all the packages to a single `pip install` call (in this case, `pip install datasets s3fs`) to let `pip` properly resolve their versions.\r\n\r\nyou are awesome bro"
] |
6,764
|
load_dataset can't work with symbolic links
|
### Feature request
Enable the `load_dataset` function to load local datasets with symbolic links.
E.g, this dataset can be loaded:
├── example_dataset/
│ ├── data/
│ │ ├── train/
│ │ │ ├── file0
│ │ │ ├── file1
│ │ ├── dev/
│ │ │ ├── file2
│ │ │ ├── file3
│ ├── metadata.csv
while this dataset can't:
├── example_dataset_symlink/
│ ├── data/
│ │ ├── train/
│ │ │ ├── sym0 -> file0
│ │ │ ├── sym1 -> file1
│ │ ├── dev/
│ │ │ ├── sym2 -> file2
│ │ │ ├── sym3 -> file3
│ ├── metadata.csv
I have created an example dataset in order to reproduce the problem:
1. Unzip `example_dataset.zip`.
2. Run `no_symlink.sh`. Training should start without issues.
3. Run `symlink.sh`. You will see that all four examples will be in train split, instead of having two examples in train and two examples in dev. The script won't load the correct audio files.
[example_dataset.zip](https://github.com/huggingface/datasets/files/14807053/example_dataset.zip)
### Motivation
I have a very large dataset locally. Instead of initiating training on the entire dataset, I need to start training on smaller subsets of the data. Due to the purpose of the experiments I am running, I will need to create many smaller datasets with overlapping data. Instead of copying the all the files for each subset, I would prefer copying symbolic links of the data. This way, the memory usage would not significantly increase beyond the initial dataset size.
Advantages of this approach:
- It would leave a smaller memory footprint on the hard drive
- Creating smaller datasets would be much faster
### Your contribution
I would gladly contribute, if this is something useful to the community. It seems like a simple change of code, something like `file_path = os.path.realpath(file_path)` should be added before loading the files. If anyone has insights on how to incorporate this functionality, I would greatly appreciate your knowledge and input.
|
[
{
"id": 1935892871,
"node_id": "MDU6TGFiZWwxOTM1ODkyODcx",
"url": "https://api.github.com/repos/huggingface/datasets/labels/enhancement",
"name": "enhancement",
"color": "a2eeef",
"default": true,
"description": "New feature or request"
}
] |
{
"url": null,
"html_url": null,
"diff_url": null,
"patch_url": null,
"merged_at": null
}
| false
|
[] |
6,763
|
Fix issue with case sensitivity when loading dataset from local cache
|
When a dataset with upper-cases in its name is first loaded using `load_dataset()`, the local cache directory is created with all lowercase letters.
However, upon subsequent loads, the current version attempts to locate the cache directory using the dataset's original name, which includes uppercase letters. This discrepancy can lead to confusion and, particularly in offline mode, results in errors.
### Reproduce
```bash
~$ python
Python 3.9.19 (main, Mar 21 2024, 17:11:28)
[GCC 11.2.0] :: Anaconda, Inc. on linux
Type "help", "copyright", "credits" or "license" for more information.
>>> from datasets import load_dataset
>>> dataset = load_dataset("locuslab/TOFU", "full")
>>> quit()
~$ export HF_DATASETS_OFFLINE=1
~$ python
Python 3.9.19 (main, Mar 21 2024, 17:11:28)
[GCC 11.2.0] :: Anaconda, Inc. on linux
Type "help", "copyright", "credits" or "license" for more information.
>>> from datasets import load_dataset
>>> dataset = load_dataset("locuslab/TOFU", "full")
Traceback (most recent call last):
File "<stdin>", line 1, in <module>
File "xxxxxx/anaconda3/envs/llm/lib/python3.9/site-packages/datasets/load.py", line 2556, in load_dataset
builder_instance = load_dataset_builder(
File "xxxxxx/anaconda3/envs/llm/lib/python3.9/site-packages/datasets/load.py", line 2228, in load_dataset_builder
dataset_module = dataset_module_factory(
File "xxxxxx/anaconda3/envs/llm/lib/python3.9/site-packages/datasets/load.py", line 1871, in dataset_module_factory
raise ConnectionError(f"Couldn't reach the Hugging Face Hub for dataset '{path}': {e1}") from None
ConnectionError: Couldn't reach the Hugging Face Hub for dataset 'locuslab/TOFU': Offline mode is enabled.
>>>
```
I fix this issue by lowering the dataset name (`.lower()`) when generating cache_dir.
|
[] |
{
"url": "https://api.github.com/repos/huggingface/datasets/pulls/6763",
"html_url": "https://github.com/huggingface/datasets/pull/6763",
"diff_url": "https://github.com/huggingface/datasets/pull/6763.diff",
"patch_url": "https://github.com/huggingface/datasets/pull/6763.patch",
"merged_at": null
}
| true
|
[
"I also need this feature for [\"Cnam-LMSSC/vibravox \"](https://huggingface.co/datasets/Cnam-LMSSC/vibravox)\r\n\r\n\r\nEDIT: Upgrading to `2.19.0` fixed my problem thanks to [this PR](https://github.com/huggingface/datasets/pull/6754)"
] |
6,762
|
Allow polars as valid output type
|
I was trying out polars as an output for a map function and found that it wasn't a valid return type in `validate_function_output`. Thought that we should accommodate this by creating and adding it to the `allowed_processed_input_types` variable.
|
[] |
{
"url": "https://api.github.com/repos/huggingface/datasets/pulls/6762",
"html_url": "https://github.com/huggingface/datasets/pull/6762",
"diff_url": "https://github.com/huggingface/datasets/pull/6762.diff",
"patch_url": "https://github.com/huggingface/datasets/pull/6762.patch",
"merged_at": null
}
| true
|
[
"The docs for this PR live [here](https://moon-ci-docs.huggingface.co/docs/datasets/pr_6762). All of your documentation changes will be reflected on that endpoint. The docs are available until 30 days after the last update."
] |
6,761
|
Remove deprecated code
|
What does this PR do?
1. remove `list_files_info` in favor of `list_repo_tree`. As of `0.23`, `list_files_info` will be removed for good. `datasets` had a utility to support both pre-0.20 and post-0.20 versions. Since `hfh` version is already pinned to `>=0.21.2`, I removed the legacy part.
2. `preupload_lfs_files` had also a different behavior between `<0.20` and `>=0.20`. I remove it since huggingface_hub is now pinned to `>=0.21.2`
3. `hf_hub_url` is overwritten to default to the dataset repo_type. I do think it is misleading to keep the same method naming for it. I renamed it to `get_dataset_url` for clarity. Let me know if you prefer to see this change reverted.
|
[] |
{
"url": "https://api.github.com/repos/huggingface/datasets/pulls/6761",
"html_url": "https://github.com/huggingface/datasets/pull/6761",
"diff_url": "https://github.com/huggingface/datasets/pull/6761.diff",
"patch_url": "https://github.com/huggingface/datasets/pull/6761.patch",
"merged_at": "2024-03-29T13:18:13"
}
| true
|
[
"The docs for this PR live [here](https://moon-ci-docs.huggingface.co/docs/datasets/pr_6761). All of your documentation changes will be reflected on that endpoint. The docs are available until 30 days after the last update.",
"Thanks for cleaning this :) I'm also fine with renaming `hf_dataset_url` (and not `get_dataset_url` as you said in your OP)",
"(Yep, `hf_dataset_url` is fine, made a mistake writing the PR description)",
"@albertvillanova Sorry about that, tests are now fixed! :)",
"<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) | 0.005357 / 0.011353 (-0.005995) | 0.003788 / 0.011008 (-0.007220) | 0.063630 / 0.038508 (0.025122) | 0.031353 / 0.023109 (0.008244) | 0.247525 / 0.275898 (-0.028373) | 0.282052 / 0.323480 (-0.041428) | 0.004247 / 0.007986 (-0.003739) | 0.002750 / 0.004328 (-0.001579) | 0.049467 / 0.004250 (0.045217) | 0.046663 / 0.037052 (0.009610) | 0.266440 / 0.258489 (0.007951) | 0.295230 / 0.293841 (0.001389) | 0.028271 / 0.128546 (-0.100276) | 0.011116 / 0.075646 (-0.064530) | 0.222092 / 0.419271 (-0.197179) | 0.036627 / 0.043533 (-0.006906) | 0.252607 / 0.255139 (-0.002532) | 0.271231 / 0.283200 (-0.011969) | 0.019070 / 0.141683 (-0.122613) | 1.152645 / 1.452155 (-0.299509) | 1.211267 / 1.492716 (-0.281449) |\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) | 0.095002 / 0.018006 (0.076996) | 0.304054 / 0.000490 (0.303564) | 0.000212 / 0.000200 (0.000012) | 0.000056 / 0.000054 (0.000001) |\n\n### Benchmark: benchmark_indices_mapping.json\n\n| metric | select | shard | shuffle | sort | train_test_split |\n|--------|---|---|---|---|---|\n| new / old (diff) | 0.018251 / 0.037411 (-0.019161) | 0.061929 / 0.014526 (0.047403) | 0.074641 / 0.176557 (-0.101916) | 0.122643 / 0.737135 (-0.614492) | 0.076744 / 0.296338 (-0.219594) |\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) | 0.284605 / 0.215209 (0.069396) | 2.774638 / 2.077655 (0.696984) | 1.473907 / 1.504120 (-0.030213) | 1.351054 / 1.541195 (-0.190141) | 1.348840 / 1.468490 (-0.119650) | 0.576243 / 4.584777 (-4.008534) | 2.444110 / 3.745712 (-1.301602) | 2.814741 / 5.269862 (-2.455121) | 1.762666 / 4.565676 (-2.803010) | 0.063959 / 0.424275 (-0.360316) | 0.005011 / 0.007607 (-0.002596) | 0.338406 / 0.226044 (0.112361) | 3.361213 / 2.268929 (1.092284) | 1.832674 / 55.444624 (-53.611950) | 1.564229 / 6.876477 (-5.312248) | 1.570843 / 2.142072 (-0.571230) | 0.657134 / 4.805227 (-4.148093) | 0.120041 / 6.500664 (-6.380623) | 0.048594 / 0.075469 (-0.026875) |\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) | 0.965328 / 1.841788 (-0.876460) | 11.704441 / 8.074308 (3.630133) | 9.895462 / 10.191392 (-0.295930) | 0.131913 / 0.680424 (-0.548511) | 0.015175 / 0.534201 (-0.519026) | 0.292022 / 0.579283 (-0.287261) | 0.269752 / 0.434364 (-0.164612) | 0.330453 / 0.540337 (-0.209884) | 0.421659 / 1.386936 (-0.965277) |\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) | 0.005472 / 0.011353 (-0.005881) | 0.003809 / 0.011008 (-0.007199) | 0.049594 / 0.038508 (0.011086) | 0.031858 / 0.023109 (0.008748) | 0.277622 / 0.275898 (0.001724) | 0.296092 / 0.323480 (-0.027388) | 0.004209 / 0.007986 (-0.003777) | 0.002726 / 0.004328 (-0.001603) | 0.048057 / 0.004250 (0.043806) | 0.043317 / 0.037052 (0.006265) | 0.288371 / 0.258489 (0.029882) | 0.312847 / 0.293841 (0.019007) | 0.029110 / 0.128546 (-0.099437) | 0.010792 / 0.075646 (-0.064854) | 0.058694 / 0.419271 (-0.360577) | 0.033315 / 0.043533 (-0.010218) | 0.281225 / 0.255139 (0.026086) | 0.297044 / 0.283200 (0.013844) | 0.018897 / 0.141683 (-0.122786) | 1.156417 / 1.452155 (-0.295738) | 1.221393 / 1.492716 (-0.271323) |\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) | 0.095065 / 0.018006 (0.077059) | 0.304107 / 0.000490 (0.303618) | 0.000213 / 0.000200 (0.000014) | 0.000043 / 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) | 0.021658 / 0.037411 (-0.015753) | 0.075948 / 0.014526 (0.061423) | 0.087019 / 0.176557 (-0.089537) | 0.127309 / 0.737135 (-0.609827) | 0.092251 / 0.296338 (-0.204087) |\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) | 0.291906 / 0.215209 (0.076697) | 2.865007 / 2.077655 (0.787352) | 1.591647 / 1.504120 (0.087527) | 1.474499 / 1.541195 (-0.066696) | 1.496644 / 1.468490 (0.028154) | 0.575337 / 4.584777 (-4.009440) | 2.569426 / 3.745712 (-1.176287) | 2.872611 / 5.269862 (-2.397251) | 1.804278 / 4.565676 (-2.761399) | 0.064225 / 0.424275 (-0.360050) | 0.005574 / 0.007607 (-0.002033) | 0.347724 / 0.226044 (0.121680) | 3.426418 / 2.268929 (1.157490) | 1.966270 / 55.444624 (-53.478355) | 1.687790 / 6.876477 (-5.188686) | 1.728530 / 2.142072 (-0.413542) | 0.650251 / 4.805227 (-4.154977) | 0.118381 / 6.500664 (-6.382283) | 0.041693 / 0.075469 (-0.033776) |\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) | 1.014203 / 1.841788 (-0.827585) | 12.219496 / 8.074308 (4.145188) | 10.469677 / 10.191392 (0.278285) | 0.141840 / 0.680424 (-0.538584) | 0.015104 / 0.534201 (-0.519097) | 0.288453 / 0.579283 (-0.290830) | 0.287467 / 0.434364 (-0.146897) | 0.331046 / 0.540337 (-0.209292) | 0.423731 / 1.386936 (-0.963205) |\n\n</details>\n</details>\n\n![](https://cml.dev/watermark.png#66d6242626eada79cfba4df39d99cd2bacb1cbea \"CML watermark\")\n"
] |
6,760
|
Load codeparrot/apps raising UnicodeDecodeError in datasets-2.18.0
|
### Describe the bug
This happens with datasets-2.18.0; I downgraded the version to 2.14.6 fixing this temporarily.
```
Traceback (most recent call last):
File "/home/xxx/miniconda3/envs/py310/lib/python3.10/site-packages/datasets/load.py", line 2556, in load_dataset
builder_instance = load_dataset_builder(
File "/home/xxx/miniconda3/envs/py310/lib/python3.10/site-packages/datasets/load.py", line 2228, in load_dataset_builder
dataset_module = dataset_module_factory(
File "/home/xxx/miniconda3/envs/py310/lib/python3.10/site-packages/datasets/load.py", line 1879, in dataset_module_factory
raise e1 from None
File "/home/xxx/miniconda3/envs/py310/lib/python3.10/site-packages/datasets/load.py", line 1831, in dataset_module_factory
can_load_config_from_parquet_export = "DEFAULT_CONFIG_NAME" not in f.read()
File "/home/xxx/miniconda3/envs/py310/lib/python3.10/codecs.py", line 322, in decode
(result, consumed) = self._buffer_decode(data, self.errors, final)
UnicodeDecodeError: 'utf-8' codec can't decode byte 0x8b in position 1: invalid start byte
```
### Steps to reproduce the bug
1. Using Python3.10/3.11
2. Install datasets-2.18.0
3. test with
```
from datasets import load_dataset
dataset = load_dataset("codeparrot/apps")
```
### Expected behavior
Normally it should manage to download and load the dataset without such error.
### Environment info
Ubuntu, Python3.10/3.11
|
[] |
{
"url": null,
"html_url": null,
"diff_url": null,
"patch_url": null,
"merged_at": null
}
| false
|
[
"The same error with mteb datasets.",
"Unfortunately, I'm unable to reproduce this error locally or on Colab.",
"Here is the requirements.txt from a clean virtual environment (managed by conda) where I only install `datasets` by \r\n`pip install datasets`. \r\nThe pip list:\r\n```\r\naiohttp==3.9.3\r\naiosignal==1.3.1\r\nattrs==23.2.0\r\ncertifi==2024.2.2\r\ncharset-normalizer==3.3.2\r\ndatasets==2.18.0\r\ndill==0.3.8\r\nfilelock==3.13.3\r\nfrozenlist==1.4.1\r\nfsspec==2024.2.0\r\nhuggingface-hub==0.22.2\r\nidna==3.6\r\nmultidict==6.0.5\r\nmultiprocess==0.70.16\r\nnumpy==1.26.4\r\npackaging==24.0\r\npandas==2.2.1\r\npyarrow==15.0.2\r\npyarrow-hotfix==0.6\r\npython-dateutil==2.9.0.post0\r\npytz==2024.1\r\nPyYAML==6.0.1\r\nrequests==2.31.0\r\nsix==1.16.0\r\ntqdm==4.66.2\r\ntyping_extensions==4.11.0\r\ntzdata==2024.1\r\nurllib3==2.2.1\r\nxxhash==3.4.1\r\nyarl==1.9.4\r\n```\r\nAnd the error can be reproduced.\r\n\r\nDowngrading to datasets==2.14.6 changes some packages' versions:\r\n\r\n```\r\nSuccessfully installed datasets-2.14.6 dill-0.3.7 fsspec-2023.10.0 multiprocess-0.70.15\r\n```\r\nand the dataset can be downloaded and loaded. \r\n\r\nThen I upgrade the version to 2.18.0 again; now the dataset can be loaded with such a line:\r\n```Using the latest cached version of the module from /home/xxx/.cache/huggingface/modules/datasets_modules/datasets/codeparrot--apps/04ac807715d07d6e5cc580f59cdc8213cd7dc4529d0bb819cca72c9f8e8c1aa5 (last modified on Sun Apr 7 09:06:43 2024) since it couldn't be found locally at codeparrot/apps, or remotely on the Hugging Face Hub. ```\r\n\r\nSo the latest version works wrong when requesting the dataset info. \r\n\r\n**But if you cannot reproduce this, I may ignore some detailed information: I use `HF_ENDPOINT=https://hf-mirror.com` for some reason (if not use this I cannot connect to huggingface resources) and the error occurs when requesting the dataset's info card.** \r\nMaybe the error is caused by this environment variable.\r\nI'll open an issue in the author's repo now.",
"> Here is the requirements.txt from a clean virtual environment (managed by conda) where I only install `datasets` by `pip install datasets`. The pip list:\r\n> \r\n> ```\r\n> aiohttp==3.9.3\r\n> aiosignal==1.3.1\r\n> attrs==23.2.0\r\n> certifi==2024.2.2\r\n> charset-normalizer==3.3.2\r\n> datasets==2.18.0\r\n> dill==0.3.8\r\n> filelock==3.13.3\r\n> frozenlist==1.4.1\r\n> fsspec==2024.2.0\r\n> huggingface-hub==0.22.2\r\n> idna==3.6\r\n> multidict==6.0.5\r\n> multiprocess==0.70.16\r\n> numpy==1.26.4\r\n> packaging==24.0\r\n> pandas==2.2.1\r\n> pyarrow==15.0.2\r\n> pyarrow-hotfix==0.6\r\n> python-dateutil==2.9.0.post0\r\n> pytz==2024.1\r\n> PyYAML==6.0.1\r\n> requests==2.31.0\r\n> six==1.16.0\r\n> tqdm==4.66.2\r\n> typing_extensions==4.11.0\r\n> tzdata==2024.1\r\n> urllib3==2.2.1\r\n> xxhash==3.4.1\r\n> yarl==1.9.4\r\n> ```\r\n> \r\n> And the error can be reproduced.\r\n> \r\n> Downgrading to datasets==2.14.6 changes some packages' versions:\r\n> \r\n> ```\r\n> Successfully installed datasets-2.14.6 dill-0.3.7 fsspec-2023.10.0 multiprocess-0.70.15\r\n> ```\r\n> \r\n> and the dataset can be downloaded and loaded.\r\n> \r\n> Then I upgrade the version to 2.18.0 again; now the dataset can be loaded with such a line: `Using the latest cached version of the module from /home/xxx/.cache/huggingface/modules/datasets_modules/datasets/codeparrot--apps/04ac807715d07d6e5cc580f59cdc8213cd7dc4529d0bb819cca72c9f8e8c1aa5 (last modified on Sun Apr 7 09:06:43 2024) since it couldn't be found locally at codeparrot/apps, or remotely on the Hugging Face Hub. `\r\n> \r\n> So the latest version works wrong when requesting the dataset info.\r\n> \r\n> **But if you cannot reproduce this, I may ignore some detailed information: I use `HF_ENDPOINT=https://hf-mirror.com` for some reason (if not use this I cannot connect to huggingface resources) and the error occurs when requesting the dataset's info card.** Maybe the error is caused by this environment variable. I'll open an issue in the author's repo now.\r\n\r\nThis is useful and my same error is settled!!!"
] |
6,759
|
Persistent multi-process Pool
|
### Feature request
Running .map and filter functions with `num_procs` consecutively instantiates several multiprocessing pools iteratively.
As instantiating a Pool is very resource intensive it can be a bottleneck to performing iteratively filtering.
My ideas:
1. There should be an option to declare `persistent_workers` similar to pytorch DataLoader. Downside would be that would be complex to determine the correct resource allocation and deallocation of the pool. i.e. the dataset can outlive the utility of the pool.
2. Provide a pool as an argument. Downside would be the expertise required by the user. Upside, is that there is better resource management.
### Motivation
Is really slow to iteratively perform map and filter operations on a dataset.
### Your contribution
If approved I could integrate it. I would need to know what method would be most suitable to implement from the two options above.
|
[
{
"id": 1935892871,
"node_id": "MDU6TGFiZWwxOTM1ODkyODcx",
"url": "https://api.github.com/repos/huggingface/datasets/labels/enhancement",
"name": "enhancement",
"color": "a2eeef",
"default": true,
"description": "New feature or request"
}
] |
{
"url": null,
"html_url": null,
"diff_url": null,
"patch_url": null,
"merged_at": null
}
| false
|
[] |
6,758
|
Passing `sample_by` to `load_dataset` when loading text data does not work
|
### Describe the bug
I have a dataset that consists of a bunch of text files, each representing an example. There is an undocumented `sample_by` argument for the `TextConfig` class that is used by `Text` to decide whether to split files into lines, paragraphs or take them whole. Passing `sample_by=“document”` to `load_dataset` results in files getting split into lines regardless. I have edited `src/datasets/packaged_modules/text/text.py` for myself to switch the default and it works fine.
As a side note, the `if-else` for `sample_by` will silently load an empty dataset if someone makes a typo in the argument, which is not ideal.
### Steps to reproduce the bug
1. Prepare data as a bunch of files in a directory.
2. Load that data via `load_dataset(“text”, data_files=<data_dir>/<files_glob>, …, sample_by=“document”)`.
3. Inspect the resultant dataset — every item should have the form of `{“text”: <a line from a file>}`.
### Expected behavior
`load_dataset(“text”, data_files=<data_dir>/<files_glob>, …, sample_by=“document”)` should result in a dataset with items of the form `{“text”: <one document>}`.
### Environment info
- `datasets` version: 2.18.0
- Platform: Linux-5.15.0-1046-nvidia-x86_64-with-glibc2.35
- Python version: 3.11.8
- `huggingface_hub` version: 0.21.4
- PyArrow version: 15.0.2
- Pandas version: 2.2.1
- `fsspec` version: 2024.2.0
|
[] |
{
"url": null,
"html_url": null,
"diff_url": null,
"patch_url": null,
"merged_at": null
}
| false
|
[
"Thanks for reporting! We are working on a fix."
] |
6,757
|
Test disabling transformers containers in docs CI
|
Related to https://github.com/huggingface/doc-builder/pull/487 and [internal slack thread](https://huggingface.slack.com/archives/C04F8N7FQNL/p1711384899462349?thread_ts=1711041424.720769&cid=C04F8N7FQNL). There is now a `custom_container` option when building docs in CI. When set to `""` (instead of `"huggingface/transformers-doc-builder"` by default), we don't run the CI inside a container, therefore saving ~2min of download time. The plan is to test disabling the transformers container on a few "big" repo and if everything works correctly, we will stop making it the default container. More details on https://github.com/huggingface/doc-builder/pull/487.
cc @mishig25
|
[] |
{
"url": "https://api.github.com/repos/huggingface/datasets/pulls/6757",
"html_url": "https://github.com/huggingface/datasets/pull/6757",
"diff_url": "https://github.com/huggingface/datasets/pull/6757.diff",
"patch_url": "https://github.com/huggingface/datasets/pull/6757.patch",
"merged_at": null
}
| true
|
[
"The docs for this PR live [here](https://moon-ci-docs.huggingface.co/docs/datasets/pr_6757). All of your documentation changes will be reflected on that endpoint. The docs are available until 30 days after the last update.",
"On slack it was mentioned that it was actually slower for `datasets`, should we close this one or am I missing something ?",
"@lhoestq I converted to draft. Want to make some more tests and will let you know"
] |
6,756
|
Support SQLite files?
|
### Feature request
Support loading a dataset from a SQLite file
https://huggingface.co/datasets/severo/test_iris_sqlite/tree/main
### Motivation
SQLite is a popular file format.
### Your contribution
See discussion on slack: https://huggingface.slack.com/archives/C04L6P8KNQ5/p1702481859117909 (internal)
In particular: a SQLite file can contain multiple tables, which might be matched to multiple configs. Maybe the detail of splits and configs should be defined in the README YAML, or use the same format as for ZIP files: `Iris.sqlite::Iris`.
See dataset here: https://huggingface.co/datasets/severo/test_iris_sqlite
Note: should we also support DuckDB files?
|
[
{
"id": 1935892871,
"node_id": "MDU6TGFiZWwxOTM1ODkyODcx",
"url": "https://api.github.com/repos/huggingface/datasets/labels/enhancement",
"name": "enhancement",
"color": "a2eeef",
"default": true,
"description": "New feature or request"
}
] |
{
"url": null,
"html_url": null,
"diff_url": null,
"patch_url": null,
"merged_at": null
}
| false
|
[
"You can use `Dataset.from_sql(path_to_sql_file)` already. Though we haven't added the Sql dataset builder to the `_PACKAGED_DATASETS_MODULES` list or in `_EXTENSION_TO_MODULE` to map `.sqlite` to the Sql dataset builder\r\n\r\nThis would allow to load a dataset repository with a `.sqlite` file using `load_dataset` and enable the Dataset Viewer",
"Considering `Dataset.from_sql`'s (extremely) low usage, I don't think many users are interested in using this format for their datasets. Also, SQLite files are hard/impossible to stream efficiently and require custom logic to define splits/subsets, so IMO we shouldn't encourage people to use SQLite on the Hub.\r\n\r\n@severo Do you have some real-world examples of datasets published in this format?",
"No. Indeed, it seems better to explicitly not support sqlite"
] |
6,755
|
Small typo on the documentation
|
### Describe the bug
There is a small typo on https://github.com/huggingface/datasets/blob/d5468836fe94e8be1ae093397dd43d4a2503b926/src/datasets/dataset_dict.py#L938
It should be `caching is enabled`.
### Steps to reproduce the bug
Please visit
https://github.com/huggingface/datasets/blob/d5468836fe94e8be1ae093397dd43d4a2503b926/src/datasets/dataset_dict.py#L938
### Expected behavior
`caching is enabled`
### Environment info
- `datasets` version: 2.17.1
- Platform: Linux-5.15.0-101-generic-x86_64-with-glibc2.35
- Python version: 3.11.7
- `huggingface_hub` version: 0.20.3
- PyArrow version: 15.0.0
- Pandas version: 2.2.1
- `fsspec` version: 2023.10.0
|
[
{
"id": 1935892877,
"node_id": "MDU6TGFiZWwxOTM1ODkyODc3",
"url": "https://api.github.com/repos/huggingface/datasets/labels/good%20first%20issue",
"name": "good first issue",
"color": "7057ff",
"default": true,
"description": "Good for newcomers"
}
] |
{
"url": null,
"html_url": null,
"diff_url": null,
"patch_url": null,
"merged_at": null
}
| false
|
[
"Thanks for reporting @fostiropoulos! I've edited your comment to fix the link to the problematic line.\r\n",
"@mariosasko can i take this up?",
"#self-assign"
] |
6,754
|
Fix cache path to snakecase for `CachedDatasetModuleFactory` and `Cache`
|
Fix https://github.com/huggingface/datasets/issues/6750#issuecomment-2016678729
I didn't find a guideline on how to run the tests, so i just run the following steps to make sure that this bug is fixed.
1. `python test.py`,
2. then `HF_DATASETS_OFFLINE=1 python test.py`
The `test.py` is
```
import datasets
datasets.utils.logging.set_verbosity_info()
ds = datasets.load_dataset('izhx/STS17-debug')
print(ds)
ds = datasets.load_dataset('C-MTEB/AFQMC', revision='b44c3b011063adb25877c13823db83bb193913c4')
print(ds)
```
|
[] |
{
"url": "https://api.github.com/repos/huggingface/datasets/pulls/6754",
"html_url": "https://github.com/huggingface/datasets/pull/6754",
"diff_url": "https://github.com/huggingface/datasets/pull/6754.diff",
"patch_url": "https://github.com/huggingface/datasets/pull/6754.patch",
"merged_at": "2024-04-15T15:38:51"
}
| true
|
[
"@lhoestq hi 😃, is there something else I need to do to check this change?",
"I added two tests and passed them on my server.\r\n\r\n```\r\npytest tests/packaged_modules/test_cache.py \r\n========================================================================== test session starts ==========================================================================\r\nplatform linux -- Python 3.11.5, pytest-8.1.1, pluggy-1.4.0\r\nrootdir: /mnt/nas/datasets\r\nconfigfile: pyproject.toml\r\nplugins: xdist-3.5.0, datadir-1.5.0\r\ncollected 8 items \r\n\r\ntests/packaged_modules/test_cache.py ........ [100%]\r\n\r\n========================================================================== 8 passed in 50.71s ===========================================================================\r\n\r\n```\r\n\r\n```\r\npytest tests/test_load.py\r\n========================================================================== test session starts ==========================================================================\r\nplatform linux -- Python 3.11.5, pytest-8.1.1, pluggy-1.4.0\r\nrootdir: /mnt/nas/datasets\r\nconfigfile: pyproject.toml\r\nplugins: xdist-3.5.0, datadir-1.5.0\r\ncollected 151 items \r\n\r\ntests/test_load.py .............................................................................................................................................. [ 94%]\r\n......... [100%]\r\n\r\n...\r\n\r\n============================================================= 151 passed, 29 warnings in 578.36s (0:09:38) ==============================================================\r\n```\r\n",
"The docs for this PR live [here](https://moon-ci-docs.huggingface.co/docs/datasets/pr_6754). All of your documentation changes will be reflected on that endpoint. The docs are available until 30 days after the last update.",
"Hi @izhx! I have also faced this issue, happy to see it already addressed, looking forward for PR merge :)",
"@lhoestq What do you think of these tests? 😀",
"<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) | 0.005060 / 0.011353 (-0.006293) | 0.003251 / 0.011008 (-0.007757) | 0.063538 / 0.038508 (0.025030) | 0.031178 / 0.023109 (0.008069) | 0.249971 / 0.275898 (-0.025927) | 0.284828 / 0.323480 (-0.038652) | 0.004183 / 0.007986 (-0.003802) | 0.002656 / 0.004328 (-0.001673) | 0.049585 / 0.004250 (0.045335) | 0.042656 / 0.037052 (0.005604) | 0.270962 / 0.258489 (0.012473) | 0.296091 / 0.293841 (0.002250) | 0.028065 / 0.128546 (-0.100482) | 0.010545 / 0.075646 (-0.065102) | 0.207323 / 0.419271 (-0.211948) | 0.035977 / 0.043533 (-0.007556) | 0.257315 / 0.255139 (0.002176) | 0.272238 / 0.283200 (-0.010962) | 0.017984 / 0.141683 (-0.123699) | 1.131314 / 1.452155 (-0.320840) | 1.180259 / 1.492716 (-0.312457) |\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) | 0.090977 / 0.018006 (0.072971) | 0.284021 / 0.000490 (0.283531) | 0.000264 / 0.000200 (0.000065) | 0.000044 / 0.000054 (-0.000011) |\n\n### Benchmark: benchmark_indices_mapping.json\n\n| metric | select | shard | shuffle | sort | train_test_split |\n|--------|---|---|---|---|---|\n| new / old (diff) | 0.017852 / 0.037411 (-0.019559) | 0.061288 / 0.014526 (0.046762) | 0.073844 / 0.176557 (-0.102713) | 0.121371 / 0.737135 (-0.615764) | 0.075036 / 0.296338 (-0.221303) |\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) | 0.287599 / 0.215209 (0.072390) | 2.821172 / 2.077655 (0.743517) | 1.459904 / 1.504120 (-0.044216) | 1.340224 / 1.541195 (-0.200970) | 1.357350 / 1.468490 (-0.111140) | 0.557344 / 4.584777 (-4.027433) | 2.412177 / 3.745712 (-1.333535) | 2.745126 / 5.269862 (-2.524735) | 1.754600 / 4.565676 (-2.811077) | 0.062487 / 0.424275 (-0.361788) | 0.005306 / 0.007607 (-0.002301) | 0.338856 / 0.226044 (0.112811) | 3.354953 / 2.268929 (1.086024) | 1.803208 / 55.444624 (-53.641417) | 1.553051 / 6.876477 (-5.323426) | 1.554790 / 2.142072 (-0.587282) | 0.651380 / 4.805227 (-4.153847) | 0.117777 / 6.500664 (-6.382887) | 0.041992 / 0.075469 (-0.033477) |\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) | 0.977588 / 1.841788 (-0.864200) | 11.363058 / 8.074308 (3.288750) | 9.791770 / 10.191392 (-0.399622) | 0.130708 / 0.680424 (-0.549716) | 0.013798 / 0.534201 (-0.520403) | 0.288313 / 0.579283 (-0.290970) | 0.268170 / 0.434364 (-0.166194) | 0.324815 / 0.540337 (-0.215522) | 0.419260 / 1.386936 (-0.967676) |\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) | 0.005187 / 0.011353 (-0.006166) | 0.003348 / 0.011008 (-0.007660) | 0.050309 / 0.038508 (0.011801) | 0.031334 / 0.023109 (0.008225) | 0.279542 / 0.275898 (0.003644) | 0.299608 / 0.323480 (-0.023872) | 0.004202 / 0.007986 (-0.003784) | 0.002735 / 0.004328 (-0.001593) | 0.050321 / 0.004250 (0.046070) | 0.039793 / 0.037052 (0.002740) | 0.289972 / 0.258489 (0.031483) | 0.313887 / 0.293841 (0.020046) | 0.028797 / 0.128546 (-0.099750) | 0.010166 / 0.075646 (-0.065480) | 0.059228 / 0.419271 (-0.360044) | 0.032667 / 0.043533 (-0.010866) | 0.278409 / 0.255139 (0.023270) | 0.292208 / 0.283200 (0.009008) | 0.017577 / 0.141683 (-0.124106) | 1.175046 / 1.452155 (-0.277109) | 1.200766 / 1.492716 (-0.291950) |\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) | 0.092236 / 0.018006 (0.074230) | 0.298860 / 0.000490 (0.298370) | 0.000211 / 0.000200 (0.000011) | 0.000043 / 0.000054 (-0.000011) |\n\n### Benchmark: benchmark_indices_mapping.json\n\n| metric | select | shard | shuffle | sort | train_test_split |\n|--------|---|---|---|---|---|\n| new / old (diff) | 0.021475 / 0.037411 (-0.015936) | 0.074414 / 0.014526 (0.059888) | 0.087746 / 0.176557 (-0.088811) | 0.124757 / 0.737135 (-0.612378) | 0.088513 / 0.296338 (-0.207826) |\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) | 0.296583 / 0.215209 (0.081374) | 2.894978 / 2.077655 (0.817323) | 1.590806 / 1.504120 (0.086686) | 1.463251 / 1.541195 (-0.077944) | 1.478751 / 1.468490 (0.010261) | 0.571724 / 4.584777 (-4.013053) | 2.454356 / 3.745712 (-1.291356) | 2.789275 / 5.269862 (-2.480586) | 1.753866 / 4.565676 (-2.811811) | 0.064787 / 0.424275 (-0.359488) | 0.005321 / 0.007607 (-0.002287) | 0.348454 / 0.226044 (0.122410) | 3.453052 / 2.268929 (1.184124) | 1.972237 / 55.444624 (-53.472388) | 1.677822 / 6.876477 (-5.198655) | 1.674750 / 2.142072 (-0.467322) | 0.649353 / 4.805227 (-4.155874) | 0.117135 / 6.500664 (-6.383529) | 0.040018 / 0.075469 (-0.035451) |\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) | 1.029812 / 1.841788 (-0.811976) | 11.945063 / 8.074308 (3.870755) | 10.238380 / 10.191392 (0.046988) | 0.146225 / 0.680424 (-0.534199) | 0.015262 / 0.534201 (-0.518939) | 0.286632 / 0.579283 (-0.292651) | 0.272952 / 0.434364 (-0.161412) | 0.323098 / 0.540337 (-0.217239) | 0.423549 / 1.386936 (-0.963387) |\n\n</details>\n</details>\n\n![](https://cml.dev/watermark.png#91b07b90915d7f7313d44ca3ff67673b9ad26bf4 \"CML watermark\")\n"
] |
6,753
|
Type error when importing datasets on Kaggle
|
### Describe the bug
When trying to run
```
import datasets
print(datasets.__version__)
```
It generates the following error
```
TypeError: expected string or bytes-like object
```
It looks like It cannot find the valid versions of `fsspec`
though fsspec version is fine when I checked Via command
```
import fsspec
print(fsspec.__version__)
# output: 2024.3.1
```
Detailed crash report
```
---------------------------------------------------------------------------
TypeError Traceback (most recent call last)
Cell In[1], line 1
----> 1 import datasets
2 print(datasets.__version__)
File /opt/conda/lib/python3.10/site-packages/datasets/__init__.py:18
1 # ruff: noqa
2 # Copyright 2020 The HuggingFace Datasets Authors and the TensorFlow Datasets Authors.
3 #
(...)
13 # See the License for the specific language governing permissions and
14 # limitations under the License.
16 __version__ = "2.18.0"
---> 18 from .arrow_dataset import Dataset
19 from .arrow_reader import ReadInstruction
20 from .builder import ArrowBasedBuilder, BeamBasedBuilder, BuilderConfig, DatasetBuilder, GeneratorBasedBuilder
File /opt/conda/lib/python3.10/site-packages/datasets/arrow_dataset.py:66
63 from multiprocess import Pool
64 from tqdm.contrib.concurrent import thread_map
---> 66 from . import config
67 from .arrow_reader import ArrowReader
68 from .arrow_writer import ArrowWriter, OptimizedTypedSequence
File /opt/conda/lib/python3.10/site-packages/datasets/config.py:41
39 # Imports
40 DILL_VERSION = version.parse(importlib.metadata.version("dill"))
---> 41 FSSPEC_VERSION = version.parse(importlib.metadata.version("fsspec"))
42 PANDAS_VERSION = version.parse(importlib.metadata.version("pandas"))
43 PYARROW_VERSION = version.parse(importlib.metadata.version("pyarrow"))
File /opt/conda/lib/python3.10/site-packages/packaging/version.py:49, in parse(version)
43 """
44 Parse the given version string and return either a :class:`Version` object
45 or a :class:`LegacyVersion` object depending on if the given version is
46 a valid PEP 440 version or a legacy version.
47 """
48 try:
---> 49 return Version(version)
50 except InvalidVersion:
51 return LegacyVersion(version)
File /opt/conda/lib/python3.10/site-packages/packaging/version.py:264, in Version.__init__(self, version)
261 def __init__(self, version: str) -> None:
262
263 # Validate the version and parse it into pieces
--> 264 match = self._regex.search(version)
265 if not match:
266 raise InvalidVersion(f"Invalid version: '{version}'")
TypeError: expected string or bytes-like object
```
### Steps to reproduce the bug
1. run `!pip install -U datasets` on kaggle
2. check datasets is installed via
```
import datasets
print(datasets.__version__)
```
### Expected behavior
Expected to print datasets version, like `2.18.0`
### Environment info
Running on Kaggle, latest enviornment , here is the notebook https://www.kaggle.com/code/jtv199/mistrial-7b-part2
|
[] |
{
"url": null,
"html_url": null,
"diff_url": null,
"patch_url": null,
"merged_at": null
}
| false
|
[
"I have the same problem \r\nIt seems that it only appears when you are using GPU \r\nIt seems to work fine with the 2.17 version though",
"Same here.",
"> I have the same problem\r\n> It seems that it only appears when you are using GPU\r\n> It seems to work fine with the 2.17 version though\r\n\r\nI downgraded from 2.18 to 2.17, and it works with CPU/GPU .. except now pyarrow complains\r\n\r\n```\r\n...\r\nFile /opt/conda/lib/python3.10/site-packages/pyarrow/array.pxi:830, in pyarrow.lib._PandasConvertible.to_pandas()\r\n\r\nFile /opt/conda/lib/python3.10/site-packages/pyarrow/table.pxi:3989, in pyarrow.lib.Table._to_pandas()\r\n\r\nImportError: cannot import name table_to_blockmanager\r\n```\r\n\r\nsee also https://www.kaggle.com/competitions/pii-detection-removal-from-educational-data/discussion/487474#2722594",
"Solved for me by downgrading `!pip install -U datasets==2.16.0` Works with gpu aswell",
"I think you should remain open this issue. It works at the previous version but not the latter versions. It is possible as a bug that the maintainer could take note for.",
"> Solved for me by downgrading `!pip install -U datasets==2.16.0` Works with gpu as well\r\n\r\nVerified it's working w/ GPU if I make these 3 updates.\r\n\r\n```\r\ndatasets==2.16.0\r\nfsspec==2023.10.0\r\ngcsfs==2023.10.0\r\n```\r\n\r\nbut the issue shouldn't be closed, this is just a workaround until they get the issue with 2.18.0 resolved.\r\n\r\nSee also: https://www.kaggle.com/competitions/pii-detection-removal-from-educational-data/discussion/487474",
"> > Solved for me by downgrading `!pip install -U datasets==2.16.0` Works with gpu as well\r\n> \r\n> Verified it's working w/ GPU if I make these 3 updates.\r\n> \r\n> ```\r\n> datasets==2.16.0\r\n> fsspec==2023.10.0\r\n> gcsfs==2023.10.0\r\n> ```\r\n> \r\n> but the issue shouldn't be closed, this is just a workaround until they get the issue with 2.18.0 resolved.\r\n> \r\n> See also: https://www.kaggle.com/competitions/pii-detection-removal-from-educational-data/discussion/487474\r\n\r\nThis also works for me, thanks"
] |
6,752
|
Precision being changed from float16 to float32 unexpectedly
|
### Describe the bug
I'm loading a HuggingFace Dataset for images.
I'm running a preprocessing (map operation) step that runs a few operations, one of them being conversion to float16. The Dataset features also say that the 'img' is of type float16. Whenever I take an image from that HuggingFace Dataset instance, the type turns out to be float32.
### Steps to reproduce the bug
```python
import torchvision.transforms.v2 as transforms
from datasets import load_dataset
dataset = load_dataset('cifar10', split='test')
dataset = dataset.with_format("torch")
data_transform = transforms.Compose([transforms.Resize((32, 32)),
transforms.ToDtype(torch.float16, scale=True),
transforms.Normalize([0.5, 0.5, 0.5], [0.5, 0.5, 0.5]),
])
def _preprocess(examples):
# Permutes from (BS x H x W x C) to (BS x C x H x W)
images = torch.permute(examples['img'], (0, 3, 2, 1))
examples['img'] = data_transform(images)
return examples
dataset = dataset.map(_preprocess, batched=True, batch_size=8)
```
Now at this point the dataset.features are showing float16 which is great because that's what I want.
```python
print(data_loader.features['img'])
Sequence(feature=Sequence(feature=Sequence(feature=Value(dtype='float16', id=None), length=-1, id=None), length=-1, id=None), length=-1, id=None)
```
But when I try to sample an image from this dataloader; I'm getting a float32 image, when I'm expecting float16:
```python
print(next(iter(data_loader))['img'].dtype)
torch.float32
```
### Expected behavior
I'm expecting the images loaded after the transformation to stay in float16.
### Environment info
- `datasets` version: 2.18.0
- Platform: Linux-5.15.146.1-microsoft-standard-WSL2-x86_64-with-glibc2.31
- Python version: 3.10.9
- `huggingface_hub` version: 0.21.4
- PyArrow version: 14.0.2
- Pandas version: 2.0.3
- `fsspec` version: 2023.10.0
|
[] |
{
"url": null,
"html_url": null,
"diff_url": null,
"patch_url": null,
"merged_at": null
}
| false
|
[
"This is because of the formatter (`torch` in this case).\r\nIt defaults to `float32`.\r\n\r\nYou can load it in `float16` using `dataset.set_format(\"torch\", dtype=torch.float16)`."
] |
6,751
|
Use 'with' operator for some download functions
|
Some functions in `streaming_download_manager.py` are not closing the file they open which lead to `Unclosed file` warnings in our code. This fixes a few of them.
|
[] |
{
"url": "https://api.github.com/repos/huggingface/datasets/pulls/6751",
"html_url": "https://github.com/huggingface/datasets/pull/6751",
"diff_url": "https://github.com/huggingface/datasets/pull/6751.diff",
"patch_url": "https://github.com/huggingface/datasets/pull/6751.patch",
"merged_at": null
}
| true
|
[
"The docs for this PR live [here](https://moon-ci-docs.huggingface.co/docs/datasets/pr_6751). All of your documentation changes will be reflected on that endpoint. The docs are available until 30 days after the last update.",
"I was mistaken on the intent of those functions, closing the PR."
] |
6,750
|
`load_dataset` requires a network connection for local download?
|
### Describe the bug
Hi all - I see that in the past a network dependency has been mistakenly introduced into `load_dataset` even for local loads. Is it possible this has happened again?
### Steps to reproduce the bug
```
>>> import datasets
>>> datasets.load_dataset("hh-rlhf")
Repo card metadata block was not found. Setting CardData to empty.
*hangs bc i'm firewalled*
````
stack trace from ctrl-c:
```
^CTraceback (most recent call last):
File "<stdin>", line 1, in <module>
File "/home/jobuser/.local/lib/python3.10/site-packages/datasets/load.py", line 2582, in load_dataset
builder_instance.download_and_prepare(
output_path = get_from_cache( [0/122]
File "/home/jobuser/.local/lib/python3.10/site-packages/datasets/utils/file_utils.py", line 532, in get_from_cache
response = http_head(
File "/home/jobuser/.local/lib/python3.10/site-packages/datasets/utils/file_utils.py", line 419, in http_head
response = _request_with_retry(
File "/home/jobuser/.local/lib/python3.10/site-packages/datasets/utils/file_utils.py", line 304, in _request_with_retry
response = requests.request(method=method.upper(), url=url, timeout=timeout, **params)
File "/home/jobuser/build/lipy-flytekit-image/environments/satellites/python/lib/python3.10/site-packages/requests/api.py", line 59, in request
return session.request(method=method, url=url, **kwargs)
File "/home/jobuser/build/lipy-flytekit-image/environments/satellites/python/lib/python3.10/site-packages/requests/sessions.py", line 587, in request
resp = self.send(prep, **send_kwargs)
File "/home/jobuser/build/lipy-flytekit-image/environments/satellites/python/lib/python3.10/site-packages/requests/sessions.py", line 701, in send
r = adapter.send(request, **kwargs)
File "/home/jobuser/build/lipy-flytekit-image/environments/satellites/python/lib/python3.10/site-packages/requests/adapters.py", line 487, in send
resp = conn.urlopen(
File "/home/jobuser/build/lipy-flytekit-image/environments/satellites/python/lib/python3.10/site-packages/urllib3/connectionpool.py", line 703, in urlopen
httplib_response = self._make_request(
File "/home/jobuser/build/lipy-flytekit-image/environments/satellites/python/lib/python3.10/site-packages/urllib3/connectionpool.py", line 386, in _make_request
self._validate_conn(conn)
File "/home/jobuser/build/lipy-flytekit-image/environments/satellites/python/lib/python3.10/site-packages/urllib3/connectionpool.py", line 1042, in _validate_conn
conn.connect()
File "/home/jobuser/build/lipy-flytekit-image/environments/satellites/python/lib/python3.10/site-packages/urllib3/connection.py", line 363, in connect
self.sock = conn = self._new_conn()
File "/home/jobuser/build/lipy-flytekit-image/environments/satellites/python/lib/python3.10/site-packages/urllib3/connection.py", line 174, in _new_conn
conn = connection.create_connection(
File "/home/jobuser/build/lipy-flytekit-image/environments/satellites/python/lib/python3.10/site-packages/urllib3/util/connection.py", line 85, in create_connection
sock.connect(sa)
KeyboardInterrupt
```
### Expected behavior
loads the dataset
### Environment info
```
> pip show datasets
Name: datasets
Version: 2.18.0
```
Python 3.10.2
|
[] |
{
"url": null,
"html_url": null,
"diff_url": null,
"patch_url": null,
"merged_at": null
}
| false
|
[
"Are you using `HF_DATASETS_OFFLINE=1` ?",
"> Are you using `HF_DATASETS_OFFLINE=1` ?\r\n\r\nThis doesn't work for me. `datasets=2.18.0`\r\n\r\n`test.py`:\r\n```\r\nimport datasets\r\n\r\ndatasets.utils.logging.set_verbosity_info()\r\n\r\nds = datasets.load_dataset('C-MTEB/AFQMC', revision='b44c3b011063adb25877c13823db83bb193913c4')\r\n\r\nprint(ds)\r\n```\r\n\r\nrun `python test.py`\r\n```\r\nGenerating dataset afqmc (/home/data/.cache/huggingface/datasets/C-MTEB___afqmc/default/0.0.0/b44c3b011063adb25877c13823db83bb193913c4)\r\nDownloading and preparing dataset afqmc/default to /home/data/.cache/huggingface/datasets/C-MTEB___afqmc/default/0.0.0/b44c3b011063adb25877c13823db83bb193913c4...\r\nDataset not on Hf google storage. Downloading and preparing it from source\r\nhf://datasets/C-MTEB/AFQMC@b44c3b011063adb25877c13823db83bb193913c4/data/validation-00000-of-00001-b8fc393b5ddedac7.parquet not found in cache or force_download set to True, downloading to /home/data/.cache/huggingface/datasets/downloads/78949f93104662359f4f3d5a2f7ec1ae37af5a5af44420a51212ea08c0be966b.incomplete\r\nDownloading data: 100%|████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 240k/240k [00:01<00:00, 178kB/s]\r\nstoring hf://datasets/C-MTEB/AFQMC@b44c3b011063adb25877c13823db83bb193913c4/data/validation-00000-of-00001-b8fc393b5ddedac7.parquet in cache at /home/data/.cache/huggingface/datasets/downloads/78949f93104662359f4f3d5a2f7ec1ae37af5a5af44420a51212ea08c0be966b\r\ncreating metadata file for /home/data/.cache/huggingface/datasets/downloads/78949f93104662359f4f3d5a2f7ec1ae37af5a5af44420a51212ea08c0be966b\r\nDownloading took 0.0 min\r\nChecksum Computation took 0.0 min\r\nGenerating test split\r\nGenerating test split: 100%|████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 3861/3861 [00:00<00:00, 3972.00 examples/s]\r\nGenerating train split\r\nGenerating train split: 100%|████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 34334/34334 [00:00<00:00, 34355.50 examples/s]\r\nGenerating validation split\r\nGenerating validation split: 100%|██████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 4316/4316 [00:00<00:00, 4477.00 examples/s]\r\nAll the splits matched successfully.\r\nDataset afqmc downloaded and prepared to /home/data/.cache/huggingface/datasets/C-MTEB___afqmc/default/0.0.0/b44c3b011063adb25877c13823db83bb193913c4. Subsequent calls will reuse this data.\r\nDatasetDict({\r\n test: Dataset({\r\n features: ['sentence1', 'sentence2', 'score', 'idx'],\r\n num_rows: 3861\r\n })\r\n train: Dataset({\r\n features: ['sentence1', 'sentence2', 'score', 'idx'],\r\n num_rows: 34334\r\n })\r\n validation: Dataset({\r\n features: ['sentence1', 'sentence2', 'score', 'idx'],\r\n num_rows: 4316\r\n })\r\n})\r\n```\r\n\r\nThen run `HF_DATASETS_OFFLINE=1 python test.py`\r\n```\r\nTraceback (most recent call last):\r\n File \"test.py\", line 9, in <module>\r\n ds = datasets.load_dataset('C-MTEB/AFQMC', revision='b44c3b011063adb25877c13823db83bb193913c4')\r\n File \"/dev/shm/tmp_env/lib/python3.10/site-packages/datasets/load.py\", line 2556, in load_dataset\r\n builder_instance = load_dataset_builder(\r\n File \"/dev/shm/tmp_env/lib/python3.10/site-packages/datasets/load.py\", line 2228, in load_dataset_builder\r\n dataset_module = dataset_module_factory(\r\n File \"/dev/shm/tmp_env/lib/python3.10/site-packages/datasets/load.py\", line 1871, in dataset_module_factory\r\n raise ConnectionError(f\"Couldn't reach the Hugging Face Hub for dataset '{path}': {e1}\") from None\r\nConnectionError: Couldn't reach the Hugging Face Hub for dataset 'C-MTEB/AFQMC': Offline mode is enabled.\r\n```\r\n\r\n",
"I was having similar inexplicable issues.\r\n\r\nDoing this I *think* helped, but, `datasets` still *clearly* does not want to respect the cache:\r\n\r\n```python\r\npip install --upgrade datasets # now it is 2.18.0\r\nHF_DATASETS_OFFLINE=\"1\" python blah.py\r\n```\r\n\r\nOr similarly, I must spacify that env var to resuse the cache, IE, no arg to `load_dataset` helps it reuse the cache:\r\n\r\n```python\r\n\r\nimport os\r\nos.environ[\"HF_DATASETS_OFFLINE\"] = \"1\"\r\n\r\nimport logging\r\nlogging.basicConfig(level=logging.DEBUG)\r\n\r\nimport datasets\r\n# >>> datasets.__version__\r\n# '2.18.0'\r\n\r\ndatasets.utils.logging.set_verbosity_info()\r\ndata = datasets.load_dataset(\"c-s-ale/dolly-15k-instruction-alpaca-format\")\r\n```"
] |
6,749
|
Fix fsspec tqdm callback
|
Following changes at https://github.com/fsspec/filesystem_spec/pull/1497 for `fsspec>=2024.2.0`
|
[] |
{
"url": "https://api.github.com/repos/huggingface/datasets/pulls/6749",
"html_url": "https://github.com/huggingface/datasets/pull/6749",
"diff_url": "https://github.com/huggingface/datasets/pull/6749.diff",
"patch_url": "https://github.com/huggingface/datasets/pull/6749.patch",
"merged_at": "2024-03-22T14:45:39"
}
| true
|
[
"The docs for this PR live [here](https://moon-ci-docs.huggingface.co/docs/datasets/pr_6749). All of your documentation changes will be reflected on that endpoint. The docs are available until 30 days after the last update.",
"<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) | 0.005017 / 0.011353 (-0.006336) | 0.002958 / 0.011008 (-0.008050) | 0.063455 / 0.038508 (0.024946) | 0.028206 / 0.023109 (0.005096) | 0.230884 / 0.275898 (-0.045014) | 0.252688 / 0.323480 (-0.070792) | 0.002995 / 0.007986 (-0.004991) | 0.002613 / 0.004328 (-0.001716) | 0.046477 / 0.004250 (0.042226) | 0.040662 / 0.037052 (0.003609) | 0.241824 / 0.258489 (-0.016665) | 0.269063 / 0.293841 (-0.024778) | 0.027336 / 0.128546 (-0.101210) | 0.010614 / 0.075646 (-0.065032) | 0.216087 / 0.419271 (-0.203184) | 0.035667 / 0.043533 (-0.007866) | 0.238657 / 0.255139 (-0.016482) | 0.253433 / 0.283200 (-0.029767) | 0.017433 / 0.141683 (-0.124250) | 1.120856 / 1.452155 (-0.331299) | 1.157415 / 1.492716 (-0.335302) |\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) | 0.088028 / 0.018006 (0.070022) | 0.277368 / 0.000490 (0.276878) | 0.000204 / 0.000200 (0.000004) | 0.000049 / 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) | 0.017956 / 0.037411 (-0.019455) | 0.061061 / 0.014526 (0.046535) | 0.073323 / 0.176557 (-0.103234) | 0.119254 / 0.737135 (-0.617881) | 0.074308 / 0.296338 (-0.222031) |\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) | 0.285118 / 0.215209 (0.069908) | 2.785796 / 2.077655 (0.708142) | 1.476436 / 1.504120 (-0.027684) | 1.356505 / 1.541195 (-0.184690) | 1.362505 / 1.468490 (-0.105985) | 0.554064 / 4.584777 (-4.030713) | 2.395774 / 3.745712 (-1.349938) | 2.713703 / 5.269862 (-2.556159) | 1.701020 / 4.565676 (-2.864657) | 0.062370 / 0.424275 (-0.361905) | 0.004944 / 0.007607 (-0.002663) | 0.327948 / 0.226044 (0.101904) | 3.243739 / 2.268929 (0.974811) | 1.803881 / 55.444624 (-53.640743) | 1.551635 / 6.876477 (-5.324841) | 1.560627 / 2.142072 (-0.581446) | 0.628187 / 4.805227 (-4.177040) | 0.115824 / 6.500664 (-6.384840) | 0.041655 / 0.075469 (-0.033814) |\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) | 0.968797 / 1.841788 (-0.872991) | 11.220905 / 8.074308 (3.146597) | 9.322584 / 10.191392 (-0.868808) | 0.139629 / 0.680424 (-0.540795) | 0.013823 / 0.534201 (-0.520378) | 0.286700 / 0.579283 (-0.292583) | 0.263517 / 0.434364 (-0.170847) | 0.341264 / 0.540337 (-0.199074) | 0.418834 / 1.386936 (-0.968102) |\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) | 0.005404 / 0.011353 (-0.005949) | 0.003630 / 0.011008 (-0.007378) | 0.048977 / 0.038508 (0.010469) | 0.029980 / 0.023109 (0.006871) | 0.274671 / 0.275898 (-0.001227) | 0.295671 / 0.323480 (-0.027808) | 0.004230 / 0.007986 (-0.003756) | 0.002656 / 0.004328 (-0.001672) | 0.048603 / 0.004250 (0.044353) | 0.044323 / 0.037052 (0.007271) | 0.286499 / 0.258489 (0.028010) | 0.313199 / 0.293841 (0.019358) | 0.030079 / 0.128546 (-0.098468) | 0.010480 / 0.075646 (-0.065166) | 0.058226 / 0.419271 (-0.361045) | 0.054920 / 0.043533 (0.011387) | 0.274921 / 0.255139 (0.019783) | 0.296559 / 0.283200 (0.013360) | 0.019164 / 0.141683 (-0.122519) | 1.154703 / 1.452155 (-0.297452) | 1.207015 / 1.492716 (-0.285701) |\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) | 0.089368 / 0.018006 (0.071362) | 0.301196 / 0.000490 (0.300706) | 0.000208 / 0.000200 (0.000008) | 0.000047 / 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) | 0.021355 / 0.037411 (-0.016056) | 0.074688 / 0.014526 (0.060162) | 0.085840 / 0.176557 (-0.090716) | 0.125784 / 0.737135 (-0.611351) | 0.087103 / 0.296338 (-0.209235) |\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) | 0.296727 / 0.215209 (0.081518) | 2.884922 / 2.077655 (0.807267) | 1.586515 / 1.504120 (0.082395) | 1.474417 / 1.541195 (-0.066777) | 1.492105 / 1.468490 (0.023615) | 0.570016 / 4.584777 (-4.014761) | 2.435760 / 3.745712 (-1.309952) | 2.657999 / 5.269862 (-2.611863) | 1.740160 / 4.565676 (-2.825516) | 0.063743 / 0.424275 (-0.360532) | 0.005048 / 0.007607 (-0.002559) | 0.341279 / 0.226044 (0.115235) | 3.396185 / 2.268929 (1.127256) | 1.952825 / 55.444624 (-53.491800) | 1.676669 / 6.876477 (-5.199808) | 1.773158 / 2.142072 (-0.368915) | 0.650664 / 4.805227 (-4.154563) | 0.116815 / 6.500664 (-6.383849) | 0.040813 / 0.075469 (-0.034656) |\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) | 0.999836 / 1.841788 (-0.841952) | 11.854540 / 8.074308 (3.780232) | 10.245516 / 10.191392 (0.054124) | 0.141235 / 0.680424 (-0.539189) | 0.015562 / 0.534201 (-0.518639) | 0.287556 / 0.579283 (-0.291727) | 0.274946 / 0.434364 (-0.159418) | 0.324652 / 0.540337 (-0.215685) | 0.449204 / 1.386936 (-0.937733) |\n\n</details>\n</details>\n\n![](https://cml.dev/watermark.png#ed2b406d045349dad16738985c947fe743260710 \"CML watermark\")\n"
] |
6,748
|
Strange slicing behavior
|
### Describe the bug
I have loaded a dataset, and then slice first 300 samples using `:` ops, however, the resulting dataset is not expected, as the output below:
```bash
len(dataset)=1050324
len(dataset[:300])=2
len(dataset[0:300])=2
len(dataset.select(range(300)))=300
```
### Steps to reproduce the bug
load a dataset then:
```bash
dataset = load_from_disk(args.train_data_dir)
print(f"{len(dataset)=}", flush=True)
print(f"{len(dataset[:300])=}", flush=True)
print(f"{len(dataset[0:300])=}", flush=True)
print(f"{len(dataset.select(range(300)))=}", flush=True)
```
### Expected behavior
```bash
len(dataset)=1050324
len(dataset[:300])=300
len(dataset[0:300])=300
len(dataset.select(range(300)))=300
```
### Environment info
- `datasets` version: 2.16.1
- Platform: Linux-5.15.0-60-generic-x86_64-with-glibc2.35
- Python version: 3.10.11
- `huggingface_hub` version: 0.20.2
- PyArrow version: 10.0.1
- Pandas version: 1.5.3
- `fsspec` version: 2023.10.0
|
[] |
{
"url": null,
"html_url": null,
"diff_url": null,
"patch_url": null,
"merged_at": null
}
| false
|
[
"As explained in the [docs](https://huggingface.co/docs/datasets/v2.18.0/en/access#slicing), slicing a `Dataset` returns a dictionary that maps its column names to their values. So, `len(dataset[:300])=2` is expected, assuming your dataset has 2 columns (the returned dict has 2 keys, but each value in the dict has 300 items).\r\n` "
] |
6,747
|
chore(deps): bump fsspec
|
There were a few fixes released recently, some DVC ecosystem packages require newer version of `fsspec`.
|
[] |
{
"url": "https://api.github.com/repos/huggingface/datasets/pulls/6747",
"html_url": "https://github.com/huggingface/datasets/pull/6747",
"diff_url": "https://github.com/huggingface/datasets/pull/6747.diff",
"patch_url": "https://github.com/huggingface/datasets/pull/6747.patch",
"merged_at": "2024-03-22T16:28:40"
}
| true
|
[
"The docs for this PR live [here](https://moon-ci-docs.huggingface.co/docs/datasets/pr_6747). All of your documentation changes will be reflected on that endpoint. The docs are available until 30 days after the last update.",
"<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) | 0.005129 / 0.011353 (-0.006224) | 0.003788 / 0.011008 (-0.007220) | 0.063456 / 0.038508 (0.024948) | 0.029079 / 0.023109 (0.005969) | 0.237228 / 0.275898 (-0.038670) | 0.260554 / 0.323480 (-0.062926) | 0.003090 / 0.007986 (-0.004895) | 0.002730 / 0.004328 (-0.001599) | 0.049040 / 0.004250 (0.044789) | 0.042432 / 0.037052 (0.005380) | 0.256954 / 0.258489 (-0.001535) | 0.285912 / 0.293841 (-0.007929) | 0.027568 / 0.128546 (-0.100978) | 0.010402 / 0.075646 (-0.065245) | 0.206773 / 0.419271 (-0.212499) | 0.035381 / 0.043533 (-0.008152) | 0.243147 / 0.255139 (-0.011992) | 0.259419 / 0.283200 (-0.023781) | 0.019503 / 0.141683 (-0.122180) | 1.145537 / 1.452155 (-0.306618) | 1.204070 / 1.492716 (-0.288646) |\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) | 0.092298 / 0.018006 (0.074291) | 0.300042 / 0.000490 (0.299553) | 0.000236 / 0.000200 (0.000036) | 0.000052 / 0.000054 (-0.000002) |\n\n### Benchmark: benchmark_indices_mapping.json\n\n| metric | select | shard | shuffle | sort | train_test_split |\n|--------|---|---|---|---|---|\n| new / old (diff) | 0.018624 / 0.037411 (-0.018788) | 0.063832 / 0.014526 (0.049306) | 0.075849 / 0.176557 (-0.100707) | 0.120919 / 0.737135 (-0.616216) | 0.075878 / 0.296338 (-0.220461) |\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) | 0.275545 / 0.215209 (0.060336) | 2.706004 / 2.077655 (0.628349) | 1.406398 / 1.504120 (-0.097722) | 1.287154 / 1.541195 (-0.254041) | 1.298278 / 1.468490 (-0.170212) | 0.559763 / 4.584777 (-4.025014) | 2.434104 / 3.745712 (-1.311608) | 2.786338 / 5.269862 (-2.483523) | 1.720951 / 4.565676 (-2.844726) | 0.062082 / 0.424275 (-0.362193) | 0.004931 / 0.007607 (-0.002676) | 0.329998 / 0.226044 (0.103954) | 3.222105 / 2.268929 (0.953176) | 1.777539 / 55.444624 (-53.667085) | 1.533845 / 6.876477 (-5.342632) | 1.520357 / 2.142072 (-0.621715) | 0.638850 / 4.805227 (-4.166377) | 0.116718 / 6.500664 (-6.383946) | 0.042215 / 0.075469 (-0.033254) |\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) | 0.962791 / 1.841788 (-0.878997) | 11.509889 / 8.074308 (3.435581) | 9.507676 / 10.191392 (-0.683716) | 0.140780 / 0.680424 (-0.539644) | 0.014187 / 0.534201 (-0.520014) | 0.286363 / 0.579283 (-0.292920) | 0.263316 / 0.434364 (-0.171048) | 0.322099 / 0.540337 (-0.218239) | 0.415602 / 1.386936 (-0.971334) |\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) | 0.005175 / 0.011353 (-0.006178) | 0.003631 / 0.011008 (-0.007377) | 0.050277 / 0.038508 (0.011769) | 0.031879 / 0.023109 (0.008770) | 0.269966 / 0.275898 (-0.005933) | 0.297229 / 0.323480 (-0.026251) | 0.004278 / 0.007986 (-0.003707) | 0.002936 / 0.004328 (-0.001393) | 0.048686 / 0.004250 (0.044436) | 0.044262 / 0.037052 (0.007209) | 0.284578 / 0.258489 (0.026089) | 0.313681 / 0.293841 (0.019840) | 0.029064 / 0.128546 (-0.099482) | 0.010700 / 0.075646 (-0.064946) | 0.058366 / 0.419271 (-0.360905) | 0.051341 / 0.043533 (0.007809) | 0.271262 / 0.255139 (0.016123) | 0.290791 / 0.283200 (0.007591) | 0.019044 / 0.141683 (-0.122639) | 1.149514 / 1.452155 (-0.302641) | 1.209277 / 1.492716 (-0.283439) |\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) | 0.094879 / 0.018006 (0.076872) | 0.302196 / 0.000490 (0.301707) | 0.000217 / 0.000200 (0.000018) | 0.000052 / 0.000054 (-0.000002) |\n\n### Benchmark: benchmark_indices_mapping.json\n\n| metric | select | shard | shuffle | sort | train_test_split |\n|--------|---|---|---|---|---|\n| new / old (diff) | 0.021715 / 0.037411 (-0.015696) | 0.075122 / 0.014526 (0.060596) | 0.087393 / 0.176557 (-0.089164) | 0.125583 / 0.737135 (-0.611553) | 0.088722 / 0.296338 (-0.207617) |\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) | 0.295158 / 0.215209 (0.079949) | 2.930208 / 2.077655 (0.852553) | 1.590197 / 1.504120 (0.086077) | 1.459038 / 1.541195 (-0.082156) | 1.471690 / 1.468490 (0.003200) | 0.570279 / 4.584777 (-4.014498) | 2.456971 / 3.745712 (-1.288741) | 2.675315 / 5.269862 (-2.594547) | 1.750122 / 4.565676 (-2.815554) | 0.062905 / 0.424275 (-0.361370) | 0.005118 / 0.007607 (-0.002489) | 0.344263 / 0.226044 (0.118219) | 3.472460 / 2.268929 (1.203532) | 1.931707 / 55.444624 (-53.512917) | 1.658537 / 6.876477 (-5.217939) | 1.785794 / 2.142072 (-0.356278) | 0.637149 / 4.805227 (-4.168078) | 0.115838 / 6.500664 (-6.384826) | 0.040771 / 0.075469 (-0.034698) |\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) | 1.002869 / 1.841788 (-0.838919) | 12.048825 / 8.074308 (3.974517) | 10.407979 / 10.191392 (0.216587) | 0.150300 / 0.680424 (-0.530124) | 0.015299 / 0.534201 (-0.518902) | 0.286277 / 0.579283 (-0.293006) | 0.312186 / 0.434364 (-0.122178) | 0.322633 / 0.540337 (-0.217704) | 0.438431 / 1.386936 (-0.948505) |\n\n</details>\n</details>\n\n![](https://cml.dev/watermark.png#d5468836fe94e8be1ae093397dd43d4a2503b926 \"CML watermark\")\n"
] |
6,746
|
ExpectedMoreSplits error when loading C4 dataset
|
### Describe the bug
I encounter bug when running the example command line
```python
python main.py \
--model decapoda-research/llama-7b-hf \
--prune_method wanda \
--sparsity_ratio 0.5 \
--sparsity_type unstructured \
--save out/llama_7b/unstructured/wanda/
```
The bug occurred at these lines of code (when loading c4 dataset)
```python
traindata = load_dataset('allenai/c4', 'allenai--c4', data_files={'train': 'en/c4-train.00000-of-01024.json.gz'}, split='train')
valdata = load_dataset('allenai/c4', 'allenai--c4', data_files={'validation': 'en/c4-validation.00000-of-00008.json.gz'}, split='validation')
```
The error message states:
```
raise ExpectedMoreSplits(str(set(expected_splits) - set(recorded_splits)))
datasets.utils.info_utils.ExpectedMoreSplits: {'validation'}
```
### Steps to reproduce the bug
1. I encounter bug when running the example command line
### Expected behavior
The error message states:
```
raise ExpectedMoreSplits(str(set(expected_splits) - set(recorded_splits)))
datasets.utils.info_utils.ExpectedMoreSplits: {'validation'}
```
### Environment info
I'm using cuda 12.4, so I use ```pip install pytorch``` instead of conda provided in install.md
Also, I've tried another environment using the same commands in install.md, but the same bug occured
|
[] |
{
"url": null,
"html_url": null,
"diff_url": null,
"patch_url": null,
"merged_at": null
}
| false
|
[
"Hi ! We updated the `allenai/c4` repository to allow people to specify which language to load easily (the the [c4 dataset page](https://huggingface.co/datasets/allenai/c4))\r\n\r\nTo fix this issue **you can update** `datasets` and remove the mention of the legacy configuration name \"allenai--c4\":\r\n\r\n```python\r\ntraindata = load_dataset('allenai/c4', data_files={'train': 'en/c4-train.00000-of-01024.json.gz'}, split='train')\r\nvaldata = load_dataset('allenai/c4', data_files={'validation': 'en/c4-validation.00000-of-00008.json.gz'}, split='validation')\r\n```",
"Did you solve this problem？I have the same bug.It is no use to delete \"allenai--c4\".",
"Did you solve it? I met this problem too.",
"But after I romove allenai--c4,it still fails",
"For me it works this way. I'm using datasets version 2.17.0",
"First, pip install --upgrade datasets.\r\nSecond, Update the following two lines of code in data.py (in lib)\r\ntraindata = load_dataset('allenai/c4', data_files={'train': 'en/c4-train.00000-of-01024.json.gz'}, split='train')\r\nvaldata = load_dataset('allenai/c4', data_files={'validation': 'en/c4-validation.00000-of-00008.json.gz'}, split='validation')",
"The error is in the Wanda repository: https://github.com/locuslab/wanda\r\n- https://github.com/locuslab/wanda/issues/57\r\n\r\nConcretely, in these code lines:\r\nhttps://github.com/locuslab/wanda/blob/8e8fc87b4a2f9955baa7e76e64d5fce7fa8724a6/lib/data.py#L43-L44\r\n\r\nPlease report there and/or make the fix in their code."
] |
6,745
|
Scraping the whole of github including private repos is bad; kindly stop
|
### Feature request
https://github.com/bigcode-project/opt-out-v2 - opt out is not consent. kindly quit this ridiculous nonsense.
### Motivation
[EDITED: insults not tolerated]
### Your contribution
[EDITED: insults not tolerated]
|
[
{
"id": 1935892871,
"node_id": "MDU6TGFiZWwxOTM1ODkyODcx",
"url": "https://api.github.com/repos/huggingface/datasets/labels/enhancement",
"name": "enhancement",
"color": "a2eeef",
"default": true,
"description": "New feature or request"
}
] |
{
"url": null,
"html_url": null,
"diff_url": null,
"patch_url": null,
"merged_at": null
}
| false
|
[
"It's not twitter here"
] |
6,744
|
Option to disable file locking
|
### Feature request
Commands such as `load_dataset` creates file locks with `filelock.FileLock`. It would be good if there was a way to disable this.
### Motivation
File locking doesn't work on all file-systems (in my case NFS mounted Weka). If the `cache_dir` only had small files then it would be possible to point to local disk and the problem would be solved. However, as cache_dir is both where the small info files are written and the processed datasets are put this isn't a feasible solution.
Considering https://github.com/huggingface/datasets/issues/6395 I still do think this is something that belongs in HuggingFace. The possibility to control packages separately is valuable. It might be that a user has their dataset on a file-system that doesn't support file-locking while they are using file locking on local disk to control some other type of access.
### Your contribution
My suggested solution:
```
diff --git a/src/datasets/utils/_filelock.py b/src/datasets/utils/_filelock.py
index 19620e6e..58f41a02 100644
--- a/src/datasets/utils/_filelock.py
+++ b/src/datasets/utils/_filelock.py
@@ -18,11 +18,15 @@
import os
from filelock import FileLock as FileLock_
-from filelock import UnixFileLock
+from filelock import SoftFileLock, UnixFileLock
from filelock import __version__ as _filelock_version
from packaging import version
+if os.getenv('HF_USE_SOFTFILELOCK', 'false').lower() in ('true', '1'):
+ FileLock_ = SoftFileLock
+
+
class FileLock(FileLock_):
"""
A `filelock.FileLock` initializer that handles long paths.
```
|
[
{
"id": 1935892871,
"node_id": "MDU6TGFiZWwxOTM1ODkyODcx",
"url": "https://api.github.com/repos/huggingface/datasets/labels/enhancement",
"name": "enhancement",
"color": "a2eeef",
"default": true,
"description": "New feature or request"
}
] |
{
"url": null,
"html_url": null,
"diff_url": null,
"patch_url": null,
"merged_at": null
}
| false
|
[] |
6,743
|
Allow null values in dict columns
|
Fix #6738
|
[] |
{
"url": "https://api.github.com/repos/huggingface/datasets/pulls/6743",
"html_url": "https://github.com/huggingface/datasets/pull/6743",
"diff_url": "https://github.com/huggingface/datasets/pull/6743.diff",
"patch_url": "https://github.com/huggingface/datasets/pull/6743.patch",
"merged_at": "2024-03-19T20:05:19"
}
| true
|
[
"The docs for this PR live [here](https://moon-ci-docs.huggingface.co/docs/datasets/pr_6743). All of your documentation changes will be reflected on that endpoint. The docs are available until 30 days after the last update.",
"<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) | 0.005013 / 0.011353 (-0.006340) | 0.003228 / 0.011008 (-0.007780) | 0.062763 / 0.038508 (0.024255) | 0.028937 / 0.023109 (0.005828) | 0.240777 / 0.275898 (-0.035121) | 0.266972 / 0.323480 (-0.056508) | 0.003073 / 0.007986 (-0.004913) | 0.002769 / 0.004328 (-0.001560) | 0.049265 / 0.004250 (0.045015) | 0.042061 / 0.037052 (0.005009) | 0.261714 / 0.258489 (0.003225) | 0.284896 / 0.293841 (-0.008944) | 0.027717 / 0.128546 (-0.100829) | 0.010430 / 0.075646 (-0.065216) | 0.209022 / 0.419271 (-0.210249) | 0.035941 / 0.043533 (-0.007591) | 0.246849 / 0.255139 (-0.008290) | 0.263205 / 0.283200 (-0.019994) | 0.019489 / 0.141683 (-0.122193) | 1.102595 / 1.452155 (-0.349559) | 1.170493 / 1.492716 (-0.322223) |\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) | 0.093611 / 0.018006 (0.075604) | 0.302041 / 0.000490 (0.301551) | 0.000223 / 0.000200 (0.000023) | 0.000052 / 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) | 0.018720 / 0.037411 (-0.018692) | 0.062199 / 0.014526 (0.047673) | 0.074888 / 0.176557 (-0.101669) | 0.120184 / 0.737135 (-0.616951) | 0.076756 / 0.296338 (-0.219583) |\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) | 0.287484 / 0.215209 (0.072275) | 2.787777 / 2.077655 (0.710123) | 1.488957 / 1.504120 (-0.015163) | 1.362678 / 1.541195 (-0.178517) | 1.364571 / 1.468490 (-0.103919) | 0.563139 / 4.584777 (-4.021638) | 2.422224 / 3.745712 (-1.323488) | 2.798011 / 5.269862 (-2.471850) | 1.751159 / 4.565676 (-2.814517) | 0.062740 / 0.424275 (-0.361536) | 0.004918 / 0.007607 (-0.002689) | 0.338285 / 0.226044 (0.112240) | 3.316012 / 2.268929 (1.047083) | 1.845975 / 55.444624 (-53.598650) | 1.553187 / 6.876477 (-5.323290) | 1.564582 / 2.142072 (-0.577490) | 0.645987 / 4.805227 (-4.159240) | 0.118216 / 6.500664 (-6.382448) | 0.041243 / 0.075469 (-0.034226) |\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) | 0.970265 / 1.841788 (-0.871522) | 11.783152 / 8.074308 (3.708844) | 9.516584 / 10.191392 (-0.674808) | 0.148086 / 0.680424 (-0.532338) | 0.013689 / 0.534201 (-0.520512) | 0.289657 / 0.579283 (-0.289626) | 0.265966 / 0.434364 (-0.168398) | 0.328483 / 0.540337 (-0.211854) | 0.433544 / 1.386936 (-0.953392) |\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) | 0.005235 / 0.011353 (-0.006118) | 0.003515 / 0.011008 (-0.007493) | 0.049484 / 0.038508 (0.010976) | 0.029264 / 0.023109 (0.006154) | 0.278518 / 0.275898 (0.002620) | 0.298948 / 0.323480 (-0.024532) | 0.004308 / 0.007986 (-0.003678) | 0.002751 / 0.004328 (-0.001577) | 0.048952 / 0.004250 (0.044701) | 0.045379 / 0.037052 (0.008327) | 0.292633 / 0.258489 (0.034144) | 0.319405 / 0.293841 (0.025564) | 0.030201 / 0.128546 (-0.098345) | 0.010657 / 0.075646 (-0.064990) | 0.057842 / 0.419271 (-0.361430) | 0.053359 / 0.043533 (0.009826) | 0.281136 / 0.255139 (0.025997) | 0.295388 / 0.283200 (0.012188) | 0.018786 / 0.141683 (-0.122897) | 1.187181 / 1.452155 (-0.264974) | 1.198394 / 1.492716 (-0.294323) |\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) | 0.093861 / 0.018006 (0.075855) | 0.304019 / 0.000490 (0.303529) | 0.000220 / 0.000200 (0.000020) | 0.000053 / 0.000054 (-0.000002) |\n\n### Benchmark: benchmark_indices_mapping.json\n\n| metric | select | shard | shuffle | sort | train_test_split |\n|--------|---|---|---|---|---|\n| new / old (diff) | 0.021582 / 0.037411 (-0.015829) | 0.075381 / 0.014526 (0.060855) | 0.087886 / 0.176557 (-0.088671) | 0.125078 / 0.737135 (-0.612057) | 0.089339 / 0.296338 (-0.206999) |\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) | 0.295797 / 0.215209 (0.080588) | 2.912021 / 2.077655 (0.834367) | 1.592191 / 1.504120 (0.088071) | 1.471270 / 1.541195 (-0.069925) | 1.475535 / 1.468490 (0.007045) | 0.564114 / 4.584777 (-4.020663) | 2.442882 / 3.745712 (-1.302830) | 2.679433 / 5.269862 (-2.590428) | 1.752097 / 4.565676 (-2.813579) | 0.062748 / 0.424275 (-0.361527) | 0.005068 / 0.007607 (-0.002539) | 0.345554 / 0.226044 (0.119509) | 3.456929 / 2.268929 (1.188000) | 1.962781 / 55.444624 (-53.481844) | 1.688313 / 6.876477 (-5.188164) | 1.817392 / 2.142072 (-0.324681) | 0.639588 / 4.805227 (-4.165639) | 0.116148 / 6.500664 (-6.384516) | 0.040851 / 0.075469 (-0.034618) |\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) | 1.009852 / 1.841788 (-0.831936) | 12.031749 / 8.074308 (3.957440) | 10.305107 / 10.191392 (0.113715) | 0.132960 / 0.680424 (-0.547464) | 0.014779 / 0.534201 (-0.519422) | 0.288903 / 0.579283 (-0.290381) | 0.275417 / 0.434364 (-0.158947) | 0.322628 / 0.540337 (-0.217709) | 0.445060 / 1.386936 (-0.941876) |\n\n</details>\n</details>\n\n![](https://cml.dev/watermark.png#f234fce40d5ffc96fac5198d8cc89817970d87ee \"CML watermark\")\n",
"notify https://huggingface.co/datasets/chaoyi-wu/PMC-Inline/discussions/1 once it's merged in dataset-viewer"
] |
6,742
|
Fix missing download_config in get_data_patterns
|
Reported in https://github.com/huggingface/datasets-server/issues/2607
|
[] |
{
"url": "https://api.github.com/repos/huggingface/datasets/pulls/6742",
"html_url": "https://github.com/huggingface/datasets/pull/6742",
"diff_url": "https://github.com/huggingface/datasets/pull/6742.diff",
"patch_url": "https://github.com/huggingface/datasets/pull/6742.patch",
"merged_at": "2024-03-19T18:15:13"
}
| true
|
[
"The docs for this PR live [here](https://moon-ci-docs.huggingface.co/docs/datasets/pr_6742). All of your documentation changes will be reflected on that endpoint. The docs are available until 30 days after the last update.",
"<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) | 0.005394 / 0.011353 (-0.005959) | 0.003780 / 0.011008 (-0.007228) | 0.063459 / 0.038508 (0.024951) | 0.028883 / 0.023109 (0.005774) | 0.239159 / 0.275898 (-0.036739) | 0.258123 / 0.323480 (-0.065357) | 0.003134 / 0.007986 (-0.004851) | 0.003452 / 0.004328 (-0.000876) | 0.049255 / 0.004250 (0.045005) | 0.042727 / 0.037052 (0.005675) | 0.257387 / 0.258489 (-0.001102) | 0.280762 / 0.293841 (-0.013079) | 0.027921 / 0.128546 (-0.100625) | 0.010867 / 0.075646 (-0.064779) | 0.207878 / 0.419271 (-0.211393) | 0.036003 / 0.043533 (-0.007530) | 0.247457 / 0.255139 (-0.007682) | 0.260231 / 0.283200 (-0.022969) | 0.019741 / 0.141683 (-0.121942) | 1.143645 / 1.452155 (-0.308510) | 1.188789 / 1.492716 (-0.303927) |\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) | 0.092065 / 0.018006 (0.074059) | 0.286021 / 0.000490 (0.285531) | 0.000220 / 0.000200 (0.000020) | 0.000048 / 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) | 0.018934 / 0.037411 (-0.018477) | 0.062474 / 0.014526 (0.047949) | 0.073384 / 0.176557 (-0.103172) | 0.121276 / 0.737135 (-0.615860) | 0.077792 / 0.296338 (-0.218546) |\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) | 0.285352 / 0.215209 (0.070143) | 2.783110 / 2.077655 (0.705456) | 1.487983 / 1.504120 (-0.016137) | 1.364264 / 1.541195 (-0.176930) | 1.388757 / 1.468490 (-0.079733) | 0.568347 / 4.584777 (-4.016430) | 2.402451 / 3.745712 (-1.343261) | 2.835577 / 5.269862 (-2.434285) | 1.754853 / 4.565676 (-2.810824) | 0.063355 / 0.424275 (-0.360920) | 0.005010 / 0.007607 (-0.002598) | 0.332061 / 0.226044 (0.106016) | 3.287121 / 2.268929 (1.018193) | 1.829520 / 55.444624 (-53.615104) | 1.542669 / 6.876477 (-5.333808) | 1.560679 / 2.142072 (-0.581393) | 0.642371 / 4.805227 (-4.162856) | 0.118636 / 6.500664 (-6.382028) | 0.042262 / 0.075469 (-0.033207) |\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) | 0.984803 / 1.841788 (-0.856985) | 11.578044 / 8.074308 (3.503735) | 9.383428 / 10.191392 (-0.807964) | 0.141367 / 0.680424 (-0.539057) | 0.014047 / 0.534201 (-0.520154) | 0.291505 / 0.579283 (-0.287778) | 0.270199 / 0.434364 (-0.164165) | 0.329874 / 0.540337 (-0.210463) | 0.429386 / 1.386936 (-0.957550) |\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) | 0.005322 / 0.011353 (-0.006031) | 0.004023 / 0.011008 (-0.006986) | 0.050126 / 0.038508 (0.011618) | 0.029937 / 0.023109 (0.006828) | 0.275985 / 0.275898 (0.000087) | 0.297965 / 0.323480 (-0.025515) | 0.004429 / 0.007986 (-0.003557) | 0.002729 / 0.004328 (-0.001599) | 0.048995 / 0.004250 (0.044744) | 0.044940 / 0.037052 (0.007888) | 0.288397 / 0.258489 (0.029908) | 0.317716 / 0.293841 (0.023875) | 0.029705 / 0.128546 (-0.098841) | 0.010972 / 0.075646 (-0.064674) | 0.058592 / 0.419271 (-0.360680) | 0.054640 / 0.043533 (0.011108) | 0.276456 / 0.255139 (0.021317) | 0.295119 / 0.283200 (0.011919) | 0.020032 / 0.141683 (-0.121651) | 1.175740 / 1.452155 (-0.276415) | 1.227246 / 1.492716 (-0.265471) |\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) | 0.092204 / 0.018006 (0.074197) | 0.300344 / 0.000490 (0.299855) | 0.000213 / 0.000200 (0.000013) | 0.000050 / 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) | 0.021540 / 0.037411 (-0.015871) | 0.076252 / 0.014526 (0.061726) | 0.087582 / 0.176557 (-0.088975) | 0.125977 / 0.737135 (-0.611159) | 0.090649 / 0.296338 (-0.205689) |\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) | 0.294544 / 0.215209 (0.079335) | 2.883736 / 2.077655 (0.806082) | 1.570932 / 1.504120 (0.066812) | 1.449082 / 1.541195 (-0.092113) | 1.463262 / 1.468490 (-0.005228) | 0.559625 / 4.584777 (-4.025152) | 2.448593 / 3.745712 (-1.297119) | 2.663857 / 5.269862 (-2.606005) | 1.757812 / 4.565676 (-2.807865) | 0.061999 / 0.424275 (-0.362276) | 0.005100 / 0.007607 (-0.002507) | 0.343620 / 0.226044 (0.117575) | 3.487059 / 2.268929 (1.218130) | 1.963078 / 55.444624 (-53.481546) | 1.661758 / 6.876477 (-5.214719) | 1.799130 / 2.142072 (-0.342942) | 0.650194 / 4.805227 (-4.155034) | 0.117375 / 6.500664 (-6.383289) | 0.040957 / 0.075469 (-0.034512) |\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) | 1.037882 / 1.841788 (-0.803906) | 12.239784 / 8.074308 (4.165476) | 10.478186 / 10.191392 (0.286794) | 0.164446 / 0.680424 (-0.515978) | 0.014901 / 0.534201 (-0.519300) | 0.302485 / 0.579283 (-0.276798) | 0.283994 / 0.434364 (-0.150370) | 0.338473 / 0.540337 (-0.201864) | 0.468901 / 1.386936 (-0.918035) |\n\n</details>\n</details>\n\n![](https://cml.dev/watermark.png#5fa934e275d240d9b1228b2f598bc96390299339 \"CML watermark\")\n"
] |
6,741
|
Fix offline mode with single config
|
Reported in https://github.com/huggingface/datasets/issues/4760
The cache was not able to reload a dataset with a single config form the cache if the config name is not specificed
For example
```python
from datasets import load_dataset, config
config.HF_DATASETS_OFFLINE = True
load_dataset("openai_humaneval")
```
This was due to a regression in https://github.com/huggingface/datasets/pull/6632
|
[] |
{
"url": "https://api.github.com/repos/huggingface/datasets/pulls/6741",
"html_url": "https://github.com/huggingface/datasets/pull/6741",
"diff_url": "https://github.com/huggingface/datasets/pull/6741.diff",
"patch_url": "https://github.com/huggingface/datasets/pull/6741.patch",
"merged_at": "2024-03-25T16:23:59"
}
| true
|
[
"The docs for this PR live [here](https://moon-ci-docs.huggingface.co/docs/datasets/pr_6741). All of your documentation changes will be reflected on that endpoint. The docs are available until 30 days after the last update.",
"<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) | 0.005093 / 0.011353 (-0.006260) | 0.003317 / 0.011008 (-0.007692) | 0.064795 / 0.038508 (0.026287) | 0.030373 / 0.023109 (0.007263) | 0.258776 / 0.275898 (-0.017122) | 0.269768 / 0.323480 (-0.053711) | 0.004186 / 0.007986 (-0.003799) | 0.002630 / 0.004328 (-0.001699) | 0.048643 / 0.004250 (0.044392) | 0.044220 / 0.037052 (0.007168) | 0.265113 / 0.258489 (0.006624) | 0.292202 / 0.293841 (-0.001639) | 0.027468 / 0.128546 (-0.101079) | 0.010123 / 0.075646 (-0.065523) | 0.226869 / 0.419271 (-0.192402) | 0.035739 / 0.043533 (-0.007794) | 0.253193 / 0.255139 (-0.001946) | 0.271002 / 0.283200 (-0.012198) | 0.017201 / 0.141683 (-0.124482) | 1.105836 / 1.452155 (-0.346318) | 1.161559 / 1.492716 (-0.331158) |\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) | 0.090481 / 0.018006 (0.072475) | 0.299013 / 0.000490 (0.298524) | 0.000220 / 0.000200 (0.000020) | 0.000047 / 0.000054 (-0.000007) |\n\n### Benchmark: benchmark_indices_mapping.json\n\n| metric | select | shard | shuffle | sort | train_test_split |\n|--------|---|---|---|---|---|\n| new / old (diff) | 0.017684 / 0.037411 (-0.019727) | 0.061580 / 0.014526 (0.047054) | 0.074370 / 0.176557 (-0.102186) | 0.119468 / 0.737135 (-0.617667) | 0.074671 / 0.296338 (-0.221668) |\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) | 0.284778 / 0.215209 (0.069569) | 2.780241 / 2.077655 (0.702586) | 1.504025 / 1.504120 (-0.000095) | 1.386644 / 1.541195 (-0.154550) | 1.402038 / 1.468490 (-0.066452) | 0.555180 / 4.584777 (-4.029597) | 2.410973 / 3.745712 (-1.334740) | 2.773252 / 5.269862 (-2.496610) | 1.722784 / 4.565676 (-2.842892) | 0.062773 / 0.424275 (-0.361502) | 0.004959 / 0.007607 (-0.002648) | 0.337163 / 0.226044 (0.111119) | 3.356947 / 2.268929 (1.088019) | 1.880953 / 55.444624 (-53.563671) | 1.556049 / 6.876477 (-5.320427) | 1.578589 / 2.142072 (-0.563483) | 0.641993 / 4.805227 (-4.163234) | 0.118624 / 6.500664 (-6.382040) | 0.042202 / 0.075469 (-0.033268) |\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) | 0.995321 / 1.841788 (-0.846467) | 12.257597 / 8.074308 (4.183289) | 9.646214 / 10.191392 (-0.545178) | 0.131124 / 0.680424 (-0.549300) | 0.014119 / 0.534201 (-0.520082) | 0.287597 / 0.579283 (-0.291686) | 0.266983 / 0.434364 (-0.167381) | 0.328165 / 0.540337 (-0.212173) | 0.422405 / 1.386936 (-0.964531) |\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) | 0.005091 / 0.011353 (-0.006262) | 0.003358 / 0.011008 (-0.007650) | 0.049136 / 0.038508 (0.010628) | 0.031075 / 0.023109 (0.007966) | 0.275047 / 0.275898 (-0.000851) | 0.296845 / 0.323480 (-0.026635) | 0.004949 / 0.007986 (-0.003037) | 0.002586 / 0.004328 (-0.001743) | 0.048164 / 0.004250 (0.043913) | 0.040754 / 0.037052 (0.003702) | 0.288715 / 0.258489 (0.030226) | 0.312383 / 0.293841 (0.018542) | 0.029372 / 0.128546 (-0.099174) | 0.010097 / 0.075646 (-0.065549) | 0.056752 / 0.419271 (-0.362520) | 0.033128 / 0.043533 (-0.010405) | 0.274986 / 0.255139 (0.019847) | 0.292692 / 0.283200 (0.009493) | 0.018309 / 0.141683 (-0.123374) | 1.190320 / 1.452155 (-0.261834) | 1.222529 / 1.492716 (-0.270188) |\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) | 0.091717 / 0.018006 (0.073711) | 0.300278 / 0.000490 (0.299788) | 0.000217 / 0.000200 (0.000017) | 0.000065 / 0.000054 (0.000011) |\n\n### Benchmark: benchmark_indices_mapping.json\n\n| metric | select | shard | shuffle | sort | train_test_split |\n|--------|---|---|---|---|---|\n| new / old (diff) | 0.021394 / 0.037411 (-0.016018) | 0.074918 / 0.014526 (0.060392) | 0.087461 / 0.176557 (-0.089095) | 0.125499 / 0.737135 (-0.611636) | 0.087484 / 0.296338 (-0.208854) |\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) | 0.296557 / 0.215209 (0.081348) | 2.905527 / 2.077655 (0.827872) | 1.624640 / 1.504120 (0.120520) | 1.505495 / 1.541195 (-0.035700) | 1.514066 / 1.468490 (0.045576) | 0.569376 / 4.584777 (-4.015401) | 2.448575 / 3.745712 (-1.297137) | 2.772805 / 5.269862 (-2.497057) | 1.757287 / 4.565676 (-2.808390) | 0.064209 / 0.424275 (-0.360066) | 0.005688 / 0.007607 (-0.001919) | 0.353175 / 0.226044 (0.127131) | 3.481591 / 2.268929 (1.212662) | 1.995384 / 55.444624 (-53.449240) | 1.684623 / 6.876477 (-5.191854) | 1.675750 / 2.142072 (-0.466323) | 0.644463 / 4.805227 (-4.160764) | 0.115393 / 6.500664 (-6.385271) | 0.040671 / 0.075469 (-0.034799) |\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) | 1.037487 / 1.841788 (-0.804301) | 11.902194 / 8.074308 (3.827886) | 10.148579 / 10.191392 (-0.042813) | 0.150261 / 0.680424 (-0.530163) | 0.015001 / 0.534201 (-0.519200) | 0.291008 / 0.579283 (-0.288275) | 0.278758 / 0.434364 (-0.155606) | 0.334037 / 0.540337 (-0.206301) | 0.419942 / 1.386936 (-0.966994) |\n\n</details>\n</details>\n\n![](https://cml.dev/watermark.png#dcd01046388fc052d37acc5a450bea69e3c57afc \"CML watermark\")\n"
] |
6,740
|
Support for loading geotiff files as a part of the ImageFolder
|
### Feature request
Request for adding rasterio support to load geotiff as a part of ImageFolder, instead of using PIL
### Motivation
As of now, there are many datasets in HuggingFace Hub which are predominantly focussed towards RemoteSensing or are from RemoteSensing. The current ImageFolder (if I have understood correctly) uses PIL. This is not really optimized because mostly these datasets have images with many channels and additional metadata. Using PIL makes one loose it unless we provide a custom script. Hence, maybe an API could be added to have this in common?
### Your contribution
If the issue is accepted - i can contribute the code, because I would like to have it automated and generalised.
|
[
{
"id": 1935892871,
"node_id": "MDU6TGFiZWwxOTM1ODkyODcx",
"url": "https://api.github.com/repos/huggingface/datasets/labels/enhancement",
"name": "enhancement",
"color": "a2eeef",
"default": true,
"description": "New feature or request"
}
] |
{
"url": null,
"html_url": null,
"diff_url": null,
"patch_url": null,
"merged_at": null
}
| false
|
[] |
6,739
|
Transpose images with EXIF Orientation tag
|
Closes https://github.com/huggingface/datasets/issues/6252
|
[] |
{
"url": "https://api.github.com/repos/huggingface/datasets/pulls/6739",
"html_url": "https://github.com/huggingface/datasets/pull/6739",
"diff_url": "https://github.com/huggingface/datasets/pull/6739.diff",
"patch_url": "https://github.com/huggingface/datasets/pull/6739.patch",
"merged_at": "2024-03-19T15:29:41"
}
| true
|
[
"The docs for this PR live [here](https://moon-ci-docs.huggingface.co/docs/datasets/pr_6739). All of your documentation changes will be reflected on that endpoint. The docs are available until 30 days after the last update.",
"<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) | 0.005295 / 0.011353 (-0.006058) | 0.003402 / 0.011008 (-0.007606) | 0.062860 / 0.038508 (0.024352) | 0.029627 / 0.023109 (0.006518) | 0.238359 / 0.275898 (-0.037539) | 0.262940 / 0.323480 (-0.060540) | 0.003077 / 0.007986 (-0.004909) | 0.002676 / 0.004328 (-0.001652) | 0.048731 / 0.004250 (0.044480) | 0.043989 / 0.037052 (0.006936) | 0.255702 / 0.258489 (-0.002787) | 0.282667 / 0.293841 (-0.011174) | 0.028019 / 0.128546 (-0.100527) | 0.010195 / 0.075646 (-0.065451) | 0.205472 / 0.419271 (-0.213800) | 0.036551 / 0.043533 (-0.006982) | 0.243282 / 0.255139 (-0.011857) | 0.261925 / 0.283200 (-0.021274) | 0.020506 / 0.141683 (-0.121177) | 1.137228 / 1.452155 (-0.314927) | 1.183935 / 1.492716 (-0.308782) |\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) | 0.100290 / 0.018006 (0.082284) | 0.316279 / 0.000490 (0.315790) | 0.000239 / 0.000200 (0.000039) | 0.000043 / 0.000054 (-0.000011) |\n\n### Benchmark: benchmark_indices_mapping.json\n\n| metric | select | shard | shuffle | sort | train_test_split |\n|--------|---|---|---|---|---|\n| new / old (diff) | 0.017979 / 0.037411 (-0.019432) | 0.061616 / 0.014526 (0.047090) | 0.072989 / 0.176557 (-0.103568) | 0.118667 / 0.737135 (-0.618468) | 0.074266 / 0.296338 (-0.222072) |\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) | 0.287971 / 0.215209 (0.072762) | 2.845235 / 2.077655 (0.767581) | 1.501983 / 1.504120 (-0.002137) | 1.389824 / 1.541195 (-0.151370) | 1.415616 / 1.468490 (-0.052874) | 0.568727 / 4.584777 (-4.016050) | 2.368330 / 3.745712 (-1.377382) | 2.844329 / 5.269862 (-2.425532) | 1.809038 / 4.565676 (-2.756639) | 0.063699 / 0.424275 (-0.360576) | 0.004972 / 0.007607 (-0.002635) | 0.340092 / 0.226044 (0.114048) | 3.369146 / 2.268929 (1.100217) | 1.863423 / 55.444624 (-53.581201) | 1.608334 / 6.876477 (-5.268142) | 1.624479 / 2.142072 (-0.517594) | 0.632439 / 4.805227 (-4.172788) | 0.116862 / 6.500664 (-6.383802) | 0.042558 / 0.075469 (-0.032911) |\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) | 0.967922 / 1.841788 (-0.873866) | 11.730612 / 8.074308 (3.656304) | 9.321333 / 10.191392 (-0.870059) | 0.142604 / 0.680424 (-0.537819) | 0.013934 / 0.534201 (-0.520267) | 0.285992 / 0.579283 (-0.293292) | 0.267639 / 0.434364 (-0.166724) | 0.324972 / 0.540337 (-0.215365) | 0.427077 / 1.386936 (-0.959859) |\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) | 0.005806 / 0.011353 (-0.005547) | 0.003771 / 0.011008 (-0.007237) | 0.049542 / 0.038508 (0.011034) | 0.030182 / 0.023109 (0.007073) | 0.303923 / 0.275898 (0.028025) | 0.325623 / 0.323480 (0.002143) | 0.004327 / 0.007986 (-0.003659) | 0.002818 / 0.004328 (-0.001510) | 0.048237 / 0.004250 (0.043987) | 0.047490 / 0.037052 (0.010437) | 0.316556 / 0.258489 (0.058067) | 0.348352 / 0.293841 (0.054512) | 0.029444 / 0.128546 (-0.099102) | 0.010544 / 0.075646 (-0.065102) | 0.057382 / 0.419271 (-0.361890) | 0.056210 / 0.043533 (0.012677) | 0.305495 / 0.255139 (0.050356) | 0.321570 / 0.283200 (0.038370) | 0.019546 / 0.141683 (-0.122137) | 1.141732 / 1.452155 (-0.310423) | 1.223626 / 1.492716 (-0.269091) |\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) | 0.093864 / 0.018006 (0.075858) | 0.309715 / 0.000490 (0.309226) | 0.000217 / 0.000200 (0.000017) | 0.000053 / 0.000054 (-0.000002) |\n\n### Benchmark: benchmark_indices_mapping.json\n\n| metric | select | shard | shuffle | sort | train_test_split |\n|--------|---|---|---|---|---|\n| new / old (diff) | 0.022047 / 0.037411 (-0.015364) | 0.074885 / 0.014526 (0.060359) | 0.088440 / 0.176557 (-0.088117) | 0.127033 / 0.737135 (-0.610103) | 0.089048 / 0.296338 (-0.207290) |\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) | 0.292624 / 0.215209 (0.077415) | 2.877592 / 2.077655 (0.799937) | 1.607036 / 1.504120 (0.102916) | 1.487819 / 1.541195 (-0.053376) | 1.517318 / 1.468490 (0.048828) | 0.553321 / 4.584777 (-4.031456) | 2.415577 / 3.745712 (-1.330135) | 2.691411 / 5.269862 (-2.578450) | 1.743395 / 4.565676 (-2.822282) | 0.062187 / 0.424275 (-0.362088) | 0.005073 / 0.007607 (-0.002534) | 0.342907 / 0.226044 (0.116863) | 3.402054 / 2.268929 (1.133126) | 1.979481 / 55.444624 (-53.465143) | 1.702885 / 6.876477 (-5.173592) | 1.868279 / 2.142072 (-0.273794) | 0.640095 / 4.805227 (-4.165132) | 0.117138 / 6.500664 (-6.383526) | 0.042197 / 0.075469 (-0.033272) |\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) | 1.007495 / 1.841788 (-0.834292) | 12.037309 / 8.074308 (3.963001) | 10.227670 / 10.191392 (0.036278) | 0.149533 / 0.680424 (-0.530891) | 0.015282 / 0.534201 (-0.518919) | 0.287357 / 0.579283 (-0.291926) | 0.285109 / 0.434364 (-0.149255) | 0.324027 / 0.540337 (-0.216311) | 0.442482 / 1.386936 (-0.944454) |\n\n</details>\n</details>\n\n![](https://cml.dev/watermark.png#19b40860acf3b3ba8db727fcf3b1b99ebb8d7e33 \"CML watermark\")\n"
] |
6,738
|
Dict feature is non-nullable while nested dict feature is
|
When i try to create a `Dataset` object with None values inside a dict column, like this:
```python
from datasets import Dataset, Features, Value
Dataset.from_dict(
{
"dict": [{"a": 0, "b": 0}, None],
}, features=Features(
{"dict": {"a": Value("int16"), "b": Value("int16")}}
)
)
```
i get `ValueError: Got None but expected a dictionary instead`.
At the same time, having None in _nested_ dict feature works, for example, this doesn't throw any errors:
```python
from datasets import Dataset, Features, Value, Sequence
dataset = Dataset.from_dict(
{
"list_dict": [[{"a": 0, "b": 0}], None],
"sequence_dict": [[{"a": 0, "b": 0}], None],
}, features=Features({
"list_dict": [{"a": Value("int16"), "b": Value("int16")}],
"sequence_dict": Sequence({"a": Value("int16"), "b": Value("int16")}),
})
)
```
Other types of features also seem to be nullable (but I haven't checked all of them).
Version of `datasets` is the latest atm (2.18.0)
Is this an expected behavior or a bug?
|
[
{
"id": 1935892857,
"node_id": "MDU6TGFiZWwxOTM1ODkyODU3",
"url": "https://api.github.com/repos/huggingface/datasets/labels/bug",
"name": "bug",
"color": "d73a4a",
"default": true,
"description": "Something isn't working"
}
] |
{
"url": null,
"html_url": null,
"diff_url": null,
"patch_url": null,
"merged_at": null
}
| false
|
[
"It looks like a bug, by default every feature should be nullable.",
"I've linked a PR with a fix :)",
"@mariosasko awesome thank you!"
] |
6,737
|
Invalid pattern: '**' can only be an entire path component
|
### Describe the bug
ValueError: Invalid pattern: '**' can only be an entire path component
when loading any dataset
### Steps to reproduce the bug
import datasets
ds = datasets.load_dataset("TokenBender/code_instructions_122k_alpaca_style")
### Expected behavior
loading the dataset successfully
### Environment info
- `datasets` version: 2.18.0
- Platform: Windows-10-10.0.22631-SP0
- Python version: 3.11.7
- `huggingface_hub` version: 0.20.3
- PyArrow version: 15.0.0
- Pandas version: 2.2.1
- `fsspec` version: 2023.12.2
|
[] |
{
"url": null,
"html_url": null,
"diff_url": null,
"patch_url": null,
"merged_at": null
}
| false
|
[
"I couldn't reproduce the issue on my side on MacOS, I guess the issue comes from the recent `fsspec` on Windows.\r\n\r\nCan you try downgrading to `fsspec==2023.9.2` for now ? It would also be great to investigate this and see if we need a fix in `datasets` or `fsspec`",
"I had the same issue! \r\nDowngrading to fsspec from 2023.10.0 to 2023.9.2 solved it for me.\r\n\r\n(env: python 3.11.7, datasets version: 2.15.0, Windows 10 22H2, Build 19045.4170)\r\n\r\nThanks a lot!",
"Ubuntu 20.04 had the same issue\r\npython 3.9 \r\n\r\nFile \"/home/delight-gpu/Workspace2/azuryl/FLAP/main.py\", line 112, in <module>\r\n main()\r\n File \"/home/delight-gpu/Workspace2/azuryl/FLAP/main.py\", line 85, in main\r\n prune_flap(args, model, tokenizer, device)\r\n File \"/home/delight-gpu/Workspace2/azuryl/FLAP/lib/prune.py\", line 294, in prune_flap\r\n dataloader, _ = get_loaders(\"wikitext2\", nsamples=args.nsamples,seed=args.seed,seqlen=model.seqlen,tokenizer=tokenizer)\r\n File \"/home/delight-gpu/Workspace2/azuryl/FLAP/lib/data.py\", line 159, in get_loaders\r\n return get_wikitext2(nsamples, seed, seqlen, tokenizer)\r\n File \"/home/delight-gpu/Workspace2/azuryl/FLAP/lib/data.py\", line 79, in get_wikitext2\r\n traindata = load_dataset('wikitext', 'wikitext-2-raw-v1', split='train')\r\n File \"/home/azuryl/anaconda3/envs/flap/lib/python3.9/site-packages/datasets/load.py\", line 1767, in load_dataset\r\n builder_instance = load_dataset_builder(\r\n File \"/home/azuryl/anaconda3/envs/flap/lib/python3.9/site-packages/datasets/load.py\", line 1498, in load_dataset_builder\r\n dataset_module = dataset_module_factory(\r\n File \"/home/azuryl/anaconda3/envs/flap/lib/python3.9/site-packages/datasets/load.py\", line 1215, in dataset_module_factory\r\n raise e1 from None\r\n File \"/home/azuryl/anaconda3/envs/flap/lib/python3.9/site-packages/datasets/load.py\", line 1192, in dataset_module_factory\r\n return HubDatasetModuleFactoryWithoutScript(\r\n File \"/home/azuryl/anaconda3/envs/flap/lib/python3.9/site-packages/datasets/load.py\", line 765, in get_module\r\n else get_data_patterns_in_dataset_repository(hfh_dataset_info, self.data_dir)\r\n File \"/home/azuryl/anaconda3/envs/flap/lib/python3.9/site-packages/datasets/data_files.py\", line 675, in get_data_patterns_in_dataset_repository\r\n return _get_data_files_patterns(resolver)\r\n File \"/home/azuryl/anaconda3/envs/flap/lib/python3.9/site-packages/datasets/data_files.py\", line 236, in _get_data_files_patterns\r\n data_files = pattern_resolver(pattern)\r\n File \"/home/azuryl/anaconda3/envs/flap/lib/python3.9/site-packages/datasets/data_files.py\", line 486, in _resolve_single_pattern_in_dataset_repository\r\n glob_iter = [PurePath(filepath) for filepath in fs.glob(PurePath(pattern).as_posix()) if fs.isfile(filepath)]\r\n File \"/home/azuryl/anaconda3/envs/flap/lib/python3.9/site-packages/fsspec/spec.py\", line 606, in glob\r\n pattern = glob_translate(path + (\"/\" if ends_with_sep else \"\"))\r\n File \"/home/azuryl/anaconda3/envs/flap/lib/python3.9/site-packages/fsspec/utils.py\", line 734, in glob_translate\r\n raise ValueError(\r\nValueError: Invalid pattern: '**' can only be an entire path component",
"on ubuntu you just need to have the latest `datasets` and `fsspec`\r\n\r\n```\r\npip install -U datasets fsspec\r\n```",
"The issue was caused by an incompatibility between the versions of `datasets`, `huggingface-hub` and `fsspec`.\r\n\r\nThe issue was fixed in:\r\n- huggingface-hub-0.21.2: https://github.com/huggingface/huggingface_hub/pull/2056\r\n- and datasets-2.18.0: https://github.com/huggingface/datasets/pull/6687\r\n - datasets-2.19.1 fixed the minimum requirement huggingface-hub >= 0.21.2: https://github.com/huggingface/datasets/pull/6713",
"@albertvillanova, thank you for this solution. I encountered the same issue and had to use:\r\n```\r\nconda install -c conda-forge huggingface_hub=0.21.2 datasets=2.19.1\r\n```\r\n\r\nCheers",
"Cheers！"
] |
6,736
|
Mosaic Streaming (MDS) Support
|
### Feature request
I'm a huge fan of the current HF Datasets `webdataset` integration (especially the built-in streaming support). However, I'd love to upload some robotics and multimodal datasets I've processed for use with [Mosaic Streaming](https://docs.mosaicml.com/projects/streaming/en/stable/), specifically their [MDS Format](https://docs.mosaicml.com/projects/streaming/en/stable/fundamentals/dataset_format.html#mds).
Because the shard files have similar semantics to WebDataset, I'm hoping that adding such support won't be too much trouble?
### Motivation
One of the downsides with WebDataset is a lack of out-of-the-box determinism (especially for large-scale training and reproducibility), easy job resumption, and the ability to quickly debug / visualize individual examples.
Mosaic Streaming provides a [great interface for this out of the box](https://docs.mosaicml.com/projects/streaming/en/stable/#key-features), so I'd love to see it supported in HF Datasets.
### Your contribution
Happy to help test things / provide example data. Can potentially submit a PR if maintainers could point me to the necessary WebDataset logic / steps for adding a new streaming format!
|
[
{
"id": 1935892871,
"node_id": "MDU6TGFiZWwxOTM1ODkyODcx",
"url": "https://api.github.com/repos/huggingface/datasets/labels/enhancement",
"name": "enhancement",
"color": "a2eeef",
"default": true,
"description": "New feature or request"
}
] |
{
"url": null,
"html_url": null,
"diff_url": null,
"patch_url": null,
"merged_at": null
}
| false
|
[
"Hi ! that would be great :) Though note that `datasets` doesn't implement format-specific resuming when streaming, so in general I think it's better if users can use the mosaic-streaming library to read their MDS datasets. I wonder if they support `hf://` paths though...\r\n\r\nAnyway for those interested, the code for WebDataset is a single file here: https://github.com/huggingface/datasets/blob/main/src/datasets/packaged_modules/webdataset/webdataset.py.\r\n\r\nIt implements `_split_generators` that downloads files and returns the lists of splits (train/validation/test) and `_split_generators` to generate examples (dicts) from the downloaded files. Streaming is automatically supported by making download steps lazy and by extending `open()` to work with remote URLs."
] |
6,735
|
Add `mode` parameter to `Image` feature
|
Fix https://github.com/huggingface/datasets/issues/6675
|
[] |
{
"url": "https://api.github.com/repos/huggingface/datasets/pulls/6735",
"html_url": "https://github.com/huggingface/datasets/pull/6735",
"diff_url": "https://github.com/huggingface/datasets/pull/6735.diff",
"patch_url": "https://github.com/huggingface/datasets/pull/6735.patch",
"merged_at": "2024-03-18T15:41:33"
}
| true
|
[
"The docs for this PR live [here](https://moon-ci-docs.huggingface.co/docs/datasets/pr_6735). All of your documentation changes will be reflected on that endpoint. The docs are available until 30 days after the last update.",
"<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) | 0.005009 / 0.011353 (-0.006344) | 0.003547 / 0.011008 (-0.007461) | 0.063014 / 0.038508 (0.024506) | 0.027699 / 0.023109 (0.004589) | 0.247140 / 0.275898 (-0.028758) | 0.273610 / 0.323480 (-0.049870) | 0.003115 / 0.007986 (-0.004871) | 0.002712 / 0.004328 (-0.001616) | 0.049134 / 0.004250 (0.044883) | 0.041582 / 0.037052 (0.004530) | 0.269992 / 0.258489 (0.011503) | 0.294516 / 0.293841 (0.000675) | 0.027818 / 0.128546 (-0.100728) | 0.010568 / 0.075646 (-0.065078) | 0.207710 / 0.419271 (-0.211561) | 0.035767 / 0.043533 (-0.007766) | 0.260058 / 0.255139 (0.004919) | 0.277615 / 0.283200 (-0.005585) | 0.020192 / 0.141683 (-0.121491) | 1.116863 / 1.452155 (-0.335292) | 1.156868 / 1.492716 (-0.335848) |\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) | 0.095087 / 0.018006 (0.077081) | 0.303249 / 0.000490 (0.302759) | 0.000215 / 0.000200 (0.000015) | 0.000053 / 0.000054 (-0.000001) |\n\n### Benchmark: benchmark_indices_mapping.json\n\n| metric | select | shard | shuffle | sort | train_test_split |\n|--------|---|---|---|---|---|\n| new / old (diff) | 0.018866 / 0.037411 (-0.018545) | 0.063853 / 0.014526 (0.049328) | 0.073863 / 0.176557 (-0.102693) | 0.121399 / 0.737135 (-0.615737) | 0.076014 / 0.296338 (-0.220325) |\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) | 0.289843 / 0.215209 (0.074634) | 2.844085 / 2.077655 (0.766431) | 1.528022 / 1.504120 (0.023902) | 1.397352 / 1.541195 (-0.143843) | 1.394676 / 1.468490 (-0.073814) | 0.555899 / 4.584777 (-4.028878) | 2.354010 / 3.745712 (-1.391702) | 2.737715 / 5.269862 (-2.532146) | 1.731260 / 4.565676 (-2.834416) | 0.062315 / 0.424275 (-0.361960) | 0.004920 / 0.007607 (-0.002687) | 0.342921 / 0.226044 (0.116877) | 3.416529 / 2.268929 (1.147600) | 1.862941 / 55.444624 (-53.581684) | 1.599661 / 6.876477 (-5.276816) | 1.617200 / 2.142072 (-0.524873) | 0.635129 / 4.805227 (-4.170099) | 0.121651 / 6.500664 (-6.379013) | 0.041867 / 0.075469 (-0.033602) |\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) | 0.990825 / 1.841788 (-0.850962) | 11.435576 / 8.074308 (3.361268) | 9.490194 / 10.191392 (-0.701198) | 0.133295 / 0.680424 (-0.547129) | 0.014061 / 0.534201 (-0.520140) | 0.288648 / 0.579283 (-0.290635) | 0.268874 / 0.434364 (-0.165490) | 0.323288 / 0.540337 (-0.217049) | 0.426090 / 1.386936 (-0.960846) |\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) | 0.006712 / 0.011353 (-0.004641) | 0.003723 / 0.011008 (-0.007285) | 0.049814 / 0.038508 (0.011306) | 0.039323 / 0.023109 (0.016213) | 0.279244 / 0.275898 (0.003346) | 0.297139 / 0.323480 (-0.026341) | 0.004197 / 0.007986 (-0.003788) | 0.002753 / 0.004328 (-0.001576) | 0.048820 / 0.004250 (0.044569) | 0.049593 / 0.037052 (0.012541) | 0.287247 / 0.258489 (0.028758) | 0.338078 / 0.293841 (0.044237) | 0.029303 / 0.128546 (-0.099243) | 0.010292 / 0.075646 (-0.065354) | 0.057852 / 0.419271 (-0.361419) | 0.053390 / 0.043533 (0.009857) | 0.275155 / 0.255139 (0.020016) | 0.292891 / 0.283200 (0.009692) | 0.020007 / 0.141683 (-0.121676) | 1.161731 / 1.452155 (-0.290424) | 1.232162 / 1.492716 (-0.260555) |\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) | 0.092848 / 0.018006 (0.074842) | 0.301180 / 0.000490 (0.300690) | 0.000236 / 0.000200 (0.000036) | 0.000050 / 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) | 0.022477 / 0.037411 (-0.014934) | 0.077012 / 0.014526 (0.062486) | 0.087335 / 0.176557 (-0.089222) | 0.126761 / 0.737135 (-0.610374) | 0.089249 / 0.296338 (-0.207090) |\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) | 0.290722 / 0.215209 (0.075513) | 2.884485 / 2.077655 (0.806830) | 1.565775 / 1.504120 (0.061656) | 1.442369 / 1.541195 (-0.098825) | 1.453995 / 1.468490 (-0.014495) | 0.563193 / 4.584777 (-4.021584) | 2.413610 / 3.745712 (-1.332102) | 2.684567 / 5.269862 (-2.585295) | 1.753322 / 4.565676 (-2.812354) | 0.061879 / 0.424275 (-0.362396) | 0.005080 / 0.007607 (-0.002527) | 0.347274 / 0.226044 (0.121229) | 3.435836 / 2.268929 (1.166907) | 1.937893 / 55.444624 (-53.506731) | 1.657824 / 6.876477 (-5.218653) | 1.777767 / 2.142072 (-0.364305) | 0.656757 / 4.805227 (-4.148471) | 0.117144 / 6.500664 (-6.383520) | 0.040691 / 0.075469 (-0.034778) |\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) | 1.012435 / 1.841788 (-0.829353) | 12.038001 / 8.074308 (3.963693) | 10.363947 / 10.191392 (0.172555) | 0.140711 / 0.680424 (-0.539713) | 0.014937 / 0.534201 (-0.519264) | 0.291070 / 0.579283 (-0.288213) | 0.277180 / 0.434364 (-0.157184) | 0.327433 / 0.540337 (-0.212904) | 0.439767 / 1.386936 (-0.947169) |\n\n</details>\n</details>\n\n![](https://cml.dev/watermark.png#0b55ec53e980855d71ae22f8b3d12b2a0d476a51 \"CML watermark\")\n"
] |
6,734
|
Tokenization slows towards end of dataset
|
### Describe the bug
Mapped tokenization slows down substantially towards end of dataset.
train set started off very slow, caught up to 20k then tapered off til the end.
what's particularly strange is that the tokenization crashed a few times before due to errors with invalid tokens somewhere or corrupted downloads, and the speed ups/downs consistently happened the same times
```bash
Running tokenizer on dataset (num_proc=48): 0%| | 847000/881416735 [12:18<252:45:45, 967.72 examples/s]
Running tokenizer on dataset (num_proc=48): 0%| | 848000/881416735 [12:19<224:16:10, 1090.66 examples/s]
Running tokenizer on dataset (num_proc=48): 10%|▉ | 84964000/881416735 [3:48:00<11:21:34, 19476.01 examples/s]
Running tokenizer on dataset (num_proc=48): 10%|▉ | 84967000/881416735 [3:48:00<12:04:01, 18333.79 examples/s]
Running tokenizer on dataset (num_proc=48): 61%|██████ | 538631977/881416735 [13:46:40<27:50:04, 3420.84 examples/s]
Running tokenizer on dataset (num_proc=48): 61%|██████ | 538632977/881416735 [13:46:40<23:48:20, 3999.77 examples/s]
Running tokenizer on dataset (num_proc=48): 100%|█████████▉| 881365886/881416735 [38:30:19<04:34, 185.10 examples/s]
Running tokenizer on dataset (num_proc=48): 100%|█████████▉| 881366886/881416735 [38:30:25<04:36, 180.57 examples/s]
```
and validation set as well
```bash
Running tokenizer on dataset (num_proc=48): 90%|████████▉ | 41544000/46390354 [28:44<02:37, 30798.76 examples/s]
Running tokenizer on dataset (num_proc=48): 90%|████████▉ | 41550000/46390354 [28:44<02:08, 37698.08 examples/s]
Running tokenizer on dataset (num_proc=48): 96%|█████████▋| 44747422/46390354 [2:15:48<12:22:44, 36.87 examples/s]
Running tokenizer on dataset (num_proc=48): 96%|█████████▋| 44747422/46390354 [2:16:00<12:22:44, 36.87 examples/s]
```
### Steps to reproduce the bug
using the following kwargs
```python
with accelerator.main_process_first():
lm_datasets = tokenized_datasets.map(
group_texts,
batched=True,
num_proc=48
load_from_cache_file=True,
desc=f"Grouping texts in chunks of {block_size}",
)
```
running through slurm script
```bash
#SBATCH --partition=gpu-nvidia-a100
#SBATCH --nodes=1
#SBATCH --ntasks=1
#SBATCH --gpus-per-task=8
#SBATCH --cpus-per-task=96
```
using this dataset https://huggingface.co/datasets/togethercomputer/RedPajama-Data-1T
### Expected behavior
Constant speed throughout
### Environment info
- `datasets` version: 2.15.0
- Platform: Linux-5.15.0-1049-aws-x86_64-with-glibc2.10
- Python version: 3.8.18
- `huggingface_hub` version: 0.19.4
- PyArrow version: 14.0.1
- Pandas version: 2.0.3
- `fsspec` version: 2023.10.0
|
[] |
{
"url": null,
"html_url": null,
"diff_url": null,
"patch_url": null,
"merged_at": null
}
| false
|
[
"Hi ! First note that if the dataset is not heterogeneous / shuffled, there might be places in the data with shorter texts that are faster to tokenize.\r\n\r\nMoreover, the way `num_proc` works is by slicing the dataset and passing each slice to a process to run the `map()` function. So at the very end of `map()`, some processes might have finished transforming their slice of data while others are still running, causing the throughput to become lower.",
"I did see some comments about how num_proc=None could help and outputting numpy arrays can also help in the docs, but this seems quite odd now dropping down to 1it/s\r\n\r\n```bash\r\nRunning tokenizer on dataset (num_proc=48): 99%|█████████▉| 46048888/46390354 [12:33:30<4:20:32, 21.84 examples/s]\r\nRunning tokenizer on dataset (num_proc=48): 99%|█████████▉| 46049888/46390354 [12:36:11<8:37:59, 10.95 examples/s]\r\nRunning tokenizer on dataset (num_proc=48): 99%|█████████▉| 46050888/46390354 [12:46:35<24:56:56, 3.78 examples/s]\r\nRunning tokenizer on dataset (num_proc=48): 99%|█████████▉| 46051888/46390354 [12:56:43<35:08:10, 2.68 examples/s]\r\nRunning tokenizer on dataset (num_proc=48): 99%|█████████▉| 46052888/46390354 [13:06:58<42:05:41, 2.23 examples/s]\r\nRunning tokenizer on dataset (num_proc=48): 99%|█████████▉| 46053888/46390354 [13:16:01<44:40:18, 2.09 examples/s]\r\nRunning tokenizer on dataset (num_proc=48): 99%|█████████▉| 46054888/46390354 [13:25:11<46:35:28, 2.00 examples/s]\r\nRunning tokenizer on dataset (num_proc=48): 99%|█████████▉| 46055888/46390354 [13:34:23<47:55:34, 1.94 examples/s]\r\n```\r\n\r\n",
"@ethansmith2000 Hi, did you solve this problem? I'm strugging with the same problem now."
] |
6,733
|
EmptyDatasetError when loading dataset downloaded with HuggingFace cli
|
### Describe the bug
I am using a cluster that does not have access to the internet when given a job. I tried downloading the dataset using the huggingface-cli command and then loading it with load_dataset but I get an error:
```raise EmptyDatasetError(f"The directory at {base_path} doesn't contain any data files") from None```
The dataset I'm using is "lmsys/chatbot_arena_conversations". The folder structure is
- README.md
- data
- train-00000-of-00001-cced8514c7ed782a.parquet
### Steps to reproduce the bug
1. Download dataset using HuggingFace CLI: ```huggingface-cli download lmsys/chatbot_arena_conversations --local-dir ./lmsys/chatbot_arena_conversations```
2. In Python
```
from datasets import load_dataset
load_dataset("lmsys/chatbot_arena_conversations")
```
### Expected behavior
Should return a Dataset Dict in the form of
```
DatasetDict({
train: Dataset({
features: [...],
num_rows: 33,000
})
})
```
### Environment info
Python 3.11.5
Datasets 2.18.0
Transformers 4.38.2
Pytorch 2.2.0
Pyarrow 15.0.1
Rocky Linux release 8.9 (Green Obsidian)
|
[] |
{
"url": null,
"html_url": null,
"diff_url": null,
"patch_url": null,
"merged_at": null
}
| false
|
[
"Hi! `datasets` is not compatible with `huggingface_hub`'s cache structure, hence the error.\r\n\r\nYou can track https://github.com/huggingface/datasets/issues/5080 to get notified when this is implemented."
] |
6,731
|
Unexpected behavior when using load_dataset with streaming=True in a for loop
|
### Describe the bug
### My Code
```
from datasets import load_dataset
res=[]
for i in [0,1]:
di=load_dataset(
"json",
data_files='path_to.json',
split='train',
streaming=True,
).map(lambda x: {"source": i})
res.append(di)
for e in res[0]:
print(e)
```
### Unexpected Behavior
Data in `res[0]` has `source=1`. However the expected value is 0.
### FYI
I further switch `streaming` to `False`. And the output value is as expected (0). So there may exist bugs in setting `streaming=True` in a for loop.
### Environment
Python 3.8.0
datasets==2.18.0
transformers==4.28.1
### Steps to reproduce the bug
1. Create a Json file with any content.
2. Run the provided code.
3. Switch `streaming` to `False` and run again to see the expected behavior.
### Expected behavior
The expected behavior is the data are mapped with its corresponding value in the for loop.
### Environment info
Python 3.8.0
datasets==2.18.0
transformers==4.28.1
Ubuntu 20.04
|
[] |
{
"url": null,
"html_url": null,
"diff_url": null,
"patch_url": null,
"merged_at": null
}
| false
|
[
"This is normal behavior in python when using `lambda`: the `i` defined in your `lambda` refers to the global variable `i` in your loop, and `i` equals to `1` when you run your `for e in res[0]` line.\r\n\r\nYou should pass `fn_kwargs` that will be passed to your `lambda` instead of using the global variable:\r\n\r\n```python\r\nfrom datasets import load_dataset\r\n\r\nres=[]\r\nfor i in [0,1]:\r\n di = load_dataset(\r\n \"json\", \r\n data_files='path_to.json', \r\n split='train',\r\n streaming=True, \r\n ).map(lambda x, source: {\"source\": source}, fn_kwargs={\"source\": i})\r\n\r\n res.append(di)\r\n\r\nfor e in res[0]:\r\n print(e)\r\n```\r\n\r\nThis doesn't happen in non-streaming since in that case `map` is executed while the variable `i` has the right value. In streaming mode, `map` is executed on-the-fly when you iterate on the dataset.",
"Thank you very much for your answer. I think this issue can be closed now."
] |
6,730
|
Deprecate Pandas builder
|
The Pandas packaged builder is undocumented and relies on `pickle` to read the data, making it **unsafe**. Moreover, I haven't seen a single instance of this builder being used (not even using the GH/Hub search), so we should deprecate it.
|
[] |
{
"url": "https://api.github.com/repos/huggingface/datasets/pulls/6730",
"html_url": "https://github.com/huggingface/datasets/pull/6730",
"diff_url": "https://github.com/huggingface/datasets/pull/6730.diff",
"patch_url": "https://github.com/huggingface/datasets/pull/6730.patch",
"merged_at": "2024-03-12T17:36:24"
}
| true
|
[
"The docs for this PR live [here](https://moon-ci-docs.huggingface.co/docs/datasets/pr_6730). All of your documentation changes will be reflected on that endpoint. The docs are available until 30 days after the last update.",
"<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) | 0.005301 / 0.011353 (-0.006052) | 0.003701 / 0.011008 (-0.007307) | 0.065830 / 0.038508 (0.027322) | 0.029791 / 0.023109 (0.006682) | 0.251676 / 0.275898 (-0.024222) | 0.283824 / 0.323480 (-0.039655) | 0.003083 / 0.007986 (-0.004903) | 0.004144 / 0.004328 (-0.000185) | 0.053670 / 0.004250 (0.049419) | 0.042020 / 0.037052 (0.004968) | 0.266389 / 0.258489 (0.007899) | 0.296740 / 0.293841 (0.002900) | 0.028320 / 0.128546 (-0.100226) | 0.010604 / 0.075646 (-0.065042) | 0.219881 / 0.419271 (-0.199390) | 0.036216 / 0.043533 (-0.007317) | 0.255718 / 0.255139 (0.000579) | 0.275808 / 0.283200 (-0.007392) | 0.018407 / 0.141683 (-0.123276) | 1.140007 / 1.452155 (-0.312148) | 1.174005 / 1.492716 (-0.318711) |\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) | 0.091230 / 0.018006 (0.073224) | 0.300704 / 0.000490 (0.300215) | 0.000207 / 0.000200 (0.000007) | 0.000043 / 0.000054 (-0.000011) |\n\n### Benchmark: benchmark_indices_mapping.json\n\n| metric | select | shard | shuffle | sort | train_test_split |\n|--------|---|---|---|---|---|\n| new / old (diff) | 0.018950 / 0.037411 (-0.018461) | 0.062177 / 0.014526 (0.047651) | 0.073968 / 0.176557 (-0.102589) | 0.122161 / 0.737135 (-0.614974) | 0.075001 / 0.296338 (-0.221338) |\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) | 0.285675 / 0.215209 (0.070466) | 2.794176 / 2.077655 (0.716522) | 1.478666 / 1.504120 (-0.025454) | 1.361843 / 1.541195 (-0.179351) | 1.383847 / 1.468490 (-0.084643) | 0.568610 / 4.584777 (-4.016167) | 2.402351 / 3.745712 (-1.343361) | 2.860772 / 5.269862 (-2.409089) | 1.768588 / 4.565676 (-2.797089) | 0.063257 / 0.424275 (-0.361018) | 0.004998 / 0.007607 (-0.002609) | 0.340897 / 0.226044 (0.114853) | 3.340238 / 2.268929 (1.071310) | 1.836434 / 55.444624 (-53.608190) | 1.556844 / 6.876477 (-5.319633) | 1.610685 / 2.142072 (-0.531388) | 0.644941 / 4.805227 (-4.160286) | 0.117593 / 6.500664 (-6.383072) | 0.042803 / 0.075469 (-0.032666) |\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) | 0.979181 / 1.841788 (-0.862607) | 11.901365 / 8.074308 (3.827057) | 9.587943 / 10.191392 (-0.603449) | 0.139648 / 0.680424 (-0.540776) | 0.013904 / 0.534201 (-0.520297) | 0.291249 / 0.579283 (-0.288034) | 0.260737 / 0.434364 (-0.173627) | 0.326000 / 0.540337 (-0.214338) | 0.433459 / 1.386936 (-0.953477) |\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) | 0.005503 / 0.011353 (-0.005850) | 0.003738 / 0.011008 (-0.007270) | 0.049137 / 0.038508 (0.010629) | 0.031484 / 0.023109 (0.008374) | 0.265783 / 0.275898 (-0.010115) | 0.295125 / 0.323480 (-0.028354) | 0.004074 / 0.007986 (-0.003911) | 0.002707 / 0.004328 (-0.001622) | 0.048340 / 0.004250 (0.044089) | 0.045453 / 0.037052 (0.008401) | 0.276500 / 0.258489 (0.018011) | 0.312002 / 0.293841 (0.018162) | 0.029139 / 0.128546 (-0.099408) | 0.010445 / 0.075646 (-0.065201) | 0.057486 / 0.419271 (-0.361785) | 0.052386 / 0.043533 (0.008853) | 0.267099 / 0.255139 (0.011960) | 0.283193 / 0.283200 (-0.000007) | 0.018368 / 0.141683 (-0.123315) | 1.136207 / 1.452155 (-0.315948) | 1.178418 / 1.492716 (-0.314298) |\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) | 0.089270 / 0.018006 (0.071264) | 0.301087 / 0.000490 (0.300598) | 0.000208 / 0.000200 (0.000008) | 0.000050 / 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) | 0.021991 / 0.037411 (-0.015421) | 0.075357 / 0.014526 (0.060831) | 0.087781 / 0.176557 (-0.088775) | 0.126923 / 0.737135 (-0.610212) | 0.088491 / 0.296338 (-0.207847) |\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) | 0.293653 / 0.215209 (0.078444) | 2.872156 / 2.077655 (0.794501) | 1.559229 / 1.504120 (0.055109) | 1.441201 / 1.541195 (-0.099993) | 1.472642 / 1.468490 (0.004152) | 0.588463 / 4.584777 (-3.996314) | 2.447685 / 3.745712 (-1.298028) | 2.755752 / 5.269862 (-2.514110) | 1.796591 / 4.565676 (-2.769086) | 0.068024 / 0.424275 (-0.356252) | 0.005148 / 0.007607 (-0.002459) | 0.343572 / 0.226044 (0.117528) | 3.347856 / 2.268929 (1.078927) | 1.945977 / 55.444624 (-53.498647) | 1.648953 / 6.876477 (-5.227524) | 1.804468 / 2.142072 (-0.337604) | 0.651034 / 4.805227 (-4.154193) | 0.118130 / 6.500664 (-6.382534) | 0.041019 / 0.075469 (-0.034450) |\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) | 1.020461 / 1.841788 (-0.821327) | 12.514237 / 8.074308 (4.439929) | 10.696276 / 10.191392 (0.504884) | 0.154549 / 0.680424 (-0.525874) | 0.015964 / 0.534201 (-0.518237) | 0.290392 / 0.579283 (-0.288891) | 0.276074 / 0.434364 (-0.158290) | 0.326253 / 0.540337 (-0.214085) | 0.440383 / 1.386936 (-0.946553) |\n\n</details>\n</details>\n\n![](https://cml.dev/watermark.png#29ffc270da34de70cf8e28b2ebeadba1c06d8730 \"CML watermark\")\n"
] |
6,729
|
Support zipfiles that span multiple disks?
|
See https://huggingface.co/datasets/PhilEO-community/PhilEO-downstream
The dataset viewer gives the following error:
```
Error code: ConfigNamesError
Exception: BadZipFile
Message: zipfiles that span multiple disks are not supported
Traceback: Traceback (most recent call last):
File "/src/services/worker/src/worker/job_runners/dataset/config_names.py", line 67, in compute_config_names_response
get_dataset_config_names(
File "/src/services/worker/.venv/lib/python3.9/site-packages/datasets/inspect.py", line 347, in get_dataset_config_names
dataset_module = dataset_module_factory(
File "/src/services/worker/.venv/lib/python3.9/site-packages/datasets/load.py", line 1871, in dataset_module_factory
raise e1 from None
File "/src/services/worker/.venv/lib/python3.9/site-packages/datasets/load.py", line 1846, in dataset_module_factory
return HubDatasetModuleFactoryWithoutScript(
File "/src/services/worker/.venv/lib/python3.9/site-packages/datasets/load.py", line 1240, in get_module
module_name, default_builder_kwargs = infer_module_for_data_files(
File "/src/services/worker/.venv/lib/python3.9/site-packages/datasets/load.py", line 584, in infer_module_for_data_files
split_modules = {
File "/src/services/worker/.venv/lib/python3.9/site-packages/datasets/load.py", line 585, in <dictcomp>
split: infer_module_for_data_files_list(data_files_list, download_config=download_config)
File "/src/services/worker/.venv/lib/python3.9/site-packages/datasets/load.py", line 526, in infer_module_for_data_files_list
return infer_module_for_data_files_list_in_archives(data_files_list, download_config=download_config)
File "/src/services/worker/.venv/lib/python3.9/site-packages/datasets/load.py", line 554, in infer_module_for_data_files_list_in_archives
for f in xglob(extracted, recursive=True, download_config=download_config)[
File "/src/services/worker/.venv/lib/python3.9/site-packages/datasets/download/streaming_download_manager.py", line 576, in xglob
fs, *_ = fsspec.get_fs_token_paths(urlpath, storage_options=storage_options)
File "/src/services/worker/.venv/lib/python3.9/site-packages/fsspec/core.py", line 622, in get_fs_token_paths
fs = filesystem(protocol, **inkwargs)
File "/src/services/worker/.venv/lib/python3.9/site-packages/fsspec/registry.py", line 290, in filesystem
return cls(**storage_options)
File "/src/services/worker/.venv/lib/python3.9/site-packages/fsspec/spec.py", line 79, in __call__
obj = super().__call__(*args, **kwargs)
File "/src/services/worker/.venv/lib/python3.9/site-packages/fsspec/implementations/zip.py", line 57, in __init__
self.zip = zipfile.ZipFile(
File "/usr/local/lib/python3.9/zipfile.py", line 1266, in __init__
self._RealGetContents()
File "/usr/local/lib/python3.9/zipfile.py", line 1329, in _RealGetContents
endrec = _EndRecData(fp)
File "/usr/local/lib/python3.9/zipfile.py", line 286, in _EndRecData
return _EndRecData64(fpin, -sizeEndCentDir, endrec)
File "/usr/local/lib/python3.9/zipfile.py", line 232, in _EndRecData64
raise BadZipFile("zipfiles that span multiple disks are not supported")
zipfile.BadZipFile: zipfiles that span multiple disks are not supported
```
The files (https://huggingface.co/datasets/PhilEO-community/PhilEO-downstream/tree/main/data) are:
<img width="629" alt="Capture d’écran 2024-03-11 à 22 07 30" src="https://github.com/huggingface/datasets/assets/1676121/0bb15a51-d54f-4d73-8572-e427ea644b36">
|
[
{
"id": 1935892871,
"node_id": "MDU6TGFiZWwxOTM1ODkyODcx",
"url": "https://api.github.com/repos/huggingface/datasets/labels/enhancement",
"name": "enhancement",
"color": "a2eeef",
"default": true,
"description": "New feature or request"
},
{
"id": 1935892912,
"node_id": "MDU6TGFiZWwxOTM1ODkyOTEy",
"url": "https://api.github.com/repos/huggingface/datasets/labels/question",
"name": "question",
"color": "d876e3",
"default": true,
"description": "Further information is requested"
}
] |
{
"url": null,
"html_url": null,
"diff_url": null,
"patch_url": null,
"merged_at": null
}
| false
|
[
"@severo were you able to solve it?",
"No. cc @albertvillanova @lhoestq @polinaeterna for an evaluation of what it would take to support this feature.",
"The underlying issue issue is that the dataset repository has used split ZIP archive files: https://huggingface.co/datasets/PhilEO-community/PhilEO-downstream/tree/main/data\r\n```\r\ndownstream_dataset_patches_npzip.z01\r\ndownstream_dataset_patches_npzip.z02\r\n...\r\ndownstream_dataset_patches_npzip.zip\r\n```\r\nand these are not supported by the Python standard library package `zipfile`.",
"It's a pretty bad way to share a dataset since one needs to download the full dataset to use it.\r\n\r\nWe likely won't support this format.",
"I agree it is a format we maybe should not support: streaming is not possible.",
"I opened a PR in the reported repo to disable the viewer: https://huggingface.co/datasets/PhilEO-community/PhilEO-downstream/discussions/1"
] |
6,728
|
Issue Downloading Certain Datasets After Setting Custom `HF_ENDPOINT`
|
### Describe the bug
This bug is triggered under the following conditions:
- datasets repo ids without organization names trigger errors, such as `bookcorpus`, `gsm8k`, `wikipedia`, rather than in the form of `A/B`.
- If `HF_ENDPOINT` is set and the hostname is not in the form of `(hub-ci.)?huggingface.co`.
- This issue occurs with `datasets>2.15.0` or `huggingface-hub>0.19.4`. For example, using the latest versions: `datasets==2.18.0` and `huggingface-hub==0.21.4`,
### Steps to reproduce the bug
the issue can be reproduced with the following code:
1. install specific datasets and huggingface_hub.
```bash
pip install datasets==2.18.0
pip install huggingface_hub==0.21.4
```
2. execute python code.
```Python
import os
os.environ['HF_ENDPOINT'] = 'https://hf-mirror.com'
from datasets import load_dataset
bookcorpus = load_dataset('bookcorpus', split='train')
```
console output:
```
Traceback (most recent call last):
File "<stdin>", line 1, in <module>
File "/home/padeoe/.local/lib/python3.10/site-packages/datasets/load.py", line 2556, in load_dataset
builder_instance = load_dataset_builder(
File "/home/padeoe/.local/lib/python3.10/site-packages/datasets/load.py", line 2228, in load_dataset_builder
dataset_module = dataset_module_factory(
File "/home/padeoe/.local/lib/python3.10/site-packages/datasets/load.py", line 1879, in dataset_module_factory
raise e1 from None
File "/home/padeoe/.local/lib/python3.10/site-packages/datasets/load.py", line 1830, in dataset_module_factory
with fs.open(f"datasets/{path}/{filename}", "r", encoding="utf-8") as f:
File "/home/padeoe/.local/lib/python3.10/site-packages/fsspec/spec.py", line 1295, in open
self.open(
File "/home/padeoe/.local/lib/python3.10/site-packages/fsspec/spec.py", line 1307, in open
f = self._open(
File "/home/padeoe/.local/lib/python3.10/site-packages/huggingface_hub/hf_file_system.py", line 228, in _open
return HfFileSystemFile(self, path, mode=mode, revision=revision, block_size=block_size, **kwargs)
File "/home/padeoe/.local/lib/python3.10/site-packages/huggingface_hub/hf_file_system.py", line 615, in __init__
self.resolved_path = fs.resolve_path(path, revision=revision)
File "/home/padeoe/.local/lib/python3.10/site-packages/huggingface_hub/hf_file_system.py", line 180, in resolve_path
repo_and_revision_exist, err = self._repo_and_revision_exist(repo_type, repo_id, revision)
File "/home/padeoe/.local/lib/python3.10/site-packages/huggingface_hub/hf_file_system.py", line 117, in _repo_and_revision_exist
self._api.repo_info(repo_id, revision=revision, repo_type=repo_type)
File "/home/padeoe/.local/lib/python3.10/site-packages/huggingface_hub/utils/_validators.py", line 118, in _inner_fn
return fn(*args, **kwargs)
File "/home/padeoe/.local/lib/python3.10/site-packages/huggingface_hub/hf_api.py", line 2413, in repo_info
return method(
File "/home/padeoe/.local/lib/python3.10/site-packages/huggingface_hub/utils/_validators.py", line 118, in _inner_fn
return fn(*args, **kwargs)
File "/home/padeoe/.local/lib/python3.10/site-packages/huggingface_hub/hf_api.py", line 2286, in dataset_info
hf_raise_for_status(r)
File "/home/padeoe/.local/lib/python3.10/site-packages/huggingface_hub/utils/_errors.py", line 362, in hf_raise_for_status
raise HfHubHTTPError(str(e), response=response) from e
huggingface_hub.utils._errors.HfHubHTTPError: 401 Client Error: Unauthorized for url: https://hf-mirror.com/api/datasets/bookcorpus/bookcorpus.py (Request ID: Root=1-65ee8659-5ab10eec5960c63e71f2bb58;b00bdbea-fd6e-4a74-8fe0-bc4682ae090e)
```
### Expected behavior
The dataset was downloaded correctly without any errors.
### Environment info
datasets==2.18.0
huggingface-hub==0.21.4
|
[] |
{
"url": null,
"html_url": null,
"diff_url": null,
"patch_url": null,
"merged_at": null
}
| false
|
[
"Through debugging, I found a potential solution is to modify the code in the error handling module of `huggingface_hub`: https://github.com/huggingface/huggingface_hub/commit/56d6c798c44e83d2a3167e74c022737d8fcbe822 ",
"@Wauplin ",
"Thanks for investigating and reporting the bug @padeoe! I've opened a PR in `huggingface_hub` with your suggested fix! :) https://github.com/huggingface/huggingface_hub/pull/2119"
] |
6,727
|
Using a registry instead of calling globals for fetching feature types
|
Hello,
When working with bio-data, each feature often has metadata associated with it (e.g. species, lineage, snp position, etc). To store this, I like to use the feature classes with the added `metadata` attribute. However, when saving or loading with custom features, you get an error since that class doesn't exist in the global namespace in `datasets.features.features`. Take for example,
```python
from dataclasses import dataclass, field
from datasets import Dataset
from datasets.features.features import Value, Features
@dataclass
class FeatureA(Value):
metadata: dict = field(default=dict)
_type: str = field(default="FeatureA", init=False, repr=False)
@dataclass
class FeatureB(Value):
metadata: dict = field(default_factory=dict)
_type: str = field(default="FeatureB", init=False, repr=False)
test_data = {
"a": [1, 2, 3],
"b": [4, 5, 6],
}
test_data = Dataset.from_dict(
test_data,
features=Features({
"a": FeatureA("int32", metadata={"species": "lactobacillus acetotolerans"}),
"b": FeatureB("int32", metadata={"species": "lactobacillus iners"}),
})
)
# returns an error since FeatureA and FeatureB are not in the global namespace
test_data.save_to_disk('./test_data')
```
Saving the dataset (0/1 shards): 0%| | 0/3 [00:00<?, ? examples/s]
---------------------------------------------------------------------------
KeyError Traceback (most recent call last)
Cell In[2], line 28
19 test_data = Dataset.from_dict(
20 test_data,
21 features=Features({
(...)
24 })
25 )
27 # returns an error since FeatureA and FeatureB are not in the global namespace
---> 28 test_data.save_to_disk('./test_data')
...
File ~\Documents\datasets\src\datasets\features\features.py:1361, in generate_from_dict(obj)
1359 return {key: generate_from_dict(value) for key, value in obj.items()}
1360 obj = dict(obj)
-> 1361 class_type = globals()[obj.pop("_type")]
1363 if class_type == Sequence:
1364 return Sequence(feature=generate_from_dict(obj["feature"]), length=obj.get("length", -1))
KeyError: 'FeatureA'
We can avoid this by having a registry (like formatters) and doing
```python
from datasets.features.features import register_feature
register_feature(FeatureA, "FeatureA")
register_feature(FeatureB, "FeatureB")
test_data.save_to_disk('./test_data')
```
Saving the dataset (1/1 shards): 100%|------| 3/3 [00:00<00:00, 211.13 examples/s]
and loading from disk returns with all metadata information
```python
from datasets import load_from_disk
test_data = load_from_disk('./test_data')
test_data.features
```
{'a': FeatureA(dtype='int32', id=None, metadata={'species': 'lactobacillus acetotolerans'}),
'b': FeatureB(dtype='int32', id=None, metadata={'species': 'lactobacillus iners'})}
|
[] |
{
"url": "https://api.github.com/repos/huggingface/datasets/pulls/6727",
"html_url": "https://github.com/huggingface/datasets/pull/6727",
"diff_url": "https://github.com/huggingface/datasets/pull/6727.diff",
"patch_url": "https://github.com/huggingface/datasets/pull/6727.patch",
"merged_at": "2024-03-13T10:46:02"
}
| true
|
[
"The docs for this PR live [here](https://moon-ci-docs.huggingface.co/docs/datasets/pr_6727). All of your documentation changes will be reflected on that endpoint. The docs are available until 30 days after the last update.",
"looks like some files are missing in your google storage",
"cc @mariosasko is it related to https://github.com/huggingface/datasets/pull/6474 ? The files should ideally not move for backward compatibility anyway",
"@lhoestq All the files are still there.\r\n\r\nThe problem is that the `natural_questions` is now a no-code dataset, so the test's paths are no longer correct (unless the revision is pinned to the previous version). \r\n\r\n@psmyth94 This has been fixed on `main`, so you can make the CI tests green with the following:\r\n```python\r\ngit remote add upstream https://github.com/huggingface/datasets.git\r\ngit pull upstream main\r\ngit push\r\n```",
"Thank you @mariosasko ! I'm updating this branch if you don't mind @psmyth94 ",
"<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) | 0.004903 / 0.011353 (-0.006450) | 0.003105 / 0.011008 (-0.007903) | 0.061980 / 0.038508 (0.023471) | 0.029726 / 0.023109 (0.006617) | 0.243406 / 0.275898 (-0.032492) | 0.262530 / 0.323480 (-0.060950) | 0.003905 / 0.007986 (-0.004081) | 0.002617 / 0.004328 (-0.001712) | 0.047851 / 0.004250 (0.043601) | 0.040397 / 0.037052 (0.003345) | 0.259461 / 0.258489 (0.000972) | 0.285059 / 0.293841 (-0.008782) | 0.027321 / 0.128546 (-0.101225) | 0.009876 / 0.075646 (-0.065770) | 0.206999 / 0.419271 (-0.212273) | 0.034906 / 0.043533 (-0.008626) | 0.245120 / 0.255139 (-0.010019) | 0.270490 / 0.283200 (-0.012710) | 0.017341 / 0.141683 (-0.124342) | 1.128182 / 1.452155 (-0.323973) | 1.173024 / 1.492716 (-0.319693) |\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) | 0.089337 / 0.018006 (0.071331) | 0.298256 / 0.000490 (0.297767) | 0.000216 / 0.000200 (0.000016) | 0.000047 / 0.000054 (-0.000007) |\n\n### Benchmark: benchmark_indices_mapping.json\n\n| metric | select | shard | shuffle | sort | train_test_split |\n|--------|---|---|---|---|---|\n| new / old (diff) | 0.018179 / 0.037411 (-0.019233) | 0.061275 / 0.014526 (0.046749) | 0.073137 / 0.176557 (-0.103419) | 0.119603 / 0.737135 (-0.617532) | 0.073969 / 0.296338 (-0.222370) |\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) | 0.283109 / 0.215209 (0.067900) | 2.765441 / 2.077655 (0.687787) | 1.471276 / 1.504120 (-0.032844) | 1.346365 / 1.541195 (-0.194830) | 1.360668 / 1.468490 (-0.107822) | 0.549947 / 4.584777 (-4.034830) | 2.344213 / 3.745712 (-1.401499) | 2.700905 / 5.269862 (-2.568956) | 1.689936 / 4.565676 (-2.875741) | 0.061985 / 0.424275 (-0.362290) | 0.004923 / 0.007607 (-0.002684) | 0.329833 / 0.226044 (0.103788) | 3.277580 / 2.268929 (1.008652) | 1.833987 / 55.444624 (-53.610638) | 1.571023 / 6.876477 (-5.305454) | 1.573259 / 2.142072 (-0.568813) | 0.627504 / 4.805227 (-4.177723) | 0.114106 / 6.500664 (-6.386558) | 0.041197 / 0.075469 (-0.034272) |\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) | 0.967400 / 1.841788 (-0.874388) | 11.046527 / 8.074308 (2.972219) | 9.542214 / 10.191392 (-0.649178) | 0.140745 / 0.680424 (-0.539679) | 0.013627 / 0.534201 (-0.520574) | 0.288429 / 0.579283 (-0.290855) | 0.260509 / 0.434364 (-0.173855) | 0.324704 / 0.540337 (-0.215633) | 0.419366 / 1.386936 (-0.967570) |\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) | 0.005123 / 0.011353 (-0.006230) | 0.003119 / 0.011008 (-0.007890) | 0.048931 / 0.038508 (0.010423) | 0.032067 / 0.023109 (0.008958) | 0.276825 / 0.275898 (0.000927) | 0.297589 / 0.323480 (-0.025890) | 0.004075 / 0.007986 (-0.003911) | 0.002579 / 0.004328 (-0.001750) | 0.047862 / 0.004250 (0.043612) | 0.044032 / 0.037052 (0.006980) | 0.289469 / 0.258489 (0.030980) | 0.327269 / 0.293841 (0.033428) | 0.029369 / 0.128546 (-0.099177) | 0.010180 / 0.075646 (-0.065466) | 0.057111 / 0.419271 (-0.362161) | 0.051046 / 0.043533 (0.007513) | 0.276758 / 0.255139 (0.021619) | 0.296084 / 0.283200 (0.012884) | 0.017376 / 0.141683 (-0.124306) | 1.154486 / 1.452155 (-0.297669) | 1.192699 / 1.492716 (-0.300018) |\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) | 0.085981 / 0.018006 (0.067974) | 0.296956 / 0.000490 (0.296466) | 0.000211 / 0.000200 (0.000011) | 0.000050 / 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) | 0.021239 / 0.037411 (-0.016172) | 0.074851 / 0.014526 (0.060326) | 0.085676 / 0.176557 (-0.090881) | 0.125876 / 0.737135 (-0.611259) | 0.087573 / 0.296338 (-0.208765) |\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) | 0.289220 / 0.215209 (0.074011) | 2.812342 / 2.077655 (0.734688) | 1.572886 / 1.504120 (0.068766) | 1.446442 / 1.541195 (-0.094752) | 1.458737 / 1.468490 (-0.009753) | 0.562010 / 4.584777 (-4.022767) | 2.422896 / 3.745712 (-1.322816) | 2.578408 / 5.269862 (-2.691454) | 1.689998 / 4.565676 (-2.875678) | 0.064782 / 0.424275 (-0.359493) | 0.005051 / 0.007607 (-0.002556) | 0.339982 / 0.226044 (0.113938) | 3.309882 / 2.268929 (1.040953) | 1.910273 / 55.444624 (-53.534351) | 1.649723 / 6.876477 (-5.226753) | 1.744073 / 2.142072 (-0.397999) | 0.651905 / 4.805227 (-4.153323) | 0.114606 / 6.500664 (-6.386058) | 0.040030 / 0.075469 (-0.035439) |\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) | 1.008374 / 1.841788 (-0.833414) | 11.547300 / 8.074308 (3.472992) | 9.966061 / 10.191392 (-0.225331) | 0.144874 / 0.680424 (-0.535550) | 0.014400 / 0.534201 (-0.519801) | 0.285435 / 0.579283 (-0.293848) | 0.274755 / 0.434364 (-0.159609) | 0.323105 / 0.540337 (-0.217232) | 0.439172 / 1.386936 (-0.947764) |\n\n</details>\n</details>\n\n![](https://cml.dev/watermark.png#4591ac120e9d6c082b2479d2005c04b9c36f539c \"CML watermark\")\n"
] |
6,726
|
Profiling for HF Filesystem shows there are easy performance gains to be made
|
### Describe the bug
# Let's make it faster
First, an evidence...
Figure 1: CProfile for loading 3 files from cerebras/SlimPajama-627B train split, and 3 files from test split using streaming=True. X axis is 1106 seconds long.
See? It's pretty slow.
What is resolve pattern doing?
```
resolve_pattern called with **/train/** and hf://datasets/cerebras/SlimPajama-627B@2d0accdd58c5d5511943ca1f5ff0e3eb5e293543
resolve_pattern took 20.815081119537354 seconds
```
Makes sense. How to improve it?
## Bigger project, biggest payoff
Databricks (and consequently, spark) store a compressed manifest file of the files contained in the remote filesystem.
Then, you download one tiny file, decompress it, and all the operations are local instead of this shenanigans.
It seems pretty straightforward to make dataset uploads compute a manifest and upload it alongside their data.
This would make resolution time so fast that nobody would ever think about it again.
It also means you either need to have the uploader compute it _every time_, or have a hook that computes it.
## Smaller project, immediate payoff: Be diligent in avoiding deepcopy
Revise the _ls_tree method to avoid deepcopy:
```
def _ls_tree(
self,
path: str,
recursive: bool = False,
refresh: bool = False,
revision: Optional[str] = None,
expand_info: bool = True,
):
..... omitted .....
for path_info in tree:
if isinstance(path_info, RepoFile):
cache_path_info = {
"name": root_path + "/" + path_info.path,
"size": path_info.size,
"type": "file",
"blob_id": path_info.blob_id,
"lfs": path_info.lfs,
"last_commit": path_info.last_commit,
"security": path_info.security,
}
else:
cache_path_info = {
"name": root_path + "/" + path_info.path,
"size": 0,
"type": "directory",
"tree_id": path_info.tree_id,
"last_commit": path_info.last_commit,
}
parent_path = self._parent(cache_path_info["name"])
self.dircache.setdefault(parent_path, []).append(cache_path_info)
out.append(cache_path_info)
return copy.deepcopy(out) # copy to not let users modify the dircache
```
Observe this deepcopy at the end. It is making a copy of a very simple data structure. We do not need to copy. We can simply generate the data structure twice instead. It will be much faster.
```
def _ls_tree(
self,
path: str,
recursive: bool = False,
refresh: bool = False,
revision: Optional[str] = None,
expand_info: bool = True,
):
..... omitted .....
def make_cache_path_info(path_info):
if isinstance(path_info, RepoFile):
return {
"name": root_path + "/" + path_info.path,
"size": path_info.size,
"type": "file",
"blob_id": path_info.blob_id,
"lfs": path_info.lfs,
"last_commit": path_info.last_commit,
"security": path_info.security,
}
else:
return {
"name": root_path + "/" + path_info.path,
"size": 0,
"type": "directory",
"tree_id": path_info.tree_id,
"last_commit": path_info.last_commit,
}
for path_info in tree:
cache_path_info = make_cache_path_info(path_info)
out_cache_path_info = make_cache_path_info(path_info) # copy to not let users modify the dircache
parent_path = self._parent(cache_path_info["name"])
self.dircache.setdefault(parent_path, []).append(cache_path_info)
out.append(out_cache_path_info)
return out
```
Note there is no longer a deepcopy in this method. We have replaced it with generating the output twice. This is substantially faster. For me, the entire resolution went from 1100s to 360s.
## Medium project, medium payoff
After the above change, we have this profile:
Figure 2: x-axis is 355 seconds. Note that globbing and _ls_tree deep copy is gone. No surprise there. It's much faster now, but we still spend ~187seconds in get_fs_token_paths.
Well get_fs_token_paths is part of fsspec. We don't need to fix that because we can trust their developers to write high performance code. Probably the caller has misconfigured something. Let's take a look at the storage_options being provided to the filesystem that is constructed during this call.
Ah yes, streaming_download_manager::_prepare_single_hop_path_and_storage_options. We know streaming download manager is not compatible with async right now, but we really need this specific part of the code to be async. We're spending so much time checking isDir on the remote filesystem, it's a huge waste.
We can make the call easily 20-30x faster by using async, removing this performance bottleneck almost entirely (and reducing the total time of this part of the code to <30s. There is no reason to block async isDir calls for streaming.
I'm not going to mess w/ this one myself; I didn't write the streaming impl, and I don't know how it works, but I know the isDir check can be async.
### Steps to reproduce the bug
```
with cProfile.Profile() as pr:
pr.enable()
# Begin Data
if not os.path.exists(data_cache_dir):
os.makedirs(data_cache_dir, exist_ok=True)
training_dataset = load_dataset(training_dataset_name, split=training_split, cache_dir=data_cache_dir, streaming=True).take(training_slice)
eval_dataset = load_dataset(eval_dataset_name, split=eval_split, cache_dir=data_cache_dir, streaming=True).take(eval_slice)
# End Data
pr.disable()
pr.create_stats()
if not os.path.exists(profiling_path):
os.makedirs(profiling_path, exist_ok=True)
pr.dump_stats(os.path.join(profiling_path, "cprofile.prof"))
```
run this code for "cerebras/SlimPajama-627B" and whatever other params
### Expected behavior
Something better.
### Environment info
- `datasets` version: 2.18.0
- Platform: Linux-5.15.146.1-microsoft-standard-WSL2-x86_64-with-glibc2.35
- Python version: 3.10.13
- `huggingface_hub` version: 0.21.3
- PyArrow version: 15.0.0
- Pandas version: 2.2.1
- `fsspec` version: 2024.2.0
|
[] |
{
"url": null,
"html_url": null,
"diff_url": null,
"patch_url": null,
"merged_at": null
}
| false
|
[
"FWIW I debugged this while waiting for it to go",
"Oh I forgot to mention you can also cache resolve_pattern, and that seemed to also substantially improves things, if you want to load a dataset twice for whatever reason."
] |
6,725
|
Request for a comparison of huggingface datasets compared with other data format especially webdataset
|
### Feature request
Request for a comparison of huggingface datasets compared with other data format especially webdataset
### Motivation
I see huggingface datasets uses Apache Arrow as its backend, it seems to be great, but I'm curious about how it is good compared with other dataset format, like webdataset, what's the pros/cons of them.
### Your contribution
More information
|
[
{
"id": 1935892871,
"node_id": "MDU6TGFiZWwxOTM1ODkyODcx",
"url": "https://api.github.com/repos/huggingface/datasets/labels/enhancement",
"name": "enhancement",
"color": "a2eeef",
"default": true,
"description": "New feature or request"
}
] |
{
"url": null,
"html_url": null,
"diff_url": null,
"patch_url": null,
"merged_at": null
}
| false
|
[] |
6,724
|
Dataset with loading script does not work in renamed repos
|
### Describe the bug
My data repository was first called `BramVanroy/hplt-mono-v1-2` but I then renamed to use underscores instead of dashes. However, it seems that `datasets` retrieves the old repo name when it checks whether the repo contains data loading scripts in this line.
https://github.com/huggingface/datasets/blob/6fb6c834f008996c994b0a86c3808d0a33d44525/src/datasets/load.py#L1845
When I print `filename` it returns `hplt-mono-v1-2.py` but the files in the repo are of course `['.gitattributes', 'README.md', 'hplt_mono_v1_2.py']`. So the `filename` is the original reponame instead of the renamed one.
I am not sure if this is a caching issue or not or how I can resolve it.
### Steps to reproduce the bug
```
from datasets import load_dataset
ds = load_dataset(
"BramVanroy/hplt-mono-v1-2",
"ky",
trust_remote_code=True
)
```
### Expected behavior
That the most recent repo name is used when `filename` is generated.
### Environment info
- `datasets` version: 2.16.1
- Platform: Linux-5.14.0-284.25.1.el9_2.x86_64-x86_64-with-glibc2.34
- Python version: 3.10.13
- `huggingface_hub` version: 0.20.2
- PyArrow version: 14.0.1
- Pandas version: 2.1.3
- `fsspec` version: 2023.10.0
|
[] |
{
"url": null,
"html_url": null,
"diff_url": null,
"patch_url": null,
"merged_at": null
}
| false
|
[] |
6,723
|
get_dataset_default_config_name docstring
|
fix https://github.com/huggingface/datasets/pull/6722
|
[] |
{
"url": "https://api.github.com/repos/huggingface/datasets/pulls/6723",
"html_url": "https://github.com/huggingface/datasets/pull/6723",
"diff_url": "https://github.com/huggingface/datasets/pull/6723.diff",
"patch_url": "https://github.com/huggingface/datasets/pull/6723.patch",
"merged_at": "2024-03-07T17:21:20"
}
| true
|
[
"The docs for this PR live [here](https://moon-ci-docs.huggingface.co/docs/datasets/pr_6723). All of your documentation changes will be reflected on that endpoint. The docs are available until 30 days after the last update.",
"<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) | 0.005658 / 0.011353 (-0.005694) | 0.003883 / 0.011008 (-0.007125) | 0.064007 / 0.038508 (0.025499) | 0.030370 / 0.023109 (0.007261) | 0.246677 / 0.275898 (-0.029221) | 0.270846 / 0.323480 (-0.052634) | 0.003102 / 0.007986 (-0.004884) | 0.002931 / 0.004328 (-0.001397) | 0.049446 / 0.004250 (0.045196) | 0.043555 / 0.037052 (0.006503) | 0.261810 / 0.258489 (0.003321) | 0.289705 / 0.293841 (-0.004136) | 0.028676 / 0.128546 (-0.099870) | 0.010778 / 0.075646 (-0.064868) | 0.210604 / 0.419271 (-0.208667) | 0.035987 / 0.043533 (-0.007546) | 0.248034 / 0.255139 (-0.007105) | 0.265019 / 0.283200 (-0.018181) | 0.018522 / 0.141683 (-0.123161) | 1.096364 / 1.452155 (-0.355791) | 1.152750 / 1.492716 (-0.339966) |\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) | 0.093987 / 0.018006 (0.075981) | 0.306143 / 0.000490 (0.305653) | 0.000218 / 0.000200 (0.000018) | 0.000045 / 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) | 0.018727 / 0.037411 (-0.018685) | 0.061983 / 0.014526 (0.047457) | 0.074254 / 0.176557 (-0.102303) | 0.121256 / 0.737135 (-0.615880) | 0.076756 / 0.296338 (-0.219582) |\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) | 0.278824 / 0.215209 (0.063615) | 2.815960 / 2.077655 (0.738305) | 1.472946 / 1.504120 (-0.031174) | 1.349722 / 1.541195 (-0.191473) | 1.327844 / 1.468490 (-0.140646) | 0.574964 / 4.584777 (-4.009813) | 2.403458 / 3.745712 (-1.342254) | 2.769293 / 5.269862 (-2.500569) | 1.736970 / 4.565676 (-2.828706) | 0.063144 / 0.424275 (-0.361131) | 0.004983 / 0.007607 (-0.002625) | 0.331212 / 0.226044 (0.105168) | 3.231496 / 2.268929 (0.962567) | 1.798487 / 55.444624 (-53.646138) | 1.523010 / 6.876477 (-5.353467) | 1.559973 / 2.142072 (-0.582099) | 0.657036 / 4.805227 (-4.148191) | 0.119084 / 6.500664 (-6.381580) | 0.042982 / 0.075469 (-0.032487) |\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) | 0.976433 / 1.841788 (-0.865355) | 11.475946 / 8.074308 (3.401638) | 9.339369 / 10.191392 (-0.852023) | 0.141761 / 0.680424 (-0.538662) | 0.014506 / 0.534201 (-0.519695) | 0.289944 / 0.579283 (-0.289340) | 0.273667 / 0.434364 (-0.160697) | 0.326682 / 0.540337 (-0.213655) | 0.458946 / 1.386936 (-0.927990) |\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) | 0.005194 / 0.011353 (-0.006159) | 0.003713 / 0.011008 (-0.007295) | 0.049297 / 0.038508 (0.010789) | 0.029723 / 0.023109 (0.006614) | 0.278664 / 0.275898 (0.002766) | 0.296387 / 0.323480 (-0.027093) | 0.004215 / 0.007986 (-0.003771) | 0.002680 / 0.004328 (-0.001648) | 0.048276 / 0.004250 (0.044025) | 0.044454 / 0.037052 (0.007402) | 0.290510 / 0.258489 (0.032021) | 0.319028 / 0.293841 (0.025187) | 0.029177 / 0.128546 (-0.099369) | 0.010361 / 0.075646 (-0.065285) | 0.056993 / 0.419271 (-0.362279) | 0.050765 / 0.043533 (0.007232) | 0.278234 / 0.255139 (0.023095) | 0.295848 / 0.283200 (0.012649) | 0.018776 / 0.141683 (-0.122906) | 1.134866 / 1.452155 (-0.317288) | 1.204083 / 1.492716 (-0.288634) |\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) | 0.094397 / 0.018006 (0.076391) | 0.304693 / 0.000490 (0.304203) | 0.000207 / 0.000200 (0.000007) | 0.000044 / 0.000054 (-0.000010) |\n\n### Benchmark: benchmark_indices_mapping.json\n\n| metric | select | shard | shuffle | sort | train_test_split |\n|--------|---|---|---|---|---|\n| new / old (diff) | 0.021322 / 0.037411 (-0.016090) | 0.075384 / 0.014526 (0.060859) | 0.086961 / 0.176557 (-0.089596) | 0.124424 / 0.737135 (-0.612711) | 0.087802 / 0.296338 (-0.208536) |\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) | 0.305542 / 0.215209 (0.090333) | 2.980678 / 2.077655 (0.903023) | 1.632348 / 1.504120 (0.128228) | 1.501466 / 1.541195 (-0.039728) | 1.517681 / 1.468490 (0.049191) | 0.579318 / 4.584777 (-4.005459) | 2.460734 / 3.745712 (-1.284978) | 2.650164 / 5.269862 (-2.619697) | 1.752061 / 4.565676 (-2.813615) | 0.064561 / 0.424275 (-0.359714) | 0.005097 / 0.007607 (-0.002510) | 0.359613 / 0.226044 (0.133569) | 3.518549 / 2.268929 (1.249620) | 1.962575 / 55.444624 (-53.482050) | 1.686108 / 6.876477 (-5.190369) | 1.787873 / 2.142072 (-0.354199) | 0.653715 / 4.805227 (-4.151512) | 0.117617 / 6.500664 (-6.383048) | 0.040359 / 0.075469 (-0.035110) |\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) | 1.021533 / 1.841788 (-0.820255) | 11.974817 / 8.074308 (3.900509) | 10.073530 / 10.191392 (-0.117862) | 0.141477 / 0.680424 (-0.538947) | 0.015081 / 0.534201 (-0.519120) | 0.292622 / 0.579283 (-0.286661) | 0.291043 / 0.434364 (-0.143321) | 0.347822 / 0.540337 (-0.192516) | 0.443647 / 1.386936 (-0.943289) |\n\n</details>\n</details>\n\n![](https://cml.dev/watermark.png#6fb6c834f008996c994b0a86c3808d0a33d44525 \"CML watermark\")\n"
] |
6,722
|
Add details in docstring
|
see https://github.com/huggingface/datasets-server/pull/2554#discussion_r1516516867
|
[] |
{
"url": "https://api.github.com/repos/huggingface/datasets/pulls/6722",
"html_url": "https://github.com/huggingface/datasets/pull/6722",
"diff_url": "https://github.com/huggingface/datasets/pull/6722.diff",
"patch_url": "https://github.com/huggingface/datasets/pull/6722.patch",
"merged_at": null
}
| true
|
[
"The docs for this PR live [here](https://moon-ci-docs.huggingface.co/docs/datasets/pr_6722). All of your documentation changes will be reflected on that endpoint. The docs are available until 30 days after the last update."
] |
6,721
|
Hi,do you know how to load the dataset from local file now?
|
Hi, if I want to load the dataset from local file, then how to specify the configuration name?
_Originally posted by @WHU-gentle in https://github.com/huggingface/datasets/issues/2976#issuecomment-1333455222_
|
[] |
{
"url": null,
"html_url": null,
"diff_url": null,
"patch_url": null,
"merged_at": null
}
| false
|
[
"\r\n@Gera001\r\n# Loading Dataset from Local Files Using 🤗Hugging Face.\r\n\r\nTo load a dataset from local files using the Hugging Face datasets library, you can use the `load_dataset` function.\r\n\r\n```\r\nfrom datasets import load_dataset\r\ndataset = load_dataset('csv', data_files={'train': 'path/to/train.csv',\r\n 'test': 'path/to/test.csv'})\r\n```\r\n\r\nReference to [HF Datasets docs for loading from local](https://huggingface.co/docs/datasets/en/loading#csv). \r\n\r\n@albertvillanova\r\nthis issue can be closed here.",
"like this: from datasets import load_from_disk\r\ndataset = load_from_disk(data_path)\r\n",
"@ge00009 \r\n> like this: from datasets import load_from_disk dataset = load_from_disk(data_path)\r\n\r\nLoads a dataset that was previously saved using `save_to_disk()`.\r\n\r\nReference link:\r\nhttps://huggingface.co/docs/datasets/en/package_reference/loading_methods#datasets.load_from_disk.example"
] |
6,720
|
TypeError: 'str' object is not callable
|
### Describe the bug
I am trying to get the HPLT datasets on the hub. Downloading/re-uploading would be too time- and resource consuming so I wrote [a dataset loader script](https://huggingface.co/datasets/BramVanroy/hplt_mono_v1_2/blob/main/hplt_mono_v1_2.py). I think I am very close but for some reason I always get the error below. It happens during the clean-up phase where the directory cannot be removed because it is not empty.
My only guess would be that this may have to do with zstandard
```
Traceback (most recent call last):
File "/home/local/vanroy/dutch-instruction-datasets/.venv/lib/python3.10/site-packages/datasets/builder.py", line 1744, in _prepare_split_single
writer.write(example, key)
File "/home/local/vanroy/dutch-instruction-datasets/.venv/lib/python3.10/site-packages/datasets/arrow_writer.py", line 492, in write
self.write_examples_on_file()
File "/home/local/vanroy/dutch-instruction-datasets/.venv/lib/python3.10/site-packages/datasets/arrow_writer.py", line 434, in write_examples_on_file
if self.schema
File "/home/local/vanroy/dutch-instruction-datasets/.venv/lib/python3.10/site-packages/datasets/arrow_writer.py", line 409, in schema
else (pa.schema(self._features.type) if self._features is not None else None)
File "/home/local/vanroy/dutch-instruction-datasets/.venv/lib/python3.10/site-packages/datasets/features/features.py", line 1643, in type
return get_nested_type(self)
File "/home/local/vanroy/dutch-instruction-datasets/.venv/lib/python3.10/site-packages/datasets/features/features.py", line 1209, in get_nested_type
{key: get_nested_type(schema[key]) for key in schema}
File "/home/local/vanroy/dutch-instruction-datasets/.venv/lib/python3.10/site-packages/datasets/features/features.py", line 1209, in <dictcomp>
{key: get_nested_type(schema[key]) for key in schema}
File "/home/local/vanroy/dutch-instruction-datasets/.venv/lib/python3.10/site-packages/datasets/features/features.py", line 1221, in get_nested_type
value_type = get_nested_type(schema.feature)
File "/home/local/vanroy/dutch-instruction-datasets/.venv/lib/python3.10/site-packages/datasets/features/features.py", line 1228, in get_nested_type
return schema()
TypeError: 'str' object is not callable
During handling of the above exception, another exception occurred:
Traceback (most recent call last):
File "/home/local/vanroy/dutch-instruction-datasets/.venv/lib/python3.10/site-packages/datasets/builder.py", line 1753, in _prepare_split_single
num_examples, num_bytes = writer.finalize()
File "/home/local/vanroy/dutch-instruction-datasets/.venv/lib/python3.10/site-packages/datasets/arrow_writer.py", line 588, in finalize
self.write_examples_on_file()
File "/home/local/vanroy/dutch-instruction-datasets/.venv/lib/python3.10/site-packages/datasets/arrow_writer.py", line 434, in write_examples_on_file
if self.schema
File "/home/local/vanroy/dutch-instruction-datasets/.venv/lib/python3.10/site-packages/datasets/arrow_writer.py", line 409, in schema
else (pa.schema(self._features.type) if self._features is not None else None)
File "/home/local/vanroy/dutch-instruction-datasets/.venv/lib/python3.10/site-packages/datasets/features/features.py", line 1643, in type
return get_nested_type(self)
File "/home/local/vanroy/dutch-instruction-datasets/.venv/lib/python3.10/site-packages/datasets/features/features.py", line 1209, in get_nested_type
{key: get_nested_type(schema[key]) for key in schema}
File "/home/local/vanroy/dutch-instruction-datasets/.venv/lib/python3.10/site-packages/datasets/features/features.py", line 1209, in <dictcomp>
{key: get_nested_type(schema[key]) for key in schema}
File "/home/local/vanroy/dutch-instruction-datasets/.venv/lib/python3.10/site-packages/datasets/features/features.py", line 1221, in get_nested_type
value_type = get_nested_type(schema.feature)
File "/home/local/vanroy/dutch-instruction-datasets/.venv/lib/python3.10/site-packages/datasets/features/features.py", line 1228, in get_nested_type
return schema()
TypeError: 'str' object is not callable
The above exception was the direct cause of the following exception:
Traceback (most recent call last):
File "/home/local/vanroy/dutch-instruction-datasets/.venv/lib/python3.10/site-packages/datasets/builder.py", line 959, in incomplete_dir
yield tmp_dir
File "/home/local/vanroy/dutch-instruction-datasets/.venv/lib/python3.10/site-packages/datasets/builder.py", line 1005, in download_and_prepare
self._download_and_prepare(
File "/home/local/vanroy/dutch-instruction-datasets/.venv/lib/python3.10/site-packages/datasets/builder.py", line 1767, in _download_and_prepare
super()._download_and_prepare(
File "/home/local/vanroy/dutch-instruction-datasets/.venv/lib/python3.10/site-packages/datasets/builder.py", line 1100, in _download_and_prepare
self._prepare_split(split_generator, **prepare_split_kwargs)
File "/home/local/vanroy/dutch-instruction-datasets/.venv/lib/python3.10/site-packages/datasets/builder.py", line 1605, in _prepare_split
for job_id, done, content in self._prepare_split_single(
File "/home/local/vanroy/dutch-instruction-datasets/.venv/lib/python3.10/site-packages/datasets/builder.py", line 1762, in _prepare_split_single
raise DatasetGenerationError("An error occurred while generating the dataset") from e
datasets.exceptions.DatasetGenerationError: An error occurred while generating the dataset
During handling of the above exception, another exception occurred:
Traceback (most recent call last):
File "/home/pricie/vanroy/.config/JetBrains/PyCharm2023.3/scratches/scratch_5.py", line 4, in <module>
ds = load_dataset(
File "/home/local/vanroy/dutch-instruction-datasets/.venv/lib/python3.10/site-packages/datasets/load.py", line 2549, in load_dataset
builder_instance.download_and_prepare(
File "/home/local/vanroy/dutch-instruction-datasets/.venv/lib/python3.10/site-packages/datasets/builder.py", line 985, in download_and_prepare
with incomplete_dir(self._output_dir) as tmp_output_dir:
File "/home/pricie/vanroy/.pyenv/versions/3.10.13/lib/python3.10/contextlib.py", line 153, in __exit__
self.gen.throw(typ, value, traceback)
File "/home/local/vanroy/dutch-instruction-datasets/.venv/lib/python3.10/site-packages/datasets/builder.py", line 966, in incomplete_dir
shutil.rmtree(tmp_dir)
File "/home/pricie/vanroy/.pyenv/versions/3.10.13/lib/python3.10/shutil.py", line 731, in rmtree
onerror(os.rmdir, path, sys.exc_info())
File "/home/pricie/vanroy/.pyenv/versions/3.10.13/lib/python3.10/shutil.py", line 729, in rmtree
os.rmdir(path)
OSError: [Errno 39] Directory not empty: '/home/pricie/vanroy/.cache/huggingface/datasets/BramVanroy___hplt_mono_v1_2/ky/1.2.0/7ab138629fe7e9e29fe93ce63d809d5ef9d963273b829f61ab538e012dc9cc47.incomplete'
```
Interestingly, though, this directory _does_ appear to be empty:
```shell
> cd /home/pricie/vanroy/.cache/huggingface/datasets/BramVanroy___hplt_mono_v1_2/ky/1.2.0/7ab138629fe7e9e29fe93ce63d809d5ef9d963273b829f61ab538e012dc9cc47.incomplete
> ls -lah
total 0
drwxr-xr-x. 1 vanroy vanroy 0 Mar 7 12:01 .
drwxr-xr-x. 1 vanroy vanroy 304 Mar 7 11:52 ..
> cd ..
> ls
7ab138629fe7e9e29fe93ce63d809d5ef9d963273b829f61ab538e012dc9cc47_builder.lock 7ab138629fe7e9e29fe93ce63d809d5ef9d963273b829f61ab538e012dc9cc47.incomplete
```
### Steps to reproduce the bug
```python
from datasets import load_dataset
ds = load_dataset(
"BramVanroy/hplt_mono_v1_2",
"ky",
trust_remote_code=True
)
```
### Expected behavior
No error.
### Environment info
- `datasets` version: 2.16.1
- Platform: Linux-5.14.0-284.25.1.el9_2.x86_64-x86_64-with-glibc2.34
- Python version: 3.10.13
- `huggingface_hub` version: 0.20.2
- PyArrow version: 14.0.1
- Pandas version: 2.1.3
- `fsspec` version: 2023.10.0
|
[] |
{
"url": null,
"html_url": null,
"diff_url": null,
"patch_url": null,
"merged_at": null
}
| false
|
[
"Hi ! I opened a PR to fix an issue in the Features defined in your code\r\n\r\nBasically changing\r\n```python\r\nSequence(\"float32\")\r\n```\r\n\r\nto\r\n```python\r\nSequence(Value(\"float32\"))\r\n```\r\n\r\n\r\nhttps://huggingface.co/datasets/BramVanroy/hplt_mono_v1_2/discussions/1",
"D'oh! Was wondering why the `str() is not callable` was in there. Glad the error is my end though, and not related to zstandard (which I had not used in the past).\r\n\r\nThanks a lot!"
] |
6,719
|
Is there any way to solve hanging of IterableDataset using split by node + filtering during inference
|
### Describe the bug
I am using an iterable dataset in a multi-node setup, trying to do training/inference while filtering the data on the fly. I usually do not use `split_dataset_by_node` but it is very slow using the IterableDatasetShard in `accelerate` and `transformers`. When I filter after applying `split_dataset_by_node`, it results in shards that are not equal sizes due to unequal samples filtered from each one.
The distributed process hangs when trying to accomplish this. Is there any way to resolve this or is it impossible to implement?
### Steps to reproduce the bug
Here is a toy example of what I am trying to do that reproduces the behavior
```
# torchrun --nproc-per-node 2 file.py
import os
import pandas as pd
import torch
from accelerate import Accelerator
from datasets import Features, Value, load_dataset
from datasets.distributed import split_dataset_by_node
from torch.utils.data import DataLoader
accelerator = Accelerator(device_placement=True, dispatch_batches=False)
if accelerator.is_main_process:
if not os.path.exists("scratch_data"):
os.mkdir("scratch_data")
n_shards = 4
for i in range(n_shards):
df = pd.DataFrame({"id": list(range(10 * i, 10 * (i + 1)))})
df.to_parquet(f"scratch_data/shard_{i}.parquet")
world_size = accelerator.num_processes
local_rank = accelerator.process_index
def collate_fn(examples):
input_ids = []
for example in examples:
input_ids.append(example["id"])
return torch.LongTensor(input_ids)
dataset = load_dataset(
"parquet", data_dir="scratch_data", split="train", streaming=True
)
dataset = (
split_dataset_by_node(dataset, rank=local_rank, world_size=world_size)
.filter(lambda x: x["id"] < 35)
.shuffle(seed=42, buffer_size=100)
)
batch_size = 2
train_dataloader = DataLoader(
dataset,
batch_size=batch_size,
collate_fn=collate_fn,
num_workers=2
)
for x in train_dataloader:
x = x.to(accelerator.device)
print({"rank": local_rank, "id": x})
y = accelerator.gather_for_metrics(x)
if accelerator.is_main_process:
print("gathered", y)
```
### Expected behavior
Is there any way to continue training/inference on the GPUs that have remaining data left without waiting for the others? Is it impossible to filter when
### Environment info
- `datasets` version: 2.18.0
- Platform: Linux-5.10.209-198.812.amzn2.x86_64-x86_64-with-glibc2.31
- Python version: 3.10.13
- `huggingface_hub` version: 0.21.3
- PyArrow version: 15.0.0
- Pandas version: 2.2.1
- `fsspec` version: 2023.6.0
|
[] |
{
"url": null,
"html_url": null,
"diff_url": null,
"patch_url": null,
"merged_at": null
}
| false
|
[] |
6,718
|
Fix concurrent script loading with force_redownload
|
I added `lock_importable_file` in `get_dataset_builder_class` and `extend_dataset_builder_for_streaming` to fix the issue, and I also added a test
cc @clefourrier
|
[] |
{
"url": "https://api.github.com/repos/huggingface/datasets/pulls/6718",
"html_url": "https://github.com/huggingface/datasets/pull/6718",
"diff_url": "https://github.com/huggingface/datasets/pull/6718.diff",
"patch_url": "https://github.com/huggingface/datasets/pull/6718.patch",
"merged_at": "2024-03-07T13:58:04"
}
| true
|
[
"The docs for this PR live [here](https://moon-ci-docs.huggingface.co/docs/datasets/pr_6718). All of your documentation changes will be reflected on that endpoint. The docs are available until 30 days after the last update.",
"<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) | 0.005074 / 0.011353 (-0.006279) | 0.003505 / 0.011008 (-0.007503) | 0.063683 / 0.038508 (0.025175) | 0.029308 / 0.023109 (0.006199) | 0.246648 / 0.275898 (-0.029250) | 0.265546 / 0.323480 (-0.057933) | 0.004108 / 0.007986 (-0.003878) | 0.002683 / 0.004328 (-0.001646) | 0.048634 / 0.004250 (0.044383) | 0.043786 / 0.037052 (0.006733) | 0.262197 / 0.258489 (0.003708) | 0.291582 / 0.293841 (-0.002259) | 0.027472 / 0.128546 (-0.101074) | 0.010213 / 0.075646 (-0.065434) | 0.206744 / 0.419271 (-0.212527) | 0.036195 / 0.043533 (-0.007337) | 0.249090 / 0.255139 (-0.006049) | 0.280002 / 0.283200 (-0.003198) | 0.018568 / 0.141683 (-0.123115) | 1.124844 / 1.452155 (-0.327311) | 1.159358 / 1.492716 (-0.333359) |\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) | 0.093186 / 0.018006 (0.075180) | 0.302331 / 0.000490 (0.301842) | 0.000217 / 0.000200 (0.000017) | 0.000046 / 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) | 0.018727 / 0.037411 (-0.018684) | 0.061730 / 0.014526 (0.047204) | 0.074330 / 0.176557 (-0.102226) | 0.119769 / 0.737135 (-0.617366) | 0.075611 / 0.296338 (-0.220727) |\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) | 0.285063 / 0.215209 (0.069854) | 2.824809 / 2.077655 (0.747155) | 1.481858 / 1.504120 (-0.022262) | 1.350193 / 1.541195 (-0.191002) | 1.358012 / 1.468490 (-0.110478) | 0.557842 / 4.584777 (-4.026935) | 2.380729 / 3.745712 (-1.364983) | 2.798891 / 5.269862 (-2.470970) | 1.719288 / 4.565676 (-2.846388) | 0.061705 / 0.424275 (-0.362570) | 0.005431 / 0.007607 (-0.002176) | 0.343233 / 0.226044 (0.117189) | 3.375223 / 2.268929 (1.106295) | 1.838188 / 55.444624 (-53.606436) | 1.570015 / 6.876477 (-5.306461) | 1.573157 / 2.142072 (-0.568915) | 0.650678 / 4.805227 (-4.154549) | 0.116412 / 6.500664 (-6.384252) | 0.041754 / 0.075469 (-0.033715) |\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) | 0.970431 / 1.841788 (-0.871357) | 11.317128 / 8.074308 (3.242819) | 9.691240 / 10.191392 (-0.500152) | 0.142260 / 0.680424 (-0.538164) | 0.014131 / 0.534201 (-0.520070) | 0.289910 / 0.579283 (-0.289373) | 0.265648 / 0.434364 (-0.168715) | 0.323130 / 0.540337 (-0.217208) | 0.447005 / 1.386936 (-0.939931) |\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) | 0.005322 / 0.011353 (-0.006031) | 0.003755 / 0.011008 (-0.007253) | 0.049646 / 0.038508 (0.011138) | 0.029669 / 0.023109 (0.006560) | 0.284151 / 0.275898 (0.008253) | 0.298351 / 0.323480 (-0.025128) | 0.004183 / 0.007986 (-0.003803) | 0.002683 / 0.004328 (-0.001645) | 0.048814 / 0.004250 (0.044563) | 0.045017 / 0.037052 (0.007965) | 0.287358 / 0.258489 (0.028869) | 0.317394 / 0.293841 (0.023553) | 0.030025 / 0.128546 (-0.098521) | 0.010854 / 0.075646 (-0.064793) | 0.058694 / 0.419271 (-0.360578) | 0.052287 / 0.043533 (0.008754) | 0.279038 / 0.255139 (0.023899) | 0.295442 / 0.283200 (0.012242) | 0.019413 / 0.141683 (-0.122270) | 1.146106 / 1.452155 (-0.306048) | 1.197777 / 1.492716 (-0.294939) |\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) | 0.092191 / 0.018006 (0.074184) | 0.302672 / 0.000490 (0.302182) | 0.000623 / 0.000200 (0.000423) | 0.000048 / 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) | 0.022067 / 0.037411 (-0.015345) | 0.081760 / 0.014526 (0.067235) | 0.087548 / 0.176557 (-0.089009) | 0.126405 / 0.737135 (-0.610730) | 0.089331 / 0.296338 (-0.207008) |\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) | 0.295821 / 0.215209 (0.080612) | 2.897930 / 2.077655 (0.820276) | 1.604500 / 1.504120 (0.100380) | 1.471502 / 1.541195 (-0.069692) | 1.497918 / 1.468490 (0.029428) | 0.576179 / 4.584777 (-4.008598) | 2.452103 / 3.745712 (-1.293609) | 2.668043 / 5.269862 (-2.601818) | 1.753544 / 4.565676 (-2.812133) | 0.064410 / 0.424275 (-0.359865) | 0.005027 / 0.007607 (-0.002580) | 0.351509 / 0.226044 (0.125465) | 3.479208 / 2.268929 (1.210280) | 1.990356 / 55.444624 (-53.454269) | 1.684920 / 6.876477 (-5.191556) | 1.794251 / 2.142072 (-0.347821) | 0.662692 / 4.805227 (-4.142535) | 0.118589 / 6.500664 (-6.382076) | 0.040813 / 0.075469 (-0.034656) |\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) | 1.002390 / 1.841788 (-0.839398) | 12.004617 / 8.074308 (3.930309) | 10.216005 / 10.191392 (0.024613) | 0.154354 / 0.680424 (-0.526070) | 0.015554 / 0.534201 (-0.518647) | 0.288741 / 0.579283 (-0.290542) | 0.276774 / 0.434364 (-0.157590) | 0.327055 / 0.540337 (-0.213282) | 0.435121 / 1.386936 (-0.951815) |\n\n</details>\n</details>\n\n![](https://cml.dev/watermark.png#f45bc6caa25115a04c41b278671a5a89457eb66c \"CML watermark\")\n"
] |
6,717
|
`remove_columns` method used with a streaming enable dataset mode produces a LibsndfileError on multichannel audio
|
### Describe the bug
When loading a HF dataset in streaming mode and removing some columns, it is impossible to load a sample if the audio contains more than one channel. I have the impression that the time axis and channels are swapped or concatenated.
### Steps to reproduce the bug
Minimal error code:
```python
from datasets import load_dataset
dataset_name = "zinc75/Vibravox_dummy"
config_name = "BWE_Larynx_microphone"
# if we use "ASR_Larynx_microphone" subset which is a monochannel audio, no error is thrown.
dataset = load_dataset(
path=dataset_name, name=config_name, split="train", streaming=True
)
dataset = dataset.remove_columns(["sensor_id"])
# dataset = dataset.map(lambda x:x, remove_columns=["sensor_id"])
# The commented version does not produce an error, but loses the dataset features.
sample = next(iter(dataset))
```
Error:
```
Traceback (most recent call last):
File "/home/julien/Bureau/github/vibravox/tmp.py", line 15, in <module>
sample = next(iter(dataset))
^^^^^^^^^^^^^^^^^^^
File "/home/julien/.pyenv/versions/vibravox/lib/python3.11/site-packages/datasets/iterable_dataset.py", line 1392, in __iter__
example = _apply_feature_types_on_example(
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/home/julien/.pyenv/versions/vibravox/lib/python3.11/site-packages/datasets/iterable_dataset.py", line 1080, in _apply_feature_types_on_example
encoded_example = features.encode_example(example)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/home/julien/.pyenv/versions/vibravox/lib/python3.11/site-packages/datasets/features/features.py", line 1889, in encode_example
return encode_nested_example(self, example)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/home/julien/.pyenv/versions/vibravox/lib/python3.11/site-packages/datasets/features/features.py", line 1244, in encode_nested_example
{k: encode_nested_example(schema[k], obj.get(k), level=level + 1) for k in schema}
File "/home/julien/.pyenv/versions/vibravox/lib/python3.11/site-packages/datasets/features/features.py", line 1244, in <dictcomp>
{k: encode_nested_example(schema[k], obj.get(k), level=level + 1) for k in schema}
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/home/julien/.pyenv/versions/vibravox/lib/python3.11/site-packages/datasets/features/features.py", line 1300, in encode_nested_example
return schema.encode_example(obj) if obj is not None else None
^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/home/julien/.pyenv/versions/vibravox/lib/python3.11/site-packages/datasets/features/audio.py", line 98, in encode_example
sf.write(buffer, value["array"], value["sampling_rate"], format="wav")
File "/home/julien/.pyenv/versions/vibravox/lib/python3.11/site-packages/soundfile.py", line 343, in write
with SoundFile(file, 'w', samplerate, channels,
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/home/julien/.pyenv/versions/vibravox/lib/python3.11/site-packages/soundfile.py", line 658, in __init__
self._file = self._open(file, mode_int, closefd)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/home/julien/.pyenv/versions/vibravox/lib/python3.11/site-packages/soundfile.py", line 1216, in _open
raise LibsndfileError(err, prefix="Error opening {0!r}: ".format(self.name))
soundfile.LibsndfileError: Error opening <_io.BytesIO object at 0x7fd795d24680>: Format not recognised.
Process finished with exit code 1
```
### Expected behavior
I would expect this code to run without error.
### Environment info
- `datasets` version: 2.18.0
- Platform: Linux-6.5.0-21-generic-x86_64-with-glibc2.35
- Python version: 3.11.0
- `huggingface_hub` version: 0.21.3
- PyArrow version: 15.0.0
- Pandas version: 2.2.1
- `fsspec` version: 2023.10.0
|
[] |
{
"url": null,
"html_url": null,
"diff_url": null,
"patch_url": null,
"merged_at": null
}
| false
|
[
"And it also works well with `dataset = dataset.select_columns([\"audio\"])`"
] |
6,716
|
Non-deterministic `Dataset.builder_name` value
|
### Describe the bug
I'm not sure if this is a bug, but `print(ds.builder_name)` in the following code sometimes prints out `rotten_tomatoes` instead of `parquet`:
```python
import datasets
for _ in range(100):
ds = datasets.load_dataset("rotten_tomatoes", split="train")
print(ds.builder_name) # prints out "rotten_tomatoes" sometimes instead of "parquet"
```
Output:
```
...
parquet
parquet
parquet
rotten_tomatoes
parquet
parquet
parquet
...
```
Here's a reproduction using GitHub Actions:
https://github.com/mlflow/mlflow/actions/runs/8153247984/job/22284263613?pr=11329#step:12:241
One of our tests is flaky because `builder_name` is not deterministic.
### Steps to reproduce the bug
1. Run the code above.
### Expected behavior
Always prints out `parquet`?
### Environment info
```
Copy-and-paste the text below in your GitHub issue.
- `datasets` version: 2.18.0
- Platform: Linux-6.5.0-1015-azure-x86_64-with-glibc2.34
- Python version: 3.8.18
- `huggingface_hub` version: 0.21.3
- PyArrow version: 15.0.0
- Pandas version: 2.0.3
- `fsspec` version: 2024.2.0
```
|
[] |
{
"url": null,
"html_url": null,
"diff_url": null,
"patch_url": null,
"merged_at": null
}
| false
|
[
"When `rotten_tomatoes` is printed out, the following warning message is also printed out:\r\n\r\n```\r\nYou can avoid this message in future by passing the argument `trust_remote_code=True`.\r\nPassing `trust_remote_code=True` will be mandatory to load this dataset from the next major release of `datasets`.\r\n```",
"Hi ! This behavior happens because the dataset was originakky created using a dataset script [rotten_tomatoes.py](https://huggingface.co/datasets/rotten_tomatoes/blob/26f40d324d7b281d8b3fb1c47f30f8b9957f206b/rotten_tomatoes.py) and because we added features recently allowing to download the dataset directly from Parquet files (parquet builder) without running the dataset script (rotten_tomatoes). The flakiness must come from the availability of the Parquet files (we automatically export them in the refs/convert/parquet branch and we recently had to move some files).\r\n\r\nAnyway the easy fix on our side is to remove the dataset script completely, let me open a PR at https://huggingface.co/datasets/rotten_tomatoes\r\n\r\nEDIT: opened https://huggingface.co/datasets/rotten_tomatoes/discussions/6, feel free to comment there if you're ok with that change",
"@lhoestq Thanks for the comment, explanation, and patch!",
"> we automatically export them in the refs/convert/parquet branch\r\n\r\nWhen this operation is in progress, the parquet files become temporarily unavailable?",
"> When this operation is in progress, the parquet files become temporarily unavailable?\r\n\r\nYes correct. I just merged the patch btw :)",
"@lhoestq Thanks for merging the PR! I think this issue can be closed."
] |
6,715
|
Fix sliced ConcatenationTable pickling with mixed schemas vertically
|
A sliced + pickled ConcatenationTable could end up with a different schema than the original schema, if the slice only contains blocks with only a subset of the columns.
This can lead to issues when saving datasets from a concatenation of datasets with mixed schemas
Reported in https://discuss.huggingface.co/t/datasetdict-save-to-disk-with-num-proc-1-seems-to-hang-with-error/75595
|
[] |
{
"url": "https://api.github.com/repos/huggingface/datasets/pulls/6715",
"html_url": "https://github.com/huggingface/datasets/pull/6715",
"diff_url": "https://github.com/huggingface/datasets/pull/6715.diff",
"patch_url": "https://github.com/huggingface/datasets/pull/6715.patch",
"merged_at": "2024-03-05T11:17:04"
}
| true
|
[
"The docs for this PR live [here](https://moon-ci-docs.huggingface.co/docs/datasets/pr_6715). All of your documentation changes will be reflected on that endpoint. The docs are available until 30 days after the last update.",
"<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) | 0.005294 / 0.011353 (-0.006059) | 0.003598 / 0.011008 (-0.007411) | 0.062798 / 0.038508 (0.024290) | 0.027479 / 0.023109 (0.004370) | 0.247146 / 0.275898 (-0.028752) | 0.272103 / 0.323480 (-0.051377) | 0.002979 / 0.007986 (-0.005007) | 0.002701 / 0.004328 (-0.001628) | 0.049384 / 0.004250 (0.045134) | 0.041562 / 0.037052 (0.004510) | 0.269924 / 0.258489 (0.011435) | 0.290749 / 0.293841 (-0.003092) | 0.028285 / 0.128546 (-0.100261) | 0.010464 / 0.075646 (-0.065183) | 0.207000 / 0.419271 (-0.212272) | 0.036186 / 0.043533 (-0.007347) | 0.254524 / 0.255139 (-0.000615) | 0.274843 / 0.283200 (-0.008356) | 0.020044 / 0.141683 (-0.121638) | 1.119223 / 1.452155 (-0.332931) | 1.156557 / 1.492716 (-0.336159) |\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) | 0.092014 / 0.018006 (0.074008) | 0.297349 / 0.000490 (0.296859) | 0.000205 / 0.000200 (0.000005) | 0.000048 / 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) | 0.018617 / 0.037411 (-0.018794) | 0.061879 / 0.014526 (0.047354) | 0.072877 / 0.176557 (-0.103680) | 0.121850 / 0.737135 (-0.615286) | 0.074686 / 0.296338 (-0.221653) |\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) | 0.281204 / 0.215209 (0.065995) | 2.728688 / 2.077655 (0.651033) | 1.469659 / 1.504120 (-0.034461) | 1.355306 / 1.541195 (-0.185889) | 1.350598 / 1.468490 (-0.117892) | 0.563669 / 4.584777 (-4.021108) | 2.377177 / 3.745712 (-1.368535) | 2.767402 / 5.269862 (-2.502460) | 1.720188 / 4.565676 (-2.845489) | 0.062594 / 0.424275 (-0.361681) | 0.005004 / 0.007607 (-0.002603) | 0.333017 / 0.226044 (0.106972) | 3.354543 / 2.268929 (1.085615) | 1.840031 / 55.444624 (-53.604593) | 1.545548 / 6.876477 (-5.330929) | 1.569858 / 2.142072 (-0.572214) | 0.642680 / 4.805227 (-4.162547) | 0.117463 / 6.500664 (-6.383201) | 0.042472 / 0.075469 (-0.032997) |\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) | 0.977436 / 1.841788 (-0.864351) | 11.285982 / 8.074308 (3.211673) | 9.441848 / 10.191392 (-0.749544) | 0.140773 / 0.680424 (-0.539650) | 0.013783 / 0.534201 (-0.520418) | 0.292304 / 0.579283 (-0.286979) | 0.275011 / 0.434364 (-0.159353) | 0.339094 / 0.540337 (-0.201244) | 0.447593 / 1.386936 (-0.939343) |\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) | 0.005258 / 0.011353 (-0.006095) | 0.003539 / 0.011008 (-0.007469) | 0.049920 / 0.038508 (0.011412) | 0.029789 / 0.023109 (0.006680) | 0.277187 / 0.275898 (0.001288) | 0.296817 / 0.323480 (-0.026663) | 0.004133 / 0.007986 (-0.003852) | 0.002679 / 0.004328 (-0.001649) | 0.048999 / 0.004250 (0.044749) | 0.044087 / 0.037052 (0.007034) | 0.290359 / 0.258489 (0.031870) | 0.319572 / 0.293841 (0.025731) | 0.030248 / 0.128546 (-0.098298) | 0.010453 / 0.075646 (-0.065194) | 0.058734 / 0.419271 (-0.360537) | 0.051216 / 0.043533 (0.007683) | 0.278667 / 0.255139 (0.023528) | 0.298792 / 0.283200 (0.015592) | 0.019131 / 0.141683 (-0.122552) | 1.131814 / 1.452155 (-0.320340) | 1.167208 / 1.492716 (-0.325508) |\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) | 0.088316 / 0.018006 (0.070309) | 0.297143 / 0.000490 (0.296653) | 0.000207 / 0.000200 (0.000007) | 0.000048 / 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) | 0.022457 / 0.037411 (-0.014954) | 0.075251 / 0.014526 (0.060726) | 0.086747 / 0.176557 (-0.089809) | 0.124975 / 0.737135 (-0.612161) | 0.087320 / 0.296338 (-0.209019) |\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) | 0.292339 / 0.215209 (0.077130) | 2.860196 / 2.077655 (0.782541) | 1.599058 / 1.504120 (0.094938) | 1.476104 / 1.541195 (-0.065091) | 1.509109 / 1.468490 (0.040619) | 0.564056 / 4.584777 (-4.020721) | 2.388870 / 3.745712 (-1.356842) | 2.582356 / 5.269862 (-2.687506) | 1.726033 / 4.565676 (-2.839644) | 0.061788 / 0.424275 (-0.362487) | 0.005021 / 0.007607 (-0.002586) | 0.345644 / 0.226044 (0.119600) | 3.384000 / 2.268929 (1.115071) | 1.946591 / 55.444624 (-53.498033) | 1.693485 / 6.876477 (-5.182992) | 1.790300 / 2.142072 (-0.351773) | 0.654637 / 4.805227 (-4.150590) | 0.116271 / 6.500664 (-6.384393) | 0.040710 / 0.075469 (-0.034759) |\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) | 1.007367 / 1.841788 (-0.834421) | 11.868065 / 8.074308 (3.793757) | 10.146212 / 10.191392 (-0.045180) | 0.128902 / 0.680424 (-0.551522) | 0.015259 / 0.534201 (-0.518942) | 0.288087 / 0.579283 (-0.291196) | 0.281516 / 0.434364 (-0.152848) | 0.325755 / 0.540337 (-0.214583) | 0.424814 / 1.386936 (-0.962122) |\n\n</details>\n</details>\n\n![](https://cml.dev/watermark.png#8247202a7ed1c3164c88f8f183513c5f003aa2af \"CML watermark\")\n"
] |
6,714
|
Expand no-code dataset info with datasets-server info
|
E.g., to have info about a dataset's number of examples for more informative TQDM bars.
|
[] |
{
"url": "https://api.github.com/repos/huggingface/datasets/pulls/6714",
"html_url": "https://github.com/huggingface/datasets/pull/6714",
"diff_url": "https://github.com/huggingface/datasets/pull/6714.diff",
"patch_url": "https://github.com/huggingface/datasets/pull/6714.patch",
"merged_at": "2024-03-04T20:22:15"
}
| true
|
[
"The docs for this PR live [here](https://moon-ci-docs.huggingface.co/docs/datasets/pr_6714). All of your documentation changes will be reflected on that endpoint. The docs are available until 30 days after the last update.",
"<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) | 0.005237 / 0.011353 (-0.006116) | 0.003614 / 0.011008 (-0.007394) | 0.063349 / 0.038508 (0.024841) | 0.027297 / 0.023109 (0.004187) | 0.236203 / 0.275898 (-0.039695) | 0.260029 / 0.323480 (-0.063451) | 0.003096 / 0.007986 (-0.004889) | 0.003342 / 0.004328 (-0.000987) | 0.048703 / 0.004250 (0.044453) | 0.043121 / 0.037052 (0.006069) | 0.257491 / 0.258489 (-0.000998) | 0.282861 / 0.293841 (-0.010980) | 0.027701 / 0.128546 (-0.100845) | 0.010634 / 0.075646 (-0.065012) | 0.207369 / 0.419271 (-0.211903) | 0.035799 / 0.043533 (-0.007734) | 0.240445 / 0.255139 (-0.014694) | 0.261977 / 0.283200 (-0.021223) | 0.018175 / 0.141683 (-0.123508) | 1.143964 / 1.452155 (-0.308191) | 1.230057 / 1.492716 (-0.262659) |\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) | 0.096656 / 0.018006 (0.078650) | 0.303434 / 0.000490 (0.302944) | 0.000225 / 0.000200 (0.000025) | 0.000051 / 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) | 0.018454 / 0.037411 (-0.018957) | 0.061792 / 0.014526 (0.047266) | 0.073384 / 0.176557 (-0.103172) | 0.120148 / 0.737135 (-0.616988) | 0.074221 / 0.296338 (-0.222118) |\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) | 0.290291 / 0.215209 (0.075082) | 2.822908 / 2.077655 (0.745254) | 1.483139 / 1.504120 (-0.020981) | 1.349619 / 1.541195 (-0.191576) | 1.356588 / 1.468490 (-0.111902) | 0.571723 / 4.584777 (-4.013054) | 2.402696 / 3.745712 (-1.343016) | 2.832215 / 5.269862 (-2.437647) | 1.794962 / 4.565676 (-2.770714) | 0.062707 / 0.424275 (-0.361568) | 0.004997 / 0.007607 (-0.002610) | 0.343093 / 0.226044 (0.117049) | 3.383028 / 2.268929 (1.114100) | 1.818624 / 55.444624 (-53.626000) | 1.549859 / 6.876477 (-5.326618) | 1.667838 / 2.142072 (-0.474235) | 0.648574 / 4.805227 (-4.156653) | 0.119181 / 6.500664 (-6.381484) | 0.042074 / 0.075469 (-0.033395) |\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) | 0.982039 / 1.841788 (-0.859748) | 11.411759 / 8.074308 (3.337451) | 9.783405 / 10.191392 (-0.407987) | 0.129577 / 0.680424 (-0.550847) | 0.014091 / 0.534201 (-0.520110) | 0.297925 / 0.579283 (-0.281358) | 0.263884 / 0.434364 (-0.170480) | 0.346032 / 0.540337 (-0.194305) | 0.444806 / 1.386936 (-0.942130) |\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) | 0.005527 / 0.011353 (-0.005826) | 0.003677 / 0.011008 (-0.007332) | 0.050245 / 0.038508 (0.011737) | 0.030070 / 0.023109 (0.006961) | 0.272640 / 0.275898 (-0.003258) | 0.296555 / 0.323480 (-0.026925) | 0.004247 / 0.007986 (-0.003738) | 0.003833 / 0.004328 (-0.000495) | 0.049341 / 0.004250 (0.045091) | 0.046604 / 0.037052 (0.009552) | 0.282765 / 0.258489 (0.024276) | 0.314924 / 0.293841 (0.021084) | 0.029749 / 0.128546 (-0.098797) | 0.010524 / 0.075646 (-0.065122) | 0.057859 / 0.419271 (-0.361412) | 0.053172 / 0.043533 (0.009640) | 0.274906 / 0.255139 (0.019767) | 0.290566 / 0.283200 (0.007366) | 0.019299 / 0.141683 (-0.122384) | 1.164092 / 1.452155 (-0.288062) | 1.205074 / 1.492716 (-0.287642) |\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) | 0.093943 / 0.018006 (0.075936) | 0.298746 / 0.000490 (0.298256) | 0.000232 / 0.000200 (0.000032) | 0.000054 / 0.000054 (-0.000000) |\n\n### Benchmark: benchmark_indices_mapping.json\n\n| metric | select | shard | shuffle | sort | train_test_split |\n|--------|---|---|---|---|---|\n| new / old (diff) | 0.022098 / 0.037411 (-0.015313) | 0.075523 / 0.014526 (0.060997) | 0.086784 / 0.176557 (-0.089773) | 0.124610 / 0.737135 (-0.612525) | 0.087743 / 0.296338 (-0.208595) |\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) | 0.298555 / 0.215209 (0.083346) | 2.951493 / 2.077655 (0.873838) | 1.611448 / 1.504120 (0.107328) | 1.481503 / 1.541195 (-0.059692) | 1.497937 / 1.468490 (0.029447) | 0.580402 / 4.584777 (-4.004375) | 2.433308 / 3.745712 (-1.312404) | 2.712717 / 5.269862 (-2.557145) | 1.766286 / 4.565676 (-2.799391) | 0.063973 / 0.424275 (-0.360303) | 0.005006 / 0.007607 (-0.002601) | 0.354541 / 0.226044 (0.128497) | 3.486448 / 2.268929 (1.217519) | 1.972779 / 55.444624 (-53.471846) | 1.709018 / 6.876477 (-5.167458) | 1.864242 / 2.142072 (-0.277831) | 0.678213 / 4.805227 (-4.127014) | 0.119525 / 6.500664 (-6.381140) | 0.041387 / 0.075469 (-0.034082) |\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) | 1.021337 / 1.841788 (-0.820451) | 12.049563 / 8.074308 (3.975255) | 10.424701 / 10.191392 (0.233309) | 0.131444 / 0.680424 (-0.548980) | 0.015644 / 0.534201 (-0.518557) | 0.293712 / 0.579283 (-0.285571) | 0.279160 / 0.434364 (-0.155204) | 0.327991 / 0.540337 (-0.212346) | 0.435455 / 1.386936 (-0.951481) |\n\n</details>\n</details>\n\n![](https://cml.dev/watermark.png#1fe9483acc1ccaf19f3c199b99391921a8526215 \"CML watermark\")\n"
] |
6,713
|
Bump huggingface-hub lower version to 0.21.2
|
This should fix the version compatibility issue when using `huggingface_hub` < 0.21.2 and latest fsspec (>=2023.12.0).
See my comment: https://github.com/huggingface/datasets/pull/6687#issuecomment-1976493336
>> EDIT: the fix has been released in `huggingface_hub` 0.21.2 - I removed my commits that were using `huggingface_hub@main`
>
>Please note that people using `huggingface_hub` < 0.21.2 and latest `fsspec` will have issues when using `datasets`:
>- https://github.com/huggingface/lighteval/actions/runs/8139147047/job/22241658122?pr=86
>- https://github.com/huggingface/lighteval/pull/84
CC: @clefourrier
|
[] |
{
"url": "https://api.github.com/repos/huggingface/datasets/pulls/6713",
"html_url": "https://github.com/huggingface/datasets/pull/6713",
"diff_url": "https://github.com/huggingface/datasets/pull/6713.diff",
"patch_url": "https://github.com/huggingface/datasets/pull/6713.patch",
"merged_at": "2024-03-04T18:06:05"
}
| true
|
[
"The docs for this PR live [here](https://moon-ci-docs.huggingface.co/docs/datasets/pr_6713). All of your documentation changes will be reflected on that endpoint. The docs are available until 30 days after the last update.",
"@lhoestq if you agree, I could make a patch release tomorrow morning.",
"sure :)",
"<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) | 0.005086 / 0.011353 (-0.006267) | 0.003695 / 0.011008 (-0.007313) | 0.063430 / 0.038508 (0.024922) | 0.026798 / 0.023109 (0.003689) | 0.253761 / 0.275898 (-0.022138) | 0.301301 / 0.323480 (-0.022179) | 0.004160 / 0.007986 (-0.003825) | 0.002783 / 0.004328 (-0.001545) | 0.050698 / 0.004250 (0.046448) | 0.040899 / 0.037052 (0.003846) | 0.269024 / 0.258489 (0.010535) | 0.323467 / 0.293841 (0.029626) | 0.027756 / 0.128546 (-0.100791) | 0.010684 / 0.075646 (-0.064963) | 0.207128 / 0.419271 (-0.212144) | 0.035874 / 0.043533 (-0.007659) | 0.251620 / 0.255139 (-0.003519) | 0.268668 / 0.283200 (-0.014532) | 0.017387 / 0.141683 (-0.124296) | 1.139230 / 1.452155 (-0.312925) | 1.183613 / 1.492716 (-0.309103) |\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) | 0.096337 / 0.018006 (0.078331) | 0.305014 / 0.000490 (0.304524) | 0.000219 / 0.000200 (0.000019) | 0.000050 / 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) | 0.018086 / 0.037411 (-0.019325) | 0.061626 / 0.014526 (0.047100) | 0.072598 / 0.176557 (-0.103959) | 0.119944 / 0.737135 (-0.617192) | 0.074549 / 0.296338 (-0.221789) |\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) | 0.282661 / 0.215209 (0.067452) | 2.804473 / 2.077655 (0.726818) | 1.444602 / 1.504120 (-0.059517) | 1.313977 / 1.541195 (-0.227217) | 1.319426 / 1.468490 (-0.149064) | 0.570176 / 4.584777 (-4.014601) | 2.397895 / 3.745712 (-1.347818) | 2.760208 / 5.269862 (-2.509654) | 1.732457 / 4.565676 (-2.833220) | 0.062743 / 0.424275 (-0.361533) | 0.004950 / 0.007607 (-0.002657) | 0.338500 / 0.226044 (0.112456) | 3.287249 / 2.268929 (1.018320) | 1.777495 / 55.444624 (-53.667130) | 1.521255 / 6.876477 (-5.355222) | 1.517317 / 2.142072 (-0.624756) | 0.642202 / 4.805227 (-4.163025) | 0.116501 / 6.500664 (-6.384163) | 0.042418 / 0.075469 (-0.033052) |\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) | 0.968966 / 1.841788 (-0.872822) | 11.490531 / 8.074308 (3.416223) | 9.507803 / 10.191392 (-0.683589) | 0.141570 / 0.680424 (-0.538854) | 0.014000 / 0.534201 (-0.520201) | 0.284237 / 0.579283 (-0.295046) | 0.269341 / 0.434364 (-0.165022) | 0.321654 / 0.540337 (-0.218683) | 0.446914 / 1.386936 (-0.940022) |\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) | 0.005280 / 0.011353 (-0.006072) | 0.003794 / 0.011008 (-0.007214) | 0.050328 / 0.038508 (0.011820) | 0.029756 / 0.023109 (0.006647) | 0.273403 / 0.275898 (-0.002495) | 0.297346 / 0.323480 (-0.026133) | 0.004310 / 0.007986 (-0.003676) | 0.002858 / 0.004328 (-0.001470) | 0.048833 / 0.004250 (0.044583) | 0.045696 / 0.037052 (0.008644) | 0.291034 / 0.258489 (0.032545) | 0.318899 / 0.293841 (0.025058) | 0.029809 / 0.128546 (-0.098737) | 0.010710 / 0.075646 (-0.064936) | 0.058183 / 0.419271 (-0.361089) | 0.051761 / 0.043533 (0.008228) | 0.275022 / 0.255139 (0.019883) | 0.291614 / 0.283200 (0.008414) | 0.017975 / 0.141683 (-0.123708) | 1.148489 / 1.452155 (-0.303666) | 1.218111 / 1.492716 (-0.274605) |\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) | 0.091806 / 0.018006 (0.073799) | 0.299413 / 0.000490 (0.298923) | 0.000219 / 0.000200 (0.000019) | 0.000044 / 0.000054 (-0.000010) |\n\n### Benchmark: benchmark_indices_mapping.json\n\n| metric | select | shard | shuffle | sort | train_test_split |\n|--------|---|---|---|---|---|\n| new / old (diff) | 0.021506 / 0.037411 (-0.015905) | 0.075537 / 0.014526 (0.061011) | 0.087020 / 0.176557 (-0.089536) | 0.125270 / 0.737135 (-0.611865) | 0.088038 / 0.296338 (-0.208300) |\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) | 0.300401 / 0.215209 (0.085192) | 2.932571 / 2.077655 (0.854916) | 1.609502 / 1.504120 (0.105383) | 1.480078 / 1.541195 (-0.061117) | 1.514902 / 1.468490 (0.046412) | 0.575591 / 4.584777 (-4.009186) | 2.461873 / 3.745712 (-1.283839) | 2.728099 / 5.269862 (-2.541762) | 1.760054 / 4.565676 (-2.805622) | 0.064371 / 0.424275 (-0.359904) | 0.004990 / 0.007607 (-0.002617) | 0.350134 / 0.226044 (0.124090) | 3.453249 / 2.268929 (1.184321) | 1.979760 / 55.444624 (-53.464865) | 1.741128 / 6.876477 (-5.135348) | 1.825734 / 2.142072 (-0.316339) | 0.654902 / 4.805227 (-4.150325) | 0.116989 / 6.500664 (-6.383676) | 0.040800 / 0.075469 (-0.034669) |\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) | 1.033352 / 1.841788 (-0.808436) | 12.196711 / 8.074308 (4.122403) | 10.315114 / 10.191392 (0.123722) | 0.132541 / 0.680424 (-0.547882) | 0.016455 / 0.534201 (-0.517746) | 0.289025 / 0.579283 (-0.290258) | 0.281464 / 0.434364 (-0.152900) | 0.325302 / 0.540337 (-0.215036) | 0.428469 / 1.386936 (-0.958467) |\n\n</details>\n</details>\n\n![](https://cml.dev/watermark.png#7093b4b1a69f413e452119c87669af9e8ceaf749 \"CML watermark\")\n"
] |
6,712
|
fix CastError pickling
|
reported in https://discuss.huggingface.co/t/datasetdict-save-to-disk-with-num-proc-1-seems-to-hang-with-error/75595
|
[] |
{
"url": "https://api.github.com/repos/huggingface/datasets/pulls/6712",
"html_url": "https://github.com/huggingface/datasets/pull/6712",
"diff_url": "https://github.com/huggingface/datasets/pull/6712.diff",
"patch_url": "https://github.com/huggingface/datasets/pull/6712.patch",
"merged_at": "2024-03-04T20:17:17"
}
| true
|
[
"The docs for this PR live [here](https://moon-ci-docs.huggingface.co/docs/datasets/pr_6712). All of your documentation changes will be reflected on that endpoint. The docs are available until 30 days after the last update.",
"<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) | 0.005151 / 0.011353 (-0.006202) | 0.003813 / 0.011008 (-0.007196) | 0.062957 / 0.038508 (0.024449) | 0.028282 / 0.023109 (0.005173) | 0.246036 / 0.275898 (-0.029862) | 0.290024 / 0.323480 (-0.033456) | 0.004009 / 0.007986 (-0.003977) | 0.002749 / 0.004328 (-0.001580) | 0.049351 / 0.004250 (0.045101) | 0.041143 / 0.037052 (0.004090) | 0.264782 / 0.258489 (0.006293) | 0.290711 / 0.293841 (-0.003130) | 0.027248 / 0.128546 (-0.101298) | 0.010691 / 0.075646 (-0.064955) | 0.205926 / 0.419271 (-0.213345) | 0.035652 / 0.043533 (-0.007880) | 0.246357 / 0.255139 (-0.008782) | 0.267851 / 0.283200 (-0.015348) | 0.018498 / 0.141683 (-0.123185) | 1.135996 / 1.452155 (-0.316159) | 1.181841 / 1.492716 (-0.310875) |\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) | 0.094054 / 0.018006 (0.076048) | 0.305470 / 0.000490 (0.304980) | 0.000225 / 0.000200 (0.000025) | 0.000043 / 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) | 0.018842 / 0.037411 (-0.018569) | 0.061532 / 0.014526 (0.047006) | 0.073483 / 0.176557 (-0.103073) | 0.119426 / 0.737135 (-0.617709) | 0.075385 / 0.296338 (-0.220954) |\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) | 0.285544 / 0.215209 (0.070335) | 2.774256 / 2.077655 (0.696601) | 1.475719 / 1.504120 (-0.028401) | 1.353841 / 1.541195 (-0.187353) | 1.381891 / 1.468490 (-0.086599) | 0.570619 / 4.584777 (-4.014158) | 2.380300 / 3.745712 (-1.365412) | 2.788767 / 5.269862 (-2.481095) | 1.741790 / 4.565676 (-2.823886) | 0.061810 / 0.424275 (-0.362465) | 0.005004 / 0.007607 (-0.002603) | 0.334963 / 0.226044 (0.108918) | 3.286388 / 2.268929 (1.017459) | 1.831669 / 55.444624 (-53.612955) | 1.523372 / 6.876477 (-5.353105) | 1.581551 / 2.142072 (-0.560521) | 0.639642 / 4.805227 (-4.165585) | 0.117356 / 6.500664 (-6.383308) | 0.043277 / 0.075469 (-0.032192) |\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) | 0.973005 / 1.841788 (-0.868782) | 11.590148 / 8.074308 (3.515839) | 9.521262 / 10.191392 (-0.670130) | 0.143243 / 0.680424 (-0.537181) | 0.013529 / 0.534201 (-0.520672) | 0.285724 / 0.579283 (-0.293559) | 0.265642 / 0.434364 (-0.168721) | 0.366098 / 0.540337 (-0.174239) | 0.444410 / 1.386936 (-0.942526) |\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) | 0.005347 / 0.011353 (-0.006006) | 0.003797 / 0.011008 (-0.007212) | 0.050441 / 0.038508 (0.011933) | 0.032812 / 0.023109 (0.009703) | 0.281278 / 0.275898 (0.005379) | 0.304524 / 0.323480 (-0.018956) | 0.005039 / 0.007986 (-0.002946) | 0.002735 / 0.004328 (-0.001594) | 0.049184 / 0.004250 (0.044933) | 0.046751 / 0.037052 (0.009698) | 0.292093 / 0.258489 (0.033604) | 0.322087 / 0.293841 (0.028246) | 0.029775 / 0.128546 (-0.098771) | 0.010540 / 0.075646 (-0.065106) | 0.057927 / 0.419271 (-0.361345) | 0.054240 / 0.043533 (0.010707) | 0.281537 / 0.255139 (0.026398) | 0.298386 / 0.283200 (0.015186) | 0.019773 / 0.141683 (-0.121910) | 1.157161 / 1.452155 (-0.294994) | 1.210395 / 1.492716 (-0.282321) |\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) | 0.095098 / 0.018006 (0.077091) | 0.306952 / 0.000490 (0.306462) | 0.000211 / 0.000200 (0.000011) | 0.000050 / 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) | 0.022602 / 0.037411 (-0.014809) | 0.075242 / 0.014526 (0.060716) | 0.087134 / 0.176557 (-0.089422) | 0.127923 / 0.737135 (-0.609212) | 0.088645 / 0.296338 (-0.207693) |\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) | 0.304187 / 0.215209 (0.088978) | 2.977120 / 2.077655 (0.899465) | 1.663592 / 1.504120 (0.159473) | 1.527601 / 1.541195 (-0.013594) | 1.540121 / 1.468490 (0.071631) | 0.562492 / 4.584777 (-4.022285) | 2.473836 / 3.745712 (-1.271876) | 2.656782 / 5.269862 (-2.613080) | 1.754212 / 4.565676 (-2.811464) | 0.062330 / 0.424275 (-0.361945) | 0.005149 / 0.007607 (-0.002459) | 0.354905 / 0.226044 (0.128860) | 3.503587 / 2.268929 (1.234659) | 2.015682 / 55.444624 (-53.428943) | 1.744421 / 6.876477 (-5.132056) | 1.923120 / 2.142072 (-0.218952) | 0.652209 / 4.805227 (-4.153018) | 0.119406 / 6.500664 (-6.381258) | 0.042840 / 0.075469 (-0.032630) |\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) | 1.009164 / 1.841788 (-0.832624) | 12.379654 / 8.074308 (4.305346) | 10.408696 / 10.191392 (0.217304) | 0.141674 / 0.680424 (-0.538750) | 0.016815 / 0.534201 (-0.517386) | 0.292453 / 0.579283 (-0.286830) | 0.277577 / 0.434364 (-0.156787) | 0.325024 / 0.540337 (-0.215313) | 0.433181 / 1.386936 (-0.953755) |\n\n</details>\n</details>\n\n![](https://cml.dev/watermark.png#b7a16a08c940e65397305aec5f1b484d91cee75a \"CML watermark\")\n"
] |
6,711
|
3x Faster Text Preprocessing
|
I was preparing some datasets for AI training and noticed that `datasets` by HuggingFace uses the conventional `open` mechanism to read the file and split it into chunks. I thought it can be significantly accelerated, and [started with a benchmark](https://gist.github.com/ashvardanian/55c2052e9f78b05b8d614aa90cb12347):
```sh
$ pip install --upgrade --force-reinstall datasets
$ python benchmark_huggingface_datasets.py xlsum.csv
Generating train split: 1004598 examples [00:47, 21116.16 examples/s]
Time taken to load the dataset: 48.66838526725769 seconds
Time taken to chunk the dataset into parts of size 10000: 0.11466407775878906 seconds
Total time taken: 48.78304934501648 seconds
```
For benchmarks I've used a [large CSV file with mixed UTF-8 content](https://github.com/ashvardanian/StringZilla/blob/main/CONTRIBUTING.md#benchmarking-datasets), most common in modern large-scale pre-training pipelines. I've later patched the `datasets` library to use `stringzilla`, which resulted in significantly lower memory consumption and in 2.9x throughput improvement on the AWS `r7iz` instances. That's using slow SSDs mounted over the network. Performance on local SSDs on something like a DGX-H100 should be even higher:
```sh
$ pip install -e .
$ python benchmark_huggingface_datasets.py xlsum.csv
Generating train split: 1004598 examples [00:15, 64529.90 examples/s]
Time taken to load the dataset: 16.45028805732727 seconds
Time taken to chunk the dataset into parts of size 10000: 0.1291060447692871 seconds
Total time taken: 16.579394102096558 seconds
```
I've already [pushed the patches to my fork](https://github.com/ashvardanian/datasets/tree/faster-text-parsers), and would love to contribute them to the upstream repository.
---
All the tests pass, but they leave a couple of important questions open. The default Python `open(..., newline=None)` uses universal newlines, where `\n`, `\r`, and `\r\n` are all converted to `\n` on the fly. I am not sure if its a good idea for a general purpose dataset preparation pipeline?
I can simulate the same behavior (which I don't yet do) for `"line"` splitter. Adjusting it for `"paragraph"`-splitter would be harder. Should we stick exactly to the old Pythonic behavior or stay closer to how C and other programming languages do that?
|
[] |
{
"url": "https://api.github.com/repos/huggingface/datasets/pulls/6711",
"html_url": "https://github.com/huggingface/datasets/pull/6711",
"diff_url": "https://github.com/huggingface/datasets/pull/6711.diff",
"patch_url": "https://github.com/huggingface/datasets/pull/6711.patch",
"merged_at": null
}
| true
|
[
"Unfortunately, that won't improve the performance. StringZilla repository has extensive benchmarks comparing against different built-in functionality of several programming languages. Using `re.finditer` for tokenization is practically the slowest anti-pattern I've encountered in any language. The gap between that and a SIMD-accelerated kernel can be as big as 10 MB/s vs 10 GB/s.\n\nI understand the need to keep the dependencies minimal. It helps the package remain small and portable. At this point, StringZilla provides 105 binaries for different OS and hardware versions (more portable than NumPy) and the [binary size generally ranges from 50 KB to 250 KB](https://pypi.org/project/stringzilla/), smaller than a single JPEG. \n",
"The `text` builder is not very popular, so I'm also not a fan of introducing a dependency for it.\r\n\r\nMoreover, I couldn't find any projects of this size/usage depending on StringZilla (with GitHub search), so we should at least wait for its greater adoption to merge this PR.\r\n",
"> Moreover, I couldn't find any projects of this size/usage depending on StringZilla (with GitHub search), so we should at least wait for its greater adoption to merge this PR.\r\n\r\nMeanwhile I understand that you want to wait for a greater adoption - if things change in the future you would be stuck with an unsupported dependency (although I think, that really applies to everything) - the performance improvement is really significant!\r\nI wonder if it's worth, perhaps, to provide an additional 'datasets.extras' library by huggingface which support these 3rd party improvements.\r\nIt would reduce the risk on the core components and, at the same time, it would definitely help on the performance side!"
] |
6,710
|
Persist IterableDataset epoch in workers
|
Use shared memory for the IterableDataset epoch.
This way calling `ds.set_epoch()` in the main process will update the epoch in the DataLoader workers as well.
This is useful especially because the epoch is used to compute the `effective_seed` used for shuffling.
I used torch's shared memory in case users want to send dataset copies without shared memory using pickle. I also find it easier to use than using `multiprocessing.shared_memory` than requires unlinking only in the main process, or `mp.Value` that is not picklable.
close https://github.com/huggingface/datasets/issues/6673
cc @rwightman
|
[] |
{
"url": "https://api.github.com/repos/huggingface/datasets/pulls/6710",
"html_url": "https://github.com/huggingface/datasets/pull/6710",
"diff_url": "https://github.com/huggingface/datasets/pull/6710.diff",
"patch_url": "https://github.com/huggingface/datasets/pull/6710.patch",
"merged_at": "2024-07-01T17:45:30"
}
| true
|
[
"The docs for this PR live [here](https://moon-ci-docs.huggingface.co/docs/datasets/pr_6710). All of your documentation changes will be reflected on that endpoint. The docs are available until 30 days after the last update.",
"<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) | 0.005283 / 0.011353 (-0.006070) | 0.003866 / 0.011008 (-0.007142) | 0.063124 / 0.038508 (0.024616) | 0.030240 / 0.023109 (0.007131) | 0.232855 / 0.275898 (-0.043043) | 0.257538 / 0.323480 (-0.065942) | 0.004165 / 0.007986 (-0.003820) | 0.002826 / 0.004328 (-0.001502) | 0.049735 / 0.004250 (0.045485) | 0.045297 / 0.037052 (0.008244) | 0.251831 / 0.258489 (-0.006658) | 0.277812 / 0.293841 (-0.016029) | 0.030004 / 0.128546 (-0.098542) | 0.012319 / 0.075646 (-0.063328) | 0.206881 / 0.419271 (-0.212391) | 0.036561 / 0.043533 (-0.006972) | 0.234364 / 0.255139 (-0.020775) | 0.258316 / 0.283200 (-0.024884) | 0.017815 / 0.141683 (-0.123867) | 1.114111 / 1.452155 (-0.338043) | 1.165428 / 1.492716 (-0.327288) |\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) | 0.099302 / 0.018006 (0.081296) | 0.309195 / 0.000490 (0.308705) | 0.000261 / 0.000200 (0.000061) | 0.000044 / 0.000054 (-0.000010) |\n\n### Benchmark: benchmark_indices_mapping.json\n\n| metric | select | shard | shuffle | sort | train_test_split |\n|--------|---|---|---|---|---|\n| new / old (diff) | 0.018765 / 0.037411 (-0.018646) | 0.063123 / 0.014526 (0.048597) | 0.075437 / 0.176557 (-0.101119) | 0.122570 / 0.737135 (-0.614566) | 0.076637 / 0.296338 (-0.219702) |\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) | 0.289965 / 0.215209 (0.074756) | 2.839053 / 2.077655 (0.761398) | 1.503463 / 1.504120 (-0.000657) | 1.390833 / 1.541195 (-0.150361) | 1.401918 / 1.468490 (-0.066572) | 0.711000 / 4.584777 (-3.873777) | 2.325513 / 3.745712 (-1.420199) | 2.831630 / 5.269862 (-2.438231) | 1.908370 / 4.565676 (-2.657307) | 0.077867 / 0.424275 (-0.346408) | 0.005509 / 0.007607 (-0.002098) | 0.336494 / 0.226044 (0.110450) | 3.358587 / 2.268929 (1.089658) | 1.901067 / 55.444624 (-53.543558) | 1.590130 / 6.876477 (-5.286347) | 1.753850 / 2.142072 (-0.388223) | 0.792458 / 4.805227 (-4.012769) | 0.135584 / 6.500664 (-6.365080) | 0.042028 / 0.075469 (-0.033441) |\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) | 0.966162 / 1.841788 (-0.875625) | 11.705310 / 8.074308 (3.631002) | 9.158842 / 10.191392 (-1.032550) | 0.128793 / 0.680424 (-0.551631) | 0.014422 / 0.534201 (-0.519779) | 0.299009 / 0.579283 (-0.280274) | 0.262873 / 0.434364 (-0.171491) | 0.340836 / 0.540337 (-0.199501) | 0.464440 / 1.386936 (-0.922496) |\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) | 0.005951 / 0.011353 (-0.005402) | 0.003984 / 0.011008 (-0.007024) | 0.051432 / 0.038508 (0.012924) | 0.033223 / 0.023109 (0.010113) | 0.263972 / 0.275898 (-0.011926) | 0.289060 / 0.323480 (-0.034420) | 0.004446 / 0.007986 (-0.003540) | 0.002891 / 0.004328 (-0.001438) | 0.049347 / 0.004250 (0.045096) | 0.041191 / 0.037052 (0.004138) | 0.278334 / 0.258489 (0.019844) | 0.314065 / 0.293841 (0.020224) | 0.032020 / 0.128546 (-0.096526) | 0.012472 / 0.075646 (-0.063174) | 0.061288 / 0.419271 (-0.357984) | 0.033489 / 0.043533 (-0.010044) | 0.266831 / 0.255139 (0.011692) | 0.283008 / 0.283200 (-0.000192) | 0.018491 / 0.141683 (-0.123192) | 1.133634 / 1.452155 (-0.318521) | 1.154627 / 1.492716 (-0.338089) |\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) | 0.101831 / 0.018006 (0.083825) | 0.317942 / 0.000490 (0.317452) | 0.000217 / 0.000200 (0.000018) | 0.000056 / 0.000054 (0.000002) |\n\n### Benchmark: benchmark_indices_mapping.json\n\n| metric | select | shard | shuffle | sort | train_test_split |\n|--------|---|---|---|---|---|\n| new / old (diff) | 0.022608 / 0.037411 (-0.014803) | 0.076776 / 0.014526 (0.062250) | 0.088686 / 0.176557 (-0.087870) | 0.129092 / 0.737135 (-0.608044) | 0.090780 / 0.296338 (-0.205558) |\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) | 0.286762 / 0.215209 (0.071553) | 2.824307 / 2.077655 (0.746652) | 1.547215 / 1.504120 (0.043095) | 1.424522 / 1.541195 (-0.116673) | 1.446414 / 1.468490 (-0.022076) | 0.723683 / 4.584777 (-3.861094) | 0.974129 / 3.745712 (-2.771583) | 2.952552 / 5.269862 (-2.317309) | 1.903663 / 4.565676 (-2.662013) | 0.078786 / 0.424275 (-0.345489) | 0.005130 / 0.007607 (-0.002477) | 0.338925 / 0.226044 (0.112881) | 3.378557 / 2.268929 (1.109629) | 1.892951 / 55.444624 (-53.551674) | 1.599844 / 6.876477 (-5.276633) | 1.611963 / 2.142072 (-0.530109) | 0.793614 / 4.805227 (-4.011613) | 0.133795 / 6.500664 (-6.366869) | 0.040777 / 0.075469 (-0.034692) |\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) | 1.001391 / 1.841788 (-0.840397) | 12.166811 / 8.074308 (4.092503) | 10.588180 / 10.191392 (0.396788) | 0.141609 / 0.680424 (-0.538815) | 0.020941 / 0.534201 (-0.513260) | 0.340149 / 0.579283 (-0.239134) | 0.122988 / 0.434364 (-0.311376) | 0.339747 / 0.540337 (-0.200591) | 0.434338 / 1.386936 (-0.952598) |\n\n</details>\n</details>\n\n![](https://cml.dev/watermark.png#4ba47a35820069e7de9f374479d06b2a7935767e \"CML watermark\")\n"
] |
6,709
|
set dev version
| null |
[] |
{
"url": "https://api.github.com/repos/huggingface/datasets/pulls/6709",
"html_url": "https://github.com/huggingface/datasets/pull/6709",
"diff_url": "https://github.com/huggingface/datasets/pull/6709.diff",
"patch_url": "https://github.com/huggingface/datasets/pull/6709.patch",
"merged_at": "2024-03-01T21:01:23"
}
| true
|
[
"The docs for this PR live [here](https://moon-ci-docs.huggingface.co/docs/datasets/pr_6709). All of your documentation changes will be reflected on that endpoint. The docs are available until 30 days after the last update.",
"<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) | 0.005081 / 0.011353 (-0.006272) | 0.004182 / 0.011008 (-0.006826) | 0.063377 / 0.038508 (0.024869) | 0.027880 / 0.023109 (0.004770) | 0.247260 / 0.275898 (-0.028638) | 0.273580 / 0.323480 (-0.049900) | 0.002995 / 0.007986 (-0.004991) | 0.002804 / 0.004328 (-0.001524) | 0.049669 / 0.004250 (0.045418) | 0.042469 / 0.037052 (0.005417) | 0.268606 / 0.258489 (0.010117) | 0.292867 / 0.293841 (-0.000973) | 0.028077 / 0.128546 (-0.100469) | 0.011031 / 0.075646 (-0.064615) | 0.210225 / 0.419271 (-0.209047) | 0.035723 / 0.043533 (-0.007810) | 0.252131 / 0.255139 (-0.003008) | 0.272895 / 0.283200 (-0.010304) | 0.019809 / 0.141683 (-0.121874) | 1.138500 / 1.452155 (-0.313655) | 1.167752 / 1.492716 (-0.324964) |\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) | 0.094881 / 0.018006 (0.076875) | 0.300168 / 0.000490 (0.299678) | 0.000207 / 0.000200 (0.000007) | 0.000050 / 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) | 0.017917 / 0.037411 (-0.019494) | 0.061854 / 0.014526 (0.047328) | 0.074481 / 0.176557 (-0.102075) | 0.120075 / 0.737135 (-0.617061) | 0.074627 / 0.296338 (-0.221711) |\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) | 0.287888 / 0.215209 (0.072679) | 2.770165 / 2.077655 (0.692510) | 1.500071 / 1.504120 (-0.004049) | 1.374857 / 1.541195 (-0.166338) | 1.427291 / 1.468490 (-0.041200) | 0.558431 / 4.584777 (-4.026346) | 2.439352 / 3.745712 (-1.306361) | 2.787471 / 5.269862 (-2.482391) | 1.742636 / 4.565676 (-2.823041) | 0.061716 / 0.424275 (-0.362559) | 0.004961 / 0.007607 (-0.002646) | 0.345209 / 0.226044 (0.119164) | 3.360253 / 2.268929 (1.091325) | 1.847945 / 55.444624 (-53.596680) | 1.595733 / 6.876477 (-5.280744) | 1.642350 / 2.142072 (-0.499723) | 0.638639 / 4.805227 (-4.166588) | 0.116918 / 6.500664 (-6.383746) | 0.042132 / 0.075469 (-0.033338) |\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) | 0.980602 / 1.841788 (-0.861185) | 11.545402 / 8.074308 (3.471094) | 9.452471 / 10.191392 (-0.738921) | 0.129930 / 0.680424 (-0.550494) | 0.014143 / 0.534201 (-0.520058) | 0.290302 / 0.579283 (-0.288981) | 0.263785 / 0.434364 (-0.170579) | 0.339580 / 0.540337 (-0.200758) | 0.450355 / 1.386936 (-0.936581) |\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) | 0.005565 / 0.011353 (-0.005788) | 0.003764 / 0.011008 (-0.007244) | 0.050082 / 0.038508 (0.011574) | 0.030354 / 0.023109 (0.007245) | 0.250609 / 0.275898 (-0.025289) | 0.277200 / 0.323480 (-0.046280) | 0.004276 / 0.007986 (-0.003710) | 0.002805 / 0.004328 (-0.001523) | 0.048765 / 0.004250 (0.044514) | 0.045477 / 0.037052 (0.008425) | 0.267704 / 0.258489 (0.009215) | 0.303214 / 0.293841 (0.009373) | 0.029393 / 0.128546 (-0.099153) | 0.010623 / 0.075646 (-0.065023) | 0.058201 / 0.419271 (-0.361070) | 0.053131 / 0.043533 (0.009599) | 0.258682 / 0.255139 (0.003543) | 0.276069 / 0.283200 (-0.007131) | 0.018260 / 0.141683 (-0.123423) | 1.141542 / 1.452155 (-0.310613) | 1.185780 / 1.492716 (-0.306936) |\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) | 0.096857 / 0.018006 (0.078850) | 0.300656 / 0.000490 (0.300167) | 0.000450 / 0.000200 (0.000250) | 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) | 0.022416 / 0.037411 (-0.014995) | 0.074781 / 0.014526 (0.060255) | 0.087299 / 0.176557 (-0.089257) | 0.127616 / 0.737135 (-0.609519) | 0.088382 / 0.296338 (-0.207957) |\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) | 0.298639 / 0.215209 (0.083430) | 2.940002 / 2.077655 (0.862347) | 1.709707 / 1.504120 (0.205587) | 1.556502 / 1.541195 (0.015307) | 1.592841 / 1.468490 (0.124351) | 0.570237 / 4.584777 (-4.014539) | 2.467576 / 3.745712 (-1.278137) | 2.741021 / 5.269862 (-2.528840) | 1.776526 / 4.565676 (-2.789151) | 0.063999 / 0.424275 (-0.360276) | 0.005068 / 0.007607 (-0.002539) | 0.360727 / 0.226044 (0.134682) | 3.535404 / 2.268929 (1.266476) | 2.035345 / 55.444624 (-53.409279) | 1.755916 / 6.876477 (-5.120561) | 1.889281 / 2.142072 (-0.252791) | 0.649025 / 4.805227 (-4.156202) | 0.118210 / 6.500664 (-6.382454) | 0.040815 / 0.075469 (-0.034654) |\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) | 1.005650 / 1.841788 (-0.836138) | 12.228314 / 8.074308 (4.154006) | 10.147363 / 10.191392 (-0.044029) | 0.159258 / 0.680424 (-0.521166) | 0.015288 / 0.534201 (-0.518913) | 0.288144 / 0.579283 (-0.291139) | 0.281319 / 0.434364 (-0.153045) | 0.323380 / 0.540337 (-0.216958) | 0.426887 / 1.386936 (-0.960049) |\n\n</details>\n</details>\n\n![](https://cml.dev/watermark.png#8b04ccb486f3831b4b0d2474119823efa3815709 \"CML watermark\")\n"
] |
6,708
|
Release: 2.18.0
| null |
[] |
{
"url": "https://api.github.com/repos/huggingface/datasets/pulls/6708",
"html_url": "https://github.com/huggingface/datasets/pull/6708",
"diff_url": "https://github.com/huggingface/datasets/pull/6708.diff",
"patch_url": "https://github.com/huggingface/datasets/pull/6708.patch",
"merged_at": "2024-03-01T20:56:50"
}
| true
|
[
"The docs for this PR live [here](https://moon-ci-docs.huggingface.co/docs/datasets/pr_6708). All of your documentation changes will be reflected on that endpoint. The docs are available until 30 days after the last update.",
"<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) | 0.005442 / 0.011353 (-0.005910) | 0.003796 / 0.011008 (-0.007213) | 0.063663 / 0.038508 (0.025155) | 0.028901 / 0.023109 (0.005792) | 0.256742 / 0.275898 (-0.019156) | 0.279555 / 0.323480 (-0.043925) | 0.004128 / 0.007986 (-0.003858) | 0.002789 / 0.004328 (-0.001539) | 0.049463 / 0.004250 (0.045213) | 0.043461 / 0.037052 (0.006409) | 0.272975 / 0.258489 (0.014486) | 0.299057 / 0.293841 (0.005216) | 0.029030 / 0.128546 (-0.099516) | 0.010453 / 0.075646 (-0.065193) | 0.207611 / 0.419271 (-0.211660) | 0.037200 / 0.043533 (-0.006332) | 0.258327 / 0.255139 (0.003188) | 0.279746 / 0.283200 (-0.003454) | 0.018940 / 0.141683 (-0.122743) | 1.150379 / 1.452155 (-0.301776) | 1.217621 / 1.492716 (-0.275095) |\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) | 0.095115 / 0.018006 (0.077109) | 0.299393 / 0.000490 (0.298903) | 0.000223 / 0.000200 (0.000023) | 0.000044 / 0.000054 (-0.000010) |\n\n### Benchmark: benchmark_indices_mapping.json\n\n| metric | select | shard | shuffle | sort | train_test_split |\n|--------|---|---|---|---|---|\n| new / old (diff) | 0.018972 / 0.037411 (-0.018439) | 0.061669 / 0.014526 (0.047143) | 0.075605 / 0.176557 (-0.100951) | 0.125695 / 0.737135 (-0.611440) | 0.076654 / 0.296338 (-0.219684) |\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) | 0.286431 / 0.215209 (0.071222) | 2.763554 / 2.077655 (0.685899) | 1.489902 / 1.504120 (-0.014218) | 1.375082 / 1.541195 (-0.166113) | 1.418903 / 1.468490 (-0.049587) | 0.555646 / 4.584777 (-4.029131) | 2.410578 / 3.745712 (-1.335134) | 2.827453 / 5.269862 (-2.442408) | 1.764381 / 4.565676 (-2.801295) | 0.062937 / 0.424275 (-0.361339) | 0.004989 / 0.007607 (-0.002619) | 0.342115 / 0.226044 (0.116071) | 3.354660 / 2.268929 (1.085732) | 1.858418 / 55.444624 (-53.586206) | 1.586403 / 6.876477 (-5.290074) | 1.625762 / 2.142072 (-0.516311) | 0.643678 / 4.805227 (-4.161550) | 0.116764 / 6.500664 (-6.383900) | 0.042198 / 0.075469 (-0.033271) |\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) | 0.974953 / 1.841788 (-0.866835) | 11.748419 / 8.074308 (3.674111) | 9.753700 / 10.191392 (-0.437692) | 0.131330 / 0.680424 (-0.549094) | 0.018876 / 0.534201 (-0.515325) | 0.290078 / 0.579283 (-0.289205) | 0.264676 / 0.434364 (-0.169688) | 0.340285 / 0.540337 (-0.200052) | 0.445340 / 1.386936 (-0.941596) |\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) | 0.005513 / 0.011353 (-0.005840) | 0.003665 / 0.011008 (-0.007344) | 0.049368 / 0.038508 (0.010860) | 0.032045 / 0.023109 (0.008936) | 0.280955 / 0.275898 (0.005057) | 0.299804 / 0.323480 (-0.023675) | 0.004391 / 0.007986 (-0.003594) | 0.002896 / 0.004328 (-0.001432) | 0.048914 / 0.004250 (0.044663) | 0.045448 / 0.037052 (0.008396) | 0.298779 / 0.258489 (0.040289) | 0.322012 / 0.293841 (0.028171) | 0.029449 / 0.128546 (-0.099097) | 0.010410 / 0.075646 (-0.065236) | 0.057867 / 0.419271 (-0.361405) | 0.053944 / 0.043533 (0.010411) | 0.278139 / 0.255139 (0.023000) | 0.297453 / 0.283200 (0.014254) | 0.018746 / 0.141683 (-0.122937) | 1.137890 / 1.452155 (-0.314264) | 1.206109 / 1.492716 (-0.286607) |\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) | 0.091741 / 0.018006 (0.073735) | 0.300415 / 0.000490 (0.299925) | 0.000214 / 0.000200 (0.000014) | 0.000044 / 0.000054 (-0.000010) |\n\n### Benchmark: benchmark_indices_mapping.json\n\n| metric | select | shard | shuffle | sort | train_test_split |\n|--------|---|---|---|---|---|\n| new / old (diff) | 0.022097 / 0.037411 (-0.015314) | 0.076853 / 0.014526 (0.062327) | 0.088440 / 0.176557 (-0.088116) | 0.127176 / 0.737135 (-0.609959) | 0.088976 / 0.296338 (-0.207363) |\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) | 0.300754 / 0.215209 (0.085545) | 2.917713 / 2.077655 (0.840058) | 1.619338 / 1.504120 (0.115218) | 1.501543 / 1.541195 (-0.039652) | 1.506032 / 1.468490 (0.037542) | 0.579481 / 4.584777 (-4.005296) | 2.458917 / 3.745712 (-1.286795) | 2.754621 / 5.269862 (-2.515241) | 1.796440 / 4.565676 (-2.769237) | 0.067547 / 0.424275 (-0.356728) | 0.005001 / 0.007607 (-0.002606) | 0.351030 / 0.226044 (0.124985) | 3.466282 / 2.268929 (1.197353) | 1.954661 / 55.444624 (-53.489964) | 1.688737 / 6.876477 (-5.187740) | 1.836762 / 2.142072 (-0.305311) | 0.656441 / 4.805227 (-4.148786) | 0.118258 / 6.500664 (-6.382406) | 0.041608 / 0.075469 (-0.033861) |\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) | 0.999696 / 1.841788 (-0.842092) | 12.383471 / 8.074308 (4.309162) | 10.338488 / 10.191392 (0.147096) | 0.150214 / 0.680424 (-0.530210) | 0.014997 / 0.534201 (-0.519204) | 0.288949 / 0.579283 (-0.290334) | 0.272012 / 0.434364 (-0.162352) | 0.327253 / 0.540337 (-0.213084) | 0.427594 / 1.386936 (-0.959342) |\n\n</details>\n</details>\n\n![](https://cml.dev/watermark.png#ca8409a8bec4508255b9c3e808d0751eb1005260 \"CML watermark\")\n"
] |
6,707
|
Silence ruff deprecation messages
| null |
[] |
{
"url": "https://api.github.com/repos/huggingface/datasets/pulls/6707",
"html_url": "https://github.com/huggingface/datasets/pull/6707",
"diff_url": "https://github.com/huggingface/datasets/pull/6707.diff",
"patch_url": "https://github.com/huggingface/datasets/pull/6707.patch",
"merged_at": "2024-03-01T17:25:46"
}
| true
|
[
"The docs for this PR live [here](https://moon-ci-docs.huggingface.co/docs/datasets/pr_6707). All of your documentation changes will be reflected on that endpoint. The docs are available until 30 days after the last update.",
"<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) | 0.004728 / 0.011353 (-0.006624) | 0.002941 / 0.011008 (-0.008067) | 0.058270 / 0.038508 (0.019762) | 0.027418 / 0.023109 (0.004309) | 0.224993 / 0.275898 (-0.050905) | 0.243103 / 0.323480 (-0.080377) | 0.004668 / 0.007986 (-0.003318) | 0.002499 / 0.004328 (-0.001829) | 0.045020 / 0.004250 (0.040770) | 0.038006 / 0.037052 (0.000953) | 0.240807 / 0.258489 (-0.017682) | 0.264554 / 0.293841 (-0.029287) | 0.027018 / 0.128546 (-0.101529) | 0.009866 / 0.075646 (-0.065780) | 0.196578 / 0.419271 (-0.222694) | 0.034536 / 0.043533 (-0.008997) | 0.236535 / 0.255139 (-0.018604) | 0.248879 / 0.283200 (-0.034321) | 0.017140 / 0.141683 (-0.124543) | 1.046927 / 1.452155 (-0.405228) | 1.121209 / 1.492716 (-0.371507) |\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) | 0.088267 / 0.018006 (0.070261) | 0.279774 / 0.000490 (0.279284) | 0.000214 / 0.000200 (0.000014) | 0.000050 / 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) | 0.016521 / 0.037411 (-0.020890) | 0.056499 / 0.014526 (0.041974) | 0.067264 / 0.176557 (-0.109293) | 0.117270 / 0.737135 (-0.619865) | 0.069284 / 0.296338 (-0.227055) |\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) | 0.260679 / 0.215209 (0.045470) | 2.608971 / 2.077655 (0.531316) | 1.363139 / 1.504120 (-0.140981) | 1.262128 / 1.541195 (-0.279067) | 1.273619 / 1.468490 (-0.194871) | 0.523417 / 4.584777 (-4.061360) | 2.291145 / 3.745712 (-1.454567) | 2.540603 / 5.269862 (-2.729258) | 1.599090 / 4.565676 (-2.966586) | 0.058170 / 0.424275 (-0.366105) | 0.004556 / 0.007607 (-0.003051) | 0.308361 / 0.226044 (0.082316) | 3.069269 / 2.268929 (0.800340) | 1.698064 / 55.444624 (-53.746560) | 1.426631 / 6.876477 (-5.449846) | 1.463913 / 2.142072 (-0.678160) | 0.595234 / 4.805227 (-4.209993) | 0.107202 / 6.500664 (-6.393462) | 0.038183 / 0.075469 (-0.037286) |\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) | 0.905999 / 1.841788 (-0.935789) | 10.828492 / 8.074308 (2.754184) | 8.705635 / 10.191392 (-1.485757) | 0.121203 / 0.680424 (-0.559221) | 0.013789 / 0.534201 (-0.520412) | 0.268172 / 0.579283 (-0.311111) | 0.254277 / 0.434364 (-0.180086) | 0.310280 / 0.540337 (-0.230057) | 0.410490 / 1.386936 (-0.976446) |\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) | 0.004696 / 0.011353 (-0.006657) | 0.002982 / 0.011008 (-0.008026) | 0.045333 / 0.038508 (0.006825) | 0.027483 / 0.023109 (0.004374) | 0.253438 / 0.275898 (-0.022460) | 0.272657 / 0.323480 (-0.050823) | 0.004060 / 0.007986 (-0.003926) | 0.002574 / 0.004328 (-0.001754) | 0.045462 / 0.004250 (0.041212) | 0.041260 / 0.037052 (0.004208) | 0.267919 / 0.258489 (0.009430) | 0.290935 / 0.293841 (-0.002906) | 0.026674 / 0.128546 (-0.101873) | 0.009370 / 0.075646 (-0.066276) | 0.053543 / 0.419271 (-0.365729) | 0.047390 / 0.043533 (0.003857) | 0.255774 / 0.255139 (0.000635) | 0.273909 / 0.283200 (-0.009291) | 0.017252 / 0.141683 (-0.124431) | 1.064298 / 1.452155 (-0.387857) | 1.125374 / 1.492716 (-0.367342) |\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) | 0.091506 / 0.018006 (0.073499) | 0.298570 / 0.000490 (0.298080) | 0.000741 / 0.000200 (0.000541) | 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) | 0.020232 / 0.037411 (-0.017179) | 0.069640 / 0.014526 (0.055114) | 0.081360 / 0.176557 (-0.095197) | 0.116955 / 0.737135 (-0.620180) | 0.080920 / 0.296338 (-0.215418) |\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) | 0.283000 / 0.215209 (0.067791) | 2.802526 / 2.077655 (0.724871) | 1.534631 / 1.504120 (0.030511) | 1.407260 / 1.541195 (-0.133935) | 1.409111 / 1.468490 (-0.059379) | 0.534892 / 4.584777 (-4.049885) | 2.350516 / 3.745712 (-1.395196) | 2.550444 / 5.269862 (-2.719418) | 1.661747 / 4.565676 (-2.903930) | 0.060978 / 0.424275 (-0.363297) | 0.005300 / 0.007607 (-0.002308) | 0.367418 / 0.226044 (0.141373) | 3.338046 / 2.268929 (1.069117) | 1.883914 / 55.444624 (-53.560710) | 1.638561 / 6.876477 (-5.237916) | 1.751547 / 2.142072 (-0.390526) | 0.633318 / 4.805227 (-4.171910) | 0.114971 / 6.500664 (-6.385693) | 0.040202 / 0.075469 (-0.035267) |\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) | 0.962053 / 1.841788 (-0.879735) | 11.600643 / 8.074308 (3.526334) | 9.526461 / 10.191392 (-0.664931) | 0.123909 / 0.680424 (-0.556515) | 0.015944 / 0.534201 (-0.518257) | 0.271542 / 0.579283 (-0.307741) | 0.254366 / 0.434364 (-0.179998) | 0.300499 / 0.540337 (-0.239838) | 0.409122 / 1.386936 (-0.977814) |\n\n</details>\n</details>\n\n![](https://cml.dev/watermark.png#0acb27347a3c03efde612023235201a777e08e72 \"CML watermark\")\n"
] |
6,706
|
Update ruff
| null |
[] |
{
"url": "https://api.github.com/repos/huggingface/datasets/pulls/6706",
"html_url": "https://github.com/huggingface/datasets/pull/6706",
"diff_url": "https://github.com/huggingface/datasets/pull/6706.diff",
"patch_url": "https://github.com/huggingface/datasets/pull/6706.patch",
"merged_at": "2024-03-01T16:52:17"
}
| true
|
[
"The docs for this PR live [here](https://moon-ci-docs.huggingface.co/docs/datasets/pr_6706). All of your documentation changes will be reflected on that endpoint. The docs are available until 30 days after the last update.",
"<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) | 0.005014 / 0.011353 (-0.006339) | 0.003324 / 0.011008 (-0.007685) | 0.062501 / 0.038508 (0.023993) | 0.027633 / 0.023109 (0.004524) | 0.245693 / 0.275898 (-0.030205) | 0.271963 / 0.323480 (-0.051517) | 0.003062 / 0.007986 (-0.004923) | 0.002646 / 0.004328 (-0.001683) | 0.049020 / 0.004250 (0.044769) | 0.042381 / 0.037052 (0.005328) | 0.269729 / 0.258489 (0.011240) | 0.289052 / 0.293841 (-0.004789) | 0.027138 / 0.128546 (-0.101408) | 0.010246 / 0.075646 (-0.065400) | 0.205378 / 0.419271 (-0.213893) | 0.035792 / 0.043533 (-0.007741) | 0.247204 / 0.255139 (-0.007935) | 0.271805 / 0.283200 (-0.011394) | 0.019541 / 0.141683 (-0.122142) | 1.129335 / 1.452155 (-0.322820) | 1.174088 / 1.492716 (-0.318629) |\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) | 0.091340 / 0.018006 (0.073334) | 0.300037 / 0.000490 (0.299547) | 0.000214 / 0.000200 (0.000014) | 0.000046 / 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) | 0.018360 / 0.037411 (-0.019051) | 0.061239 / 0.014526 (0.046713) | 0.072304 / 0.176557 (-0.104253) | 0.118883 / 0.737135 (-0.618253) | 0.073562 / 0.296338 (-0.222777) |\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) | 0.284478 / 0.215209 (0.069269) | 2.761819 / 2.077655 (0.684165) | 1.443757 / 1.504120 (-0.060363) | 1.315221 / 1.541195 (-0.225974) | 1.333930 / 1.468490 (-0.134560) | 0.581470 / 4.584777 (-4.003307) | 2.422530 / 3.745712 (-1.323183) | 2.869898 / 5.269862 (-2.399963) | 1.789159 / 4.565676 (-2.776517) | 0.063708 / 0.424275 (-0.360567) | 0.004922 / 0.007607 (-0.002685) | 0.337352 / 0.226044 (0.111307) | 3.290192 / 2.268929 (1.021263) | 1.840192 / 55.444624 (-53.604432) | 1.543008 / 6.876477 (-5.333469) | 1.548947 / 2.142072 (-0.593125) | 0.655129 / 4.805227 (-4.150098) | 0.119010 / 6.500664 (-6.381654) | 0.042583 / 0.075469 (-0.032886) |\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) | 0.981333 / 1.841788 (-0.860455) | 11.349564 / 8.074308 (3.275256) | 9.397603 / 10.191392 (-0.793789) | 0.142151 / 0.680424 (-0.538273) | 0.013850 / 0.534201 (-0.520351) | 0.286323 / 0.579283 (-0.292960) | 0.265223 / 0.434364 (-0.169141) | 0.335322 / 0.540337 (-0.205015) | 0.441727 / 1.386936 (-0.945209) |\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) | 0.005134 / 0.011353 (-0.006219) | 0.003216 / 0.011008 (-0.007792) | 0.049401 / 0.038508 (0.010893) | 0.031509 / 0.023109 (0.008400) | 0.262211 / 0.275898 (-0.013687) | 0.284814 / 0.323480 (-0.038665) | 0.004165 / 0.007986 (-0.003821) | 0.002693 / 0.004328 (-0.001636) | 0.048088 / 0.004250 (0.043838) | 0.043609 / 0.037052 (0.006557) | 0.271126 / 0.258489 (0.012637) | 0.301374 / 0.293841 (0.007533) | 0.028891 / 0.128546 (-0.099655) | 0.009911 / 0.075646 (-0.065735) | 0.057334 / 0.419271 (-0.361938) | 0.050936 / 0.043533 (0.007403) | 0.258883 / 0.255139 (0.003744) | 0.282884 / 0.283200 (-0.000315) | 0.017475 / 0.141683 (-0.124208) | 1.167562 / 1.452155 (-0.284593) | 1.214081 / 1.492716 (-0.278636) |\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) | 0.096890 / 0.018006 (0.078884) | 0.315819 / 0.000490 (0.315329) | 0.000218 / 0.000200 (0.000018) | 0.000054 / 0.000054 (0.000000) |\n\n### Benchmark: benchmark_indices_mapping.json\n\n| metric | select | shard | shuffle | sort | train_test_split |\n|--------|---|---|---|---|---|\n| new / old (diff) | 0.021775 / 0.037411 (-0.015637) | 0.075816 / 0.014526 (0.061290) | 0.086992 / 0.176557 (-0.089564) | 0.125816 / 0.737135 (-0.611319) | 0.090343 / 0.296338 (-0.205995) |\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) | 0.295204 / 0.215209 (0.079995) | 2.903129 / 2.077655 (0.825475) | 1.629838 / 1.504120 (0.125718) | 1.531862 / 1.541195 (-0.009332) | 1.504614 / 1.468490 (0.036123) | 0.572910 / 4.584777 (-4.011867) | 2.482555 / 3.745712 (-1.263157) | 2.637259 / 5.269862 (-2.632603) | 1.733049 / 4.565676 (-2.832628) | 0.063239 / 0.424275 (-0.361036) | 0.005037 / 0.007607 (-0.002570) | 0.346657 / 0.226044 (0.120612) | 3.446469 / 2.268929 (1.177540) | 2.017864 / 55.444624 (-53.426761) | 1.688704 / 6.876477 (-5.187773) | 1.790813 / 2.142072 (-0.351259) | 0.660769 / 4.805227 (-4.144458) | 0.115582 / 6.500664 (-6.385082) | 0.040111 / 0.075469 (-0.035358) |\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) | 1.058089 / 1.841788 (-0.783699) | 11.998171 / 8.074308 (3.923863) | 10.459128 / 10.191392 (0.267736) | 0.149653 / 0.680424 (-0.530771) | 0.015015 / 0.534201 (-0.519186) | 0.289973 / 0.579283 (-0.289310) | 0.274217 / 0.434364 (-0.160147) | 0.351057 / 0.540337 (-0.189281) | 0.434295 / 1.386936 (-0.952641) |\n\n</details>\n</details>\n\n![](https://cml.dev/watermark.png#cb33582eb2262cec5ec6f238b50a7d043ab3ca94 \"CML watermark\")\n"
] |
6,705
|
Fix data_files when passing data_dir
|
This code should not return empty data files
```python
from datasets import load_dataset_builder
revision = "3d406e70bc21c3ca92a9a229b4c6fc3ed88279fd"
b = load_dataset_builder("bigcode/the-stack-v2-dedup", data_dir="data/Dockerfile", revision=revision)
print(b.config.data_files)
```
Previously it would return no data files because it would apply the YAML `data_files: data/**/train-*` pattern to this directory
cc @anton-l
|
[] |
{
"url": "https://api.github.com/repos/huggingface/datasets/pulls/6705",
"html_url": "https://github.com/huggingface/datasets/pull/6705",
"diff_url": "https://github.com/huggingface/datasets/pull/6705.diff",
"patch_url": "https://github.com/huggingface/datasets/pull/6705.patch",
"merged_at": "2024-03-01T18:52:49"
}
| true
|
[
"The docs for this PR live [here](https://moon-ci-docs.huggingface.co/docs/datasets/pr_6705). All of your documentation changes will be reflected on that endpoint. The docs are available until 30 days after the last update.",
"<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) | 0.005014 / 0.011353 (-0.006339) | 0.003371 / 0.011008 (-0.007637) | 0.063622 / 0.038508 (0.025114) | 0.026551 / 0.023109 (0.003442) | 0.244602 / 0.275898 (-0.031296) | 0.269981 / 0.323480 (-0.053499) | 0.003959 / 0.007986 (-0.004027) | 0.002678 / 0.004328 (-0.001650) | 0.049421 / 0.004250 (0.045170) | 0.039926 / 0.037052 (0.002873) | 0.256609 / 0.258489 (-0.001881) | 0.281934 / 0.293841 (-0.011907) | 0.027794 / 0.128546 (-0.100752) | 0.010130 / 0.075646 (-0.065516) | 0.207471 / 0.419271 (-0.211800) | 0.035423 / 0.043533 (-0.008110) | 0.246987 / 0.255139 (-0.008152) | 0.265413 / 0.283200 (-0.017787) | 0.018287 / 0.141683 (-0.123396) | 1.117550 / 1.452155 (-0.334604) | 1.151713 / 1.492716 (-0.341003) |\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) | 0.095632 / 0.018006 (0.077626) | 0.304315 / 0.000490 (0.303825) | 0.000214 / 0.000200 (0.000014) | 0.000049 / 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) | 0.018591 / 0.037411 (-0.018820) | 0.062081 / 0.014526 (0.047555) | 0.075137 / 0.176557 (-0.101420) | 0.119116 / 0.737135 (-0.618020) | 0.075254 / 0.296338 (-0.221085) |\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) | 0.286161 / 0.215209 (0.070952) | 2.793824 / 2.077655 (0.716169) | 1.492523 / 1.504120 (-0.011597) | 1.372158 / 1.541195 (-0.169037) | 1.385921 / 1.468490 (-0.082569) | 0.568700 / 4.584777 (-4.016077) | 2.340451 / 3.745712 (-1.405262) | 2.712022 / 5.269862 (-2.557840) | 1.712479 / 4.565676 (-2.853197) | 0.060906 / 0.424275 (-0.363369) | 0.004909 / 0.007607 (-0.002698) | 0.338227 / 0.226044 (0.112182) | 3.331329 / 2.268929 (1.062400) | 1.845646 / 55.444624 (-53.598978) | 1.559384 / 6.876477 (-5.317093) | 1.577683 / 2.142072 (-0.564390) | 0.629367 / 4.805227 (-4.175860) | 0.118645 / 6.500664 (-6.382019) | 0.041517 / 0.075469 (-0.033952) |\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) | 0.962237 / 1.841788 (-0.879551) | 11.232566 / 8.074308 (3.158258) | 9.627141 / 10.191392 (-0.564251) | 0.129732 / 0.680424 (-0.550692) | 0.013701 / 0.534201 (-0.520500) | 0.291869 / 0.579283 (-0.287414) | 0.269298 / 0.434364 (-0.165066) | 0.342502 / 0.540337 (-0.197835) | 0.455891 / 1.386936 (-0.931045) |\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) | 0.005256 / 0.011353 (-0.006097) | 0.003419 / 0.011008 (-0.007589) | 0.049681 / 0.038508 (0.011173) | 0.029566 / 0.023109 (0.006457) | 0.268010 / 0.275898 (-0.007888) | 0.293721 / 0.323480 (-0.029759) | 0.004249 / 0.007986 (-0.003737) | 0.002643 / 0.004328 (-0.001685) | 0.048758 / 0.004250 (0.044508) | 0.044294 / 0.037052 (0.007241) | 0.279584 / 0.258489 (0.021095) | 0.311150 / 0.293841 (0.017309) | 0.029443 / 0.128546 (-0.099103) | 0.010314 / 0.075646 (-0.065333) | 0.057770 / 0.419271 (-0.361501) | 0.050953 / 0.043533 (0.007420) | 0.268283 / 0.255139 (0.013144) | 0.289155 / 0.283200 (0.005956) | 0.017742 / 0.141683 (-0.123941) | 1.163963 / 1.452155 (-0.288192) | 1.200580 / 1.492716 (-0.292136) |\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) | 0.096365 / 0.018006 (0.078359) | 0.307257 / 0.000490 (0.306767) | 0.000265 / 0.000200 (0.000065) | 0.000049 / 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) | 0.021862 / 0.037411 (-0.015550) | 0.075502 / 0.014526 (0.060976) | 0.087800 / 0.176557 (-0.088756) | 0.125468 / 0.737135 (-0.611667) | 0.088207 / 0.296338 (-0.208132) |\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) | 0.324184 / 0.215209 (0.108975) | 3.198442 / 2.077655 (1.120787) | 1.862801 / 1.504120 (0.358682) | 1.728637 / 1.541195 (0.187443) | 1.727997 / 1.468490 (0.259507) | 0.571590 / 4.584777 (-4.013187) | 2.448661 / 3.745712 (-1.297051) | 2.665943 / 5.269862 (-2.603919) | 1.731718 / 4.565676 (-2.833958) | 0.063644 / 0.424275 (-0.360631) | 0.004989 / 0.007607 (-0.002619) | 0.364543 / 0.226044 (0.138498) | 3.615859 / 2.268929 (1.346930) | 2.131637 / 55.444624 (-53.312987) | 1.857317 / 6.876477 (-5.019159) | 1.992813 / 2.142072 (-0.149260) | 0.654662 / 4.805227 (-4.150565) | 0.117631 / 6.500664 (-6.383034) | 0.040934 / 0.075469 (-0.034535) |\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) | 1.013802 / 1.841788 (-0.827985) | 11.899873 / 8.074308 (3.825565) | 10.291297 / 10.191392 (0.099905) | 0.155245 / 0.680424 (-0.525179) | 0.014449 / 0.534201 (-0.519752) | 0.286331 / 0.579283 (-0.292952) | 0.273111 / 0.434364 (-0.161253) | 0.321182 / 0.540337 (-0.219155) | 0.433406 / 1.386936 (-0.953530) |\n\n</details>\n</details>\n\n![](https://cml.dev/watermark.png#e5406f9a9a453f2c0614c2ee26975e5973edc278 \"CML watermark\")\n"
] |
6,704
|
Improve default patterns resolution
|
Separate the default patterns that match directories from the ones matching files and ensure directories are checked first (reverts the change from https://github.com/huggingface/datasets/pull/6244, which merged these patterns). Also, ensure that the glob patterns do not overlap to avoid duplicates in the result.
Additionally, replace `get_fs_token_paths` with `url_to_fs` to avoid [unnecessary glob calls](https://github.com/fsspec/filesystem_spec/blob/14dce8ca78f7aa509a20edb263bff83a7760c24d/fsspec/core.py#L655-L656).
fix https://github.com/huggingface/datasets/issues/6259
fix https://github.com/huggingface/datasets/issues/6272
|
[] |
{
"url": "https://api.github.com/repos/huggingface/datasets/pulls/6704",
"html_url": "https://github.com/huggingface/datasets/pull/6704",
"diff_url": "https://github.com/huggingface/datasets/pull/6704.diff",
"patch_url": "https://github.com/huggingface/datasets/pull/6704.patch",
"merged_at": "2024-03-15T15:22:03"
}
| true
|
[
"The docs for this PR live [here](https://moon-ci-docs.huggingface.co/docs/datasets/pr_6704). All of your documentation changes will be reflected on that endpoint. The docs are available until 30 days after the last update.",
"Awesome !\r\n\r\nNote that it can still create duplicates if a path matches several dir patterns, e.g.\r\n\r\n```\r\ndata/train-train/data/txt\r\n```\r\nmatches two dir patterns:\r\n```\r\n**/{keyword}[{sep}]*/**\r\n**/*[{sep}]{keyword}/**\r\n```\r\n\r\nPS: feel free to update your branch, I just updated ruff on `main`",
"Yes, I didn't mention that case on purpose 🙂. One solution would be deprecating the `**/*[{sep}]{keyword}/**` pattern (and eventually removing it). This way, the directory patterns would align more with the filename ones. Or do you think this is too big of a breaking change?",
"I think it's too big of a breaking change yes :/ (and would make the docs / logic more complex for users to get imo) Though I think your approach is already a nice step in the right direction",
"These changes to the `resolve_pattern` function lead to 20-30x faster local file resolution in my benchmarks.",
"Nice ! Though since `fsspec` caches the filesystem, is there a risk when adding new files and reloading a dataset ?\r\n\r\n\r\n```python\r\nwith open(\"my/local/dir/0000.txt\", \"w\") as f:\r\n f.write(\"Hello there\")\r\nd1 = load_dataset(\"my/local/dir\")\r\nwith open(\"my/local/dir/0001.txt\", \"w\") as f:\r\n f.write(\"General Kenobi\")\r\nd2 = load_dataset(\"my/local/dir\")\r\nassert list(d1) != list(d2)\r\n```",
"Yes. But I think I have a solution for this.",
"I'm not satisfied with the context manager approach...\r\n\r\nA clean solution would require a bigger rewrite of the resolution logic (e.g., merging `get_data_patterns` and `DataFilesDict.from_patterns` into a `get_data_files` function that would build the `DataFilesDict` by matching the paths using `fs.find` and `fsspec.utils.glob_translate` (available in `fsspec>=2023.12.0`))\r\n\r\nThe current changes make the local resolution 2-3x faster, which is good enough for now, I think.",
"<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) | 0.004888 / 0.011353 (-0.006465) | 0.003267 / 0.011008 (-0.007742) | 0.065117 / 0.038508 (0.026609) | 0.029416 / 0.023109 (0.006306) | 0.232021 / 0.275898 (-0.043877) | 0.258053 / 0.323480 (-0.065427) | 0.003971 / 0.007986 (-0.004014) | 0.002550 / 0.004328 (-0.001779) | 0.049126 / 0.004250 (0.044876) | 0.040620 / 0.037052 (0.003568) | 0.253437 / 0.258489 (-0.005052) | 0.273583 / 0.293841 (-0.020258) | 0.026775 / 0.128546 (-0.101771) | 0.010073 / 0.075646 (-0.065573) | 0.219089 / 0.419271 (-0.200183) | 0.035047 / 0.043533 (-0.008486) | 0.247661 / 0.255139 (-0.007478) | 0.258674 / 0.283200 (-0.024525) | 0.018428 / 0.141683 (-0.123255) | 1.130394 / 1.452155 (-0.321761) | 1.173167 / 1.492716 (-0.319549) |\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) | 0.092581 / 0.018006 (0.074574) | 0.303657 / 0.000490 (0.303167) | 0.000215 / 0.000200 (0.000015) | 0.000051 / 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) | 0.018640 / 0.037411 (-0.018771) | 0.062032 / 0.014526 (0.047506) | 0.073982 / 0.176557 (-0.102575) | 0.121499 / 0.737135 (-0.615636) | 0.076780 / 0.296338 (-0.219559) |\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) | 0.279411 / 0.215209 (0.064202) | 2.737977 / 2.077655 (0.660322) | 1.454135 / 1.504120 (-0.049985) | 1.343144 / 1.541195 (-0.198051) | 1.339876 / 1.468490 (-0.128614) | 0.567306 / 4.584777 (-4.017471) | 2.372569 / 3.745712 (-1.373143) | 2.716810 / 5.269862 (-2.553052) | 1.697895 / 4.565676 (-2.867782) | 0.061804 / 0.424275 (-0.362471) | 0.004986 / 0.007607 (-0.002622) | 0.332721 / 0.226044 (0.106676) | 3.274572 / 2.268929 (1.005644) | 1.789900 / 55.444624 (-53.654725) | 1.536346 / 6.876477 (-5.340131) | 1.551940 / 2.142072 (-0.590132) | 0.634539 / 4.805227 (-4.170688) | 0.115860 / 6.500664 (-6.384805) | 0.041737 / 0.075469 (-0.033732) |\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) | 1.024469 / 1.841788 (-0.817319) | 11.327496 / 8.074308 (3.253188) | 9.265855 / 10.191392 (-0.925537) | 0.142200 / 0.680424 (-0.538224) | 0.013945 / 0.534201 (-0.520256) | 0.289670 / 0.579283 (-0.289614) | 0.269240 / 0.434364 (-0.165124) | 0.324748 / 0.540337 (-0.215590) | 0.421393 / 1.386936 (-0.965543) |\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) | 0.005284 / 0.011353 (-0.006069) | 0.003351 / 0.011008 (-0.007658) | 0.049973 / 0.038508 (0.011465) | 0.030257 / 0.023109 (0.007148) | 0.273660 / 0.275898 (-0.002238) | 0.300328 / 0.323480 (-0.023152) | 0.004133 / 0.007986 (-0.003852) | 0.002614 / 0.004328 (-0.001715) | 0.048055 / 0.004250 (0.043804) | 0.044731 / 0.037052 (0.007678) | 0.290257 / 0.258489 (0.031768) | 0.321243 / 0.293841 (0.027402) | 0.029542 / 0.128546 (-0.099004) | 0.010074 / 0.075646 (-0.065573) | 0.057944 / 0.419271 (-0.361327) | 0.051267 / 0.043533 (0.007734) | 0.276278 / 0.255139 (0.021139) | 0.302464 / 0.283200 (0.019264) | 0.018231 / 0.141683 (-0.123452) | 1.140782 / 1.452155 (-0.311373) | 1.182991 / 1.492716 (-0.309725) |\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) | 0.092325 / 0.018006 (0.074319) | 0.302610 / 0.000490 (0.302121) | 0.000202 / 0.000200 (0.000002) | 0.000049 / 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) | 0.021458 / 0.037411 (-0.015954) | 0.074883 / 0.014526 (0.060357) | 0.085747 / 0.176557 (-0.090809) | 0.125506 / 0.737135 (-0.611629) | 0.086921 / 0.296338 (-0.209417) |\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) | 0.290485 / 0.215209 (0.075276) | 2.853898 / 2.077655 (0.776243) | 1.615606 / 1.504120 (0.111486) | 1.491797 / 1.541195 (-0.049397) | 1.515981 / 1.468490 (0.047491) | 0.566760 / 4.584777 (-4.018017) | 2.462593 / 3.745712 (-1.283119) | 2.765516 / 5.269862 (-2.504345) | 1.755078 / 4.565676 (-2.810598) | 0.063614 / 0.424275 (-0.360661) | 0.005040 / 0.007607 (-0.002567) | 0.347957 / 0.226044 (0.121912) | 3.464258 / 2.268929 (1.195330) | 1.992273 / 55.444624 (-53.452351) | 1.699147 / 6.876477 (-5.177330) | 1.868438 / 2.142072 (-0.273635) | 0.660756 / 4.805227 (-4.144471) | 0.118142 / 6.500664 (-6.382522) | 0.041974 / 0.075469 (-0.033495) |\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) | 1.012206 / 1.841788 (-0.829581) | 12.343735 / 8.074308 (4.269427) | 10.321975 / 10.191392 (0.130583) | 0.140007 / 0.680424 (-0.540417) | 0.015755 / 0.534201 (-0.518446) | 0.291978 / 0.579283 (-0.287305) | 0.278792 / 0.434364 (-0.155572) | 0.325366 / 0.540337 (-0.214972) | 0.439403 / 1.386936 (-0.947533) |\n\n</details>\n</details>\n\n![](https://cml.dev/watermark.png#d1d3c06a651c6ad5142f331cb5dc0008ddcade33 \"CML watermark\")\n",
"This change is breaking in \r\n\r\nhttps://github.com/huggingface/datasets/blob/f96e74d5c633cd5435dd526adb4a74631eb05c43/src/datasets/arrow_dataset.py#L1515\r\n\r\nwhen the input is `pathlib.Path`. The issue is that `url_to_fs` expects a `str` and cannot deal with `Path`. `get_fs_token_paths` converts to `str` so it is not a problem",
"I opened https://github.com/huggingface/datasets/pull/6828 to add proper Path support to save_to_disk / load_from_disk"
] |
6,703
|
Unable to load dataset that was saved with `save_to_disk`
|
### Describe the bug
I get the following error message: You are trying to load a dataset that was saved using `save_to_disk`. Please use `load_from_disk` instead.
### Steps to reproduce the bug
1. Save a dataset with `save_to_disk`
2. Try to load it with `load_datasets`
### Expected behavior
I am able to load the dataset again with `load_datasets` which most packages uses over `load_from_disk`. I want to have a workaround that allows me to create the same indexing that `push_to_hub` creates for you before using `save_to_disk` - how can that be achieved?
### Environment info
datasets 2.17.1, python 3.10
|
[] |
{
"url": null,
"html_url": null,
"diff_url": null,
"patch_url": null,
"merged_at": null
}
| false
|
[
"`save_to_disk` uses a special serialization that can only be read using `load_from_disk`.\r\n\r\nContrary to `load_dataset`, `load_from_disk` directly loads Arrow files and uses the dataset directory as cache.\r\n\r\nOn the other hand `load_dataset` does a conversion step to get Arrow files from the raw data files (could be in JSON, CSV, Parquet etc.) and caches them in the `datasets` cache directory (default is `~/.cache/huggingface/datasets`). We haven't implemented any logic in `load_dataset` to support datasets saved with `save_to_disk` because they don't use the same cache.\r\n\r\nEDIT: note that you can save your dataset in Parquet format locally using `.to.parquet()` (make sure to shard in multiple files your dataset if it's multiple GBs - you can use `.shard()` + `.to_parquet()` to do that) and you'll be able to reload it using `load_dataset`",
"@lhoestq, so is it correctly understood that if I run `to_parquet()` and then `save_to_disk()`, I can load it with `load_dataset`? If yes, then it would resolve this issue (and should probably be documented somewhere 😄)",
"Here is an example:\r\n```python\r\nds.to_parquet(\"my/local/dir/data.parquet\")\r\n\r\n# later\r\nds = load_dataset(\"my/local/dir\")\r\n```\r\n\r\nand for bigger datasets:\r\n```python\r\nnum_shards = 1024 # set number of files to save (e.g. try to have files smaller than 5GB)\r\nfor shard_idx in num_shards:\r\n shard = ds.shard(index=shard_idx, num_shards=num_shards)\r\n shard.to_parquet(f\"my/local/dir/{shard_idx:05d}.parquet\") # 00000.parquet to 01023.parquet\r\n\r\n# later\r\nds = load_dataset(\"my/local/dir\")\r\n```\r\n\r\n\r\nI hope this helps :)",
"Thanks for helping out! Does this approach work with `s3fs`? e.g. something like this:\r\n\r\n```python\r\nimport s3fs\r\ns3 = s3fs.S3FileSystem(anon=True)\r\nwith s3.open('mybucket/new-file.parquet', 'w') as f:\r\n ds.to_parquet(f)\r\n```\r\n\r\nThis is instead of `save_to_disk` to save to an S3 bucket.\r\n\r\nOtherwise, I am not sure how to make this work when saving the dataset to an S3 bucket. Would `dataset.set_format(\"arrow\")` work as a replacement?",
"`load_dataset` does't support S3 buckets unfortunately :/",
"> `load_dataset` does't support S3 buckets unfortunately :/\r\n\r\nI am aware but I have some code that downloads it to disk before using that method. The most important part is to store it in a format that load_dataset is compatible with. ",
"Feel free to use Parquet then :)",
"I ended up with this. Not ideal to save to local disk, but it works and loads via `load_datasets` after downloading from S3 with another method.\r\n\r\n```python\r\nwith tempfile.TemporaryDirectory() as dir:\r\n dataset_nbytes = ds._estimate_nbytes()\r\n max_shard_size_local = convert_file_size_to_int(max_shard_size)\r\n num_shards = int(dataset_nbytes / max_shard_size_local) + 1\r\n\r\n for shard_idx in range(num_shards):\r\n shard = ds.shard(index=shard_idx, num_shards=num_shards)\r\n shard.to_parquet(f\"{dir}/{shard_idx:05d}.parquet\")\r\n \r\n fs.upload(\r\n lpath=dir,\r\n rpath=s3_path,\r\n recursive=True,\r\n )\r\n```"
] |
6,702
|
Push samples to dataset on hub without having the dataset locally
|
### Feature request
Say I have the following code:
```
from datasets import Dataset
import pandas as pd
new_data = {
"column_1": ["value1", "value2"],
"column_2": ["value3", "value4"],
}
df_new = pd.DataFrame(new_data)
dataset_new = Dataset.from_pandas(df_new)
# add these samples to a remote dataset
```
It would be great to have a way to push dataset_new to a remote dataset that respects the same schema. This way one would not have to do the following:
```
from datasets import load_dataset
dataset = load_dataset('username/dataset_name', use_auth_token='your_hf_token_here')
updated_dataset = dataset['train'].concatenate(dataset_new)
updated_dataset.push_to_hub('username/dataset_name', use_auth_token='your_hf_token_here')
```
### Motivation
No need to download the dataset.
### Your contribution
Maybe this feature already exists, didnt see it though. I do not have the expertise to do this.
|
[
{
"id": 1935892871,
"node_id": "MDU6TGFiZWwxOTM1ODkyODcx",
"url": "https://api.github.com/repos/huggingface/datasets/labels/enhancement",
"name": "enhancement",
"color": "a2eeef",
"default": true,
"description": "New feature or request"
}
] |
{
"url": null,
"html_url": null,
"diff_url": null,
"patch_url": null,
"merged_at": null
}
| false
|
[
"Hi ! For now I would recommend creating a new Parquet file using `dataset_new.to_parquet()` and upload it to HF using `huggingface_hub` every time you get a new batch of data. You can name the Parquet files `0000.parquet`, `0001.parquet`, etc.\r\n\r\nThough maybe make sure to not upload one file per sample since that would be inefficient. You can buffer your data and upload when you have enough new samples for example",
"This is excellent, thanks!"
] |
6,701
|
Base parquet batch_size on parquet row group size
|
This allows to stream datasets like [Major-TOM/Core-S2L2A](https://huggingface.co/datasets/Major-TOM/Core-S2L2A) which have row groups with few rows (one row is ~10MB). Previously the cold start would take a lot of time and OOM because it would download many row groups before yielding the first example.
I tried on OpenOrca and imagenet-hard and it does't affect overall throughput.
Even if the overall throughput doesn't change for datasets like imagenet-hard with big rows, note that it does create shorter and more frequent pauses to download the next row group. Though I find it fine because previously the pauses were less frequent but very long (downloading multiple row groups at a time)
|
[] |
{
"url": "https://api.github.com/repos/huggingface/datasets/pulls/6701",
"html_url": "https://github.com/huggingface/datasets/pull/6701",
"diff_url": "https://github.com/huggingface/datasets/pull/6701.diff",
"patch_url": "https://github.com/huggingface/datasets/pull/6701.patch",
"merged_at": "2024-02-29T15:08:55"
}
| true
|
[
"The docs for this PR live [here](https://moon-ci-docs.huggingface.co/docs/datasets/pr_6701). All of your documentation changes will be reflected on that endpoint. The docs are available until 30 days after the last update.",
"<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) | 0.005490 / 0.011353 (-0.005863) | 0.003709 / 0.011008 (-0.007299) | 0.064192 / 0.038508 (0.025684) | 0.029581 / 0.023109 (0.006472) | 0.251086 / 0.275898 (-0.024812) | 0.267306 / 0.323480 (-0.056174) | 0.003074 / 0.007986 (-0.004912) | 0.003340 / 0.004328 (-0.000988) | 0.048820 / 0.004250 (0.044569) | 0.045370 / 0.037052 (0.008318) | 0.260384 / 0.258489 (0.001895) | 0.284558 / 0.293841 (-0.009283) | 0.027732 / 0.128546 (-0.100814) | 0.010661 / 0.075646 (-0.064986) | 0.213403 / 0.419271 (-0.205868) | 0.036283 / 0.043533 (-0.007250) | 0.250107 / 0.255139 (-0.005032) | 0.265220 / 0.283200 (-0.017980) | 0.021021 / 0.141683 (-0.120661) | 1.112058 / 1.452155 (-0.340096) | 1.169039 / 1.492716 (-0.323678) |\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) | 0.095008 / 0.018006 (0.077002) | 0.303509 / 0.000490 (0.303019) | 0.000233 / 0.000200 (0.000033) | 0.000052 / 0.000054 (-0.000002) |\n\n### Benchmark: benchmark_indices_mapping.json\n\n| metric | select | shard | shuffle | sort | train_test_split |\n|--------|---|---|---|---|---|\n| new / old (diff) | 0.018224 / 0.037411 (-0.019187) | 0.061366 / 0.014526 (0.046841) | 0.073584 / 0.176557 (-0.102972) | 0.119869 / 0.737135 (-0.617266) | 0.074228 / 0.296338 (-0.222111) |\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) | 0.288147 / 0.215209 (0.072938) | 2.824419 / 2.077655 (0.746764) | 1.478530 / 1.504120 (-0.025590) | 1.350127 / 1.541195 (-0.191067) | 1.349622 / 1.468490 (-0.118868) | 0.568058 / 4.584777 (-4.016719) | 2.377494 / 3.745712 (-1.368218) | 2.720767 / 5.269862 (-2.549094) | 1.710763 / 4.565676 (-2.854914) | 0.061498 / 0.424275 (-0.362778) | 0.004893 / 0.007607 (-0.002715) | 0.335633 / 0.226044 (0.109588) | 3.380646 / 2.268929 (1.111717) | 1.802436 / 55.444624 (-53.642188) | 1.562737 / 6.876477 (-5.313739) | 1.566267 / 2.142072 (-0.575806) | 0.629058 / 4.805227 (-4.176169) | 0.116307 / 6.500664 (-6.384357) | 0.042174 / 0.075469 (-0.033295) |\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) | 0.950945 / 1.841788 (-0.890842) | 11.279009 / 8.074308 (3.204701) | 9.433251 / 10.191392 (-0.758141) | 0.138964 / 0.680424 (-0.541460) | 0.014155 / 0.534201 (-0.520046) | 0.284065 / 0.579283 (-0.295218) | 0.263301 / 0.434364 (-0.171063) | 0.331932 / 0.540337 (-0.208406) | 0.441656 / 1.386936 (-0.945280) |\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) | 0.005132 / 0.011353 (-0.006221) | 0.003484 / 0.011008 (-0.007524) | 0.049040 / 0.038508 (0.010532) | 0.030254 / 0.023109 (0.007145) | 0.277141 / 0.275898 (0.001243) | 0.295242 / 0.323480 (-0.028238) | 0.004295 / 0.007986 (-0.003690) | 0.002632 / 0.004328 (-0.001696) | 0.048540 / 0.004250 (0.044290) | 0.044787 / 0.037052 (0.007734) | 0.287736 / 0.258489 (0.029247) | 0.313146 / 0.293841 (0.019305) | 0.029340 / 0.128546 (-0.099206) | 0.010204 / 0.075646 (-0.065442) | 0.059058 / 0.419271 (-0.360214) | 0.051033 / 0.043533 (0.007500) | 0.274086 / 0.255139 (0.018947) | 0.293048 / 0.283200 (0.009848) | 0.019573 / 0.141683 (-0.122110) | 1.174032 / 1.452155 (-0.278123) | 1.227107 / 1.492716 (-0.265609) |\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) | 0.094896 / 0.018006 (0.076890) | 0.303519 / 0.000490 (0.303029) | 0.000223 / 0.000200 (0.000023) | 0.000049 / 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) | 0.021495 / 0.037411 (-0.015917) | 0.074234 / 0.014526 (0.059708) | 0.086212 / 0.176557 (-0.090345) | 0.125052 / 0.737135 (-0.612084) | 0.087464 / 0.296338 (-0.208874) |\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) | 0.297098 / 0.215209 (0.081889) | 2.970944 / 2.077655 (0.893289) | 1.650101 / 1.504120 (0.145981) | 1.532694 / 1.541195 (-0.008501) | 1.513652 / 1.468490 (0.045162) | 0.559614 / 4.584777 (-4.025163) | 2.404848 / 3.745712 (-1.340865) | 2.627851 / 5.269862 (-2.642011) | 1.707550 / 4.565676 (-2.858127) | 0.061821 / 0.424275 (-0.362454) | 0.005012 / 0.007607 (-0.002595) | 0.342462 / 0.226044 (0.116417) | 3.401703 / 2.268929 (1.132774) | 1.991632 / 55.444624 (-53.452993) | 1.737706 / 6.876477 (-5.138771) | 1.837457 / 2.142072 (-0.304616) | 0.638845 / 4.805227 (-4.166383) | 0.114773 / 6.500664 (-6.385891) | 0.040175 / 0.075469 (-0.035294) |\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) | 1.038286 / 1.841788 (-0.803501) | 11.885757 / 8.074308 (3.811448) | 10.061530 / 10.191392 (-0.129862) | 0.140824 / 0.680424 (-0.539600) | 0.015080 / 0.534201 (-0.519121) | 0.287992 / 0.579283 (-0.291291) | 0.273498 / 0.434364 (-0.160866) | 0.326478 / 0.540337 (-0.213860) | 0.426900 / 1.386936 (-0.960036) |\n\n</details>\n</details>\n\n![](https://cml.dev/watermark.png#b02be21047087c5ffc11cf1c072a5aceab517eba \"CML watermark\")\n"
] |
6,700
|
remove_columns is not in-place but the doc shows it is in-place
|
### Describe the bug
The doc of `datasets` v2.17.0/v2.17.1 shows that `remove_columns` is in-place. [link](https://huggingface.co/docs/datasets/v2.17.1/en/package_reference/main_classes#datasets.DatasetDict.remove_columns)
In the text classification example of transformers v4.38.1, the columns are not removed.
https://github.com/huggingface/transformers/blob/a0857740c0e6127485c11476650314df3accc2b6/examples/pytorch/text-classification/run_classification.py#L421
### Steps to reproduce the bug
https://github.com/huggingface/transformers/blob/a0857740c0e6127485c11476650314df3accc2b6/examples/pytorch/text-classification/run_classification.py#L421
### Expected behavior
Actually remove the columns.
### Environment info
1. datasets v2.17.0
2. transformers v4.38.1
|
[] |
{
"url": null,
"html_url": null,
"diff_url": null,
"patch_url": null,
"merged_at": null
}
| false
|
[
"Good catch! I've opened a PR with a fix in the `transformers` repo.",
"@mariosasko Thanks!\r\n\r\nWill the doc of `datasets` be updated?\r\n\r\nI find some possible mistakes in doc about whether `remove_columns` is in-place.\r\n1. [You can also remove a column using map() with remove_columns but the present method is in-place (doesn’t copy the data to a new dataset) and is thus faster.](https://huggingface.co/docs/datasets/v2.17.1/en/package_reference/main_classes#datasets.Dataset.remove_columns)\r\n2. [You can also remove a column using Dataset.map() with remove_columns but the present method is in-place (doesn’t copy the data to a new dataset) and is thus faster.](https://huggingface.co/docs/datasets/v2.17.1/en/package_reference/main_classes#datasets.DatasetDict.remove_columns)\r\n3. [🤗 Datasets also has a remove_columns() function which is faster because it doesn’t copy the data of the remaining columns.](https://huggingface.co/docs/datasets/v2.17.1/en/process#map)",
"I've linked a PR that will fix the usage in the `datasets` docs."
] |
6,699
|
`Dataset` unexpected changed dict data and may cause error
|
### Describe the bug
Will unexpected get keys with `None` value in the parsed json dict.
### Steps to reproduce the bug
```jsonl test.jsonl
{"id": 0, "indexs": {"-1": [0, 10]}}
{"id": 1, "indexs": {"-1": [0, 10]}}
```
```python
dataset = Dataset.from_json('.test.jsonl')
print(dataset[0])
```
Result:
```
{'id': 0, 'indexs': {'-1': [...], '-2': None, '-3': None, '-4': None, '-5': None, '-6': None, '-7': None, '-8': None, '-9': None, ...}}
```
Those keys with `None` value will unexpected appear in the dict.
### Expected behavior
Result should be
```
{'id': 0, 'indexs': {'-1': [0, 10]}}
```
### Environment info
- `datasets` version: 2.16.1
- Platform: Linux-6.5.0-14-generic-x86_64-with-glibc2.35
- Python version: 3.11.6
- `huggingface_hub` version: 0.20.2
- PyArrow version: 14.0.2
- Pandas version: 2.1.4
- `fsspec` version: 2023.10.0
|
[] |
{
"url": null,
"html_url": null,
"diff_url": null,
"patch_url": null,
"merged_at": null
}
| false
|
[
"If `test.jsonl` contains more lines like:\r\n```\r\n{\"id\": 0, \"indexs\": {\"-1\": [0, 10]}}\r\n{\"id\": 1, \"indexs\": {\"-1\": [0, 10]}}\r\n{\"id\": 2, \"indexs\": {\"-2\": [0, 10]}}\r\n...\r\n{\"id\": n, \"indexs\": {\"-9999\": [0, 10]}}\r\n```\r\n\r\n`Dataset.from_json` will just raise an error:\r\n```\r\nAn error occurred while generating the dataset\r\nTypeError: Couldn't cast array of type\r\nstruct<-5942: list<item: int64>, -5943: list<item: int64>, -5944: list<item: int64>, -5945: list<item: int64>, -5946: list<item: int64>, -5947: list<item: int64>, -5948: list<item: int64>, -5949: list<item: int64>: ...\r\nto\r\n{... '-5312': Sequence(feature=Value(dtype='int64', id=None), length=-1, id=None), '-5313': Sequence(feature=Value(dtype='int64', id=None), length=-1, id=None)}\r\n\r\nThe above exception was the direct cause of the following exception:\r\n\r\nTraceback (most recent call last):\r\n File \"/home/scruel/mambaforge/envs/vae/lib/python3.11/runpy.py\", line 198, in _run_module_as_main\r\n return _run_code(code, main_globals, None,\r\n ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^\r\n File \"/home/scruel/mambaforge/envs/vae/lib/python3.11/runpy.py\", line 88, in _run_code\r\n exec(code, run_globals)\r\n File \"/home/scruel/.vscode-server/extensions/ms-python.debugpy-2024.0.0-linux-x64/bundled/libs/debugpy/adapter/../../debugpy/launcher/../../debugpy/__main__.py\", line 39, in <module>\r\n cli.main()\r\n File \"/home/scruel/.vscode-server/extensions/ms-python.debugpy-2024.0.0-linux-x64/bundled/libs/debugpy/adapter/../../debugpy/launcher/../../debugpy/../debugpy/server/cli.py\", line 430, in main\r\n run()\r\n File \"/home/scruel/.vscode-server/extensions/ms-python.debugpy-2024.0.0-linux-x64/bundled/libs/debugpy/adapter/../../debugpy/launcher/../../debugpy/../debugpy/server/cli.py\", line 284, in run_file\r\n runpy.run_path(target, run_name=\"__main__\")\r\n File \"/home/scruel/.vscode-server/extensions/ms-python.debugpy-2024.0.0-linux-x64/bundled/libs/debugpy/_vendored/pydevd/_pydevd_bundle/pydevd_runpy.py\", line 321, in run_path\r\n return _run_module_code(code, init_globals, run_name,\r\n ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^\r\n File \"/home/scruel/.vscode-server/extensions/ms-python.debugpy-2024.0.0-linux-x64/bundled/libs/debugpy/_vendored/pydevd/_pydevd_bundle/pydevd_runpy.py\", line 135, in _run_module_code\r\n _run_code(code, mod_globals, init_globals,\r\n File \"/home/scruel/.vscode-server/extensions/ms-python.debugpy-2024.0.0-linux-x64/bundled/libs/debugpy/_vendored/pydevd/_pydevd_bundle/pydevd_runpy.py\", line 124, in _run_code\r\n exec(code, run_globals)\r\n File \"/home/scruel/Code/Python/Working/llm-memory/data_reader.py\", line 120, in <module>\r\n reader = SnippetReader(jsonl_path, npy_path)\r\n ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^\r\n File \"/home/scruel/Code/Python/Working/llm-memory/data_reader.py\", line 85, in __init__\r\n self._dataset = Dataset.from_json(jsonl_path, features=)\r\n ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^\r\n File \"/home/scruel/mambaforge/envs/vae/lib/python3.11/site-packages/datasets/arrow_dataset.py\", line 1130, in from_json\r\n ).read()\r\n ^^^^^^\r\n File \"/home/scruel/mambaforge/envs/vae/lib/python3.11/site-packages/datasets/io/json.py\", line 59, in read\r\n self.builder.download_and_prepare(\r\n File \"/home/scruel/mambaforge/envs/vae/lib/python3.11/site-packages/datasets/builder.py\", line 1005, in download_and_prepare\r\n self._download_and_prepare(\r\n File \"/home/scruel/mambaforge/envs/vae/lib/python3.11/site-packages/datasets/builder.py\", line 1100, in _download_and_prepare\r\n self._prepare_split(split_generator, **prepare_split_kwargs)\r\n File \"/home/scruel/mambaforge/envs/vae/lib/python3.11/site-packages/datasets/builder.py\", line 1860, in _prepare_split\r\n for job_id, done, content in self._prepare_split_single(\r\n File \"/home/scruel/mambaforge/envs/vae/lib/python3.11/site-packages/datasets/builder.py\", line 2016, in _prepare_split_single\r\n raise DatasetGenerationError(\"An error occurred while generating the dataset\") from e\r\ndatasets.exceptions.DatasetGenerationError: An error occurred while generating the dataset\r\n```",
"Hi! Our JSON parser expects all examples/rows to share the same set of columns (applies to nested columns, too), hence the error. \r\n\r\nTo read the `index` column, we would have to manually cast the input to PyArrow's `pa.map_` type, but this requires a more thorough investigation, as `pa.map_` has limited support in PyArrow."
] |
6,698
|
Faster `xlistdir`
|
Pass `detail=False` to the `fsspec` `listdir` to avoid unnecessarily fetching expensive metadata about the paths.
|
[] |
{
"url": "https://api.github.com/repos/huggingface/datasets/pulls/6698",
"html_url": "https://github.com/huggingface/datasets/pull/6698",
"diff_url": "https://github.com/huggingface/datasets/pull/6698.diff",
"patch_url": "https://github.com/huggingface/datasets/pull/6698.patch",
"merged_at": "2024-02-27T23:38:14"
}
| true
|
[
"The docs for this PR live [here](https://moon-ci-docs.huggingface.co/docs/datasets/pr_6698). All of your documentation changes will be reflected on that endpoint. The docs are available until 30 days after the last update.",
"CI failure is unrelated to the changes.",
"<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) | 0.005499 / 0.011353 (-0.005854) | 0.003824 / 0.011008 (-0.007184) | 0.064230 / 0.038508 (0.025722) | 0.028962 / 0.023109 (0.005853) | 0.283540 / 0.275898 (0.007642) | 0.300774 / 0.323480 (-0.022706) | 0.003405 / 0.007986 (-0.004581) | 0.002796 / 0.004328 (-0.001532) | 0.049834 / 0.004250 (0.045584) | 0.045924 / 0.037052 (0.008872) | 0.274818 / 0.258489 (0.016328) | 0.306189 / 0.293841 (0.012348) | 0.028304 / 0.128546 (-0.100242) | 0.011496 / 0.075646 (-0.064150) | 0.208236 / 0.419271 (-0.211036) | 0.035720 / 0.043533 (-0.007813) | 0.261190 / 0.255139 (0.006051) | 0.281545 / 0.283200 (-0.001655) | 0.019388 / 0.141683 (-0.122295) | 1.134999 / 1.452155 (-0.317156) | 1.203053 / 1.492716 (-0.289663) |\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) | 0.096007 / 0.018006 (0.078000) | 0.316958 / 0.000490 (0.316469) | 0.000210 / 0.000200 (0.000010) | 0.000053 / 0.000054 (-0.000001) |\n\n### Benchmark: benchmark_indices_mapping.json\n\n| metric | select | shard | shuffle | sort | train_test_split |\n|--------|---|---|---|---|---|\n| new / old (diff) | 0.018330 / 0.037411 (-0.019081) | 0.063299 / 0.014526 (0.048773) | 0.073833 / 0.176557 (-0.102723) | 0.122285 / 0.737135 (-0.614850) | 0.077352 / 0.296338 (-0.218987) |\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) | 0.304487 / 0.215209 (0.089278) | 3.017666 / 2.077655 (0.940012) | 1.664292 / 1.504120 (0.160172) | 1.448446 / 1.541195 (-0.092748) | 1.435612 / 1.468490 (-0.032878) | 0.569704 / 4.584777 (-4.015073) | 2.362015 / 3.745712 (-1.383698) | 2.910380 / 5.269862 (-2.359481) | 1.814560 / 4.565676 (-2.751116) | 0.063986 / 0.424275 (-0.360289) | 0.005022 / 0.007607 (-0.002585) | 0.363528 / 0.226044 (0.137483) | 3.641940 / 2.268929 (1.373011) | 1.961589 / 55.444624 (-53.483035) | 1.603683 / 6.876477 (-5.272793) | 1.663144 / 2.142072 (-0.478928) | 0.645628 / 4.805227 (-4.159599) | 0.118759 / 6.500664 (-6.381905) | 0.042631 / 0.075469 (-0.032838) |\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) | 0.985648 / 1.841788 (-0.856140) | 13.082558 / 8.074308 (5.008250) | 9.909811 / 10.191392 (-0.281581) | 0.131340 / 0.680424 (-0.549083) | 0.013983 / 0.534201 (-0.520218) | 0.289869 / 0.579283 (-0.289414) | 0.271775 / 0.434364 (-0.162589) | 0.334853 / 0.540337 (-0.205485) | 0.457017 / 1.386936 (-0.929919) |\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) | 0.005580 / 0.011353 (-0.005773) | 0.003788 / 0.011008 (-0.007221) | 0.049401 / 0.038508 (0.010893) | 0.030372 / 0.023109 (0.007263) | 0.278554 / 0.275898 (0.002655) | 0.302462 / 0.323480 (-0.021018) | 0.004412 / 0.007986 (-0.003573) | 0.002825 / 0.004328 (-0.001504) | 0.047826 / 0.004250 (0.043576) | 0.047903 / 0.037052 (0.010851) | 0.293098 / 0.258489 (0.034609) | 0.322777 / 0.293841 (0.028936) | 0.030010 / 0.128546 (-0.098536) | 0.011187 / 0.075646 (-0.064459) | 0.057639 / 0.419271 (-0.361632) | 0.059693 / 0.043533 (0.016160) | 0.280288 / 0.255139 (0.025149) | 0.294022 / 0.283200 (0.010823) | 0.019635 / 0.141683 (-0.122048) | 1.154733 / 1.452155 (-0.297422) | 1.200808 / 1.492716 (-0.291908) |\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) | 0.099682 / 0.018006 (0.081676) | 0.319521 / 0.000490 (0.319031) | 0.000224 / 0.000200 (0.000024) | 0.000053 / 0.000054 (-0.000001) |\n\n### Benchmark: benchmark_indices_mapping.json\n\n| metric | select | shard | shuffle | sort | train_test_split |\n|--------|---|---|---|---|---|\n| new / old (diff) | 0.022042 / 0.037411 (-0.015370) | 0.078842 / 0.014526 (0.064317) | 0.088715 / 0.176557 (-0.087841) | 0.126832 / 0.737135 (-0.610303) | 0.089217 / 0.296338 (-0.207122) |\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) | 0.300099 / 0.215209 (0.084890) | 2.907746 / 2.077655 (0.830092) | 1.619418 / 1.504120 (0.115298) | 1.495693 / 1.541195 (-0.045501) | 1.544956 / 1.468490 (0.076466) | 0.556652 / 4.584777 (-4.028124) | 2.414408 / 3.745712 (-1.331304) | 2.737227 / 5.269862 (-2.532635) | 1.763187 / 4.565676 (-2.802490) | 0.062207 / 0.424275 (-0.362069) | 0.005076 / 0.007607 (-0.002531) | 0.349880 / 0.226044 (0.123836) | 3.425355 / 2.268929 (1.156427) | 1.972094 / 55.444624 (-53.472531) | 1.710650 / 6.876477 (-5.165827) | 1.902218 / 2.142072 (-0.239855) | 0.640699 / 4.805227 (-4.164529) | 0.117879 / 6.500664 (-6.382785) | 0.042412 / 0.075469 (-0.033057) |\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) | 1.030131 / 1.841788 (-0.811656) | 12.750637 / 8.074308 (4.676329) | 10.352636 / 10.191392 (0.161244) | 0.141139 / 0.680424 (-0.539285) | 0.015343 / 0.534201 (-0.518858) | 0.294931 / 0.579283 (-0.284352) | 0.275237 / 0.434364 (-0.159127) | 0.336669 / 0.540337 (-0.203668) | 0.429945 / 1.386936 (-0.956991) |\n\n</details>\n</details>\n\n![](https://cml.dev/watermark.png#9c424fa517a1b8517c89545f979e0c8c7d90c3e3 \"CML watermark\")\n"
] |
6,697
|
Unable to Load Dataset in Kaggle
|
### Describe the bug
Having installed the latest versions of transformers==4.38.1 and datasets==2.17.1 Unable to load the dataset in a kaggle notebook.
Get this Error:
```
---------------------------------------------------------------------------
ValueError Traceback (most recent call last)
Cell In[8], line 3
1 from datasets import load_dataset
----> 3 dataset = load_dataset("llm-blender/mix-instruct")
File /opt/conda/lib/python3.10/site-packages/datasets/load.py:1664, in load_dataset(path, name, data_dir, data_files, split, cache_dir, features, download_config, download_mode, ignore_verifications, keep_in_memory, save_infos, revision, use_auth_token, task, streaming, **config_kwargs)
1661 ignore_verifications = ignore_verifications or save_infos
1663 # Create a dataset builder
-> 1664 builder_instance = load_dataset_builder(
1665 path=path,
1666 name=name,
1667 data_dir=data_dir,
1668 data_files=data_files,
1669 cache_dir=cache_dir,
1670 features=features,
1671 download_config=download_config,
1672 download_mode=download_mode,
1673 revision=revision,
1674 use_auth_token=use_auth_token,
1675 **config_kwargs,
1676 )
1678 # Return iterable dataset in case of streaming
1679 if streaming:
File /opt/conda/lib/python3.10/site-packages/datasets/load.py:1490, in load_dataset_builder(path, name, data_dir, data_files, cache_dir, features, download_config, download_mode, revision, use_auth_token, **config_kwargs)
1488 download_config = download_config.copy() if download_config else DownloadConfig()
1489 download_config.use_auth_token = use_auth_token
-> 1490 dataset_module = dataset_module_factory(
1491 path,
1492 revision=revision,
1493 download_config=download_config,
1494 download_mode=download_mode,
1495 data_dir=data_dir,
1496 data_files=data_files,
1497 )
1499 # Get dataset builder class from the processing script
1500 builder_cls = import_main_class(dataset_module.module_path)
File /opt/conda/lib/python3.10/site-packages/datasets/load.py:1242, in dataset_module_factory(path, revision, download_config, download_mode, force_local_path, dynamic_modules_path, data_dir, data_files, **download_kwargs)
1237 if isinstance(e1, FileNotFoundError):
1238 raise FileNotFoundError(
1239 f"Couldn't find a dataset script at {relative_to_absolute_path(combined_path)} or any data file in the same directory. "
1240 f"Couldn't find '{path}' on the Hugging Face Hub either: {type(e1).__name__}: {e1}"
1241 ) from None
-> 1242 raise e1 from None
1243 else:
1244 raise FileNotFoundError(
1245 f"Couldn't find a dataset script at {relative_to_absolute_path(combined_path)} or any data file in the same directory."
1246 )
File /opt/conda/lib/python3.10/site-packages/datasets/load.py:1230, in dataset_module_factory(path, revision, download_config, download_mode, force_local_path, dynamic_modules_path, data_dir, data_files, **download_kwargs)
1215 return HubDatasetModuleFactoryWithScript(
1216 path,
1217 revision=revision,
(...)
1220 dynamic_modules_path=dynamic_modules_path,
1221 ).get_module()
1222 else:
1223 return HubDatasetModuleFactoryWithoutScript(
1224 path,
1225 revision=revision,
1226 data_dir=data_dir,
1227 data_files=data_files,
1228 download_config=download_config,
1229 download_mode=download_mode,
-> 1230 ).get_module()
1231 except Exception as e1: # noqa: all the attempts failed, before raising the error we should check if the module is already cached.
1232 try:
File /opt/conda/lib/python3.10/site-packages/datasets/load.py:846, in HubDatasetModuleFactoryWithoutScript.get_module(self)
836 token = self.download_config.use_auth_token
837 hfh_dataset_info = HfApi(config.HF_ENDPOINT).dataset_info(
838 self.name,
839 revision=self.revision,
840 token=token,
841 timeout=100.0,
842 )
843 patterns = (
844 sanitize_patterns(self.data_files)
845 if self.data_files is not None
--> 846 else get_patterns_in_dataset_repository(hfh_dataset_info)
847 )
848 data_files = DataFilesDict.from_hf_repo(
849 patterns,
850 dataset_info=hfh_dataset_info,
851 allowed_extensions=ALL_ALLOWED_EXTENSIONS,
852 )
853 infered_module_names = {
854 key: infer_module_for_data_files(data_files_list, use_auth_token=self.download_config.use_auth_token)
855 for key, data_files_list in data_files.items()
856 }
File /opt/conda/lib/python3.10/site-packages/datasets/data_files.py:471, in get_patterns_in_dataset_repository(dataset_info)
469 resolver = partial(_resolve_single_pattern_in_dataset_repository, dataset_info)
470 try:
--> 471 return _get_data_files_patterns(resolver)
472 except FileNotFoundError:
473 raise FileNotFoundError(
474 f"The dataset repository at '{dataset_info.id}' doesn't contain any data file."
475 ) from None
File /opt/conda/lib/python3.10/site-packages/datasets/data_files.py:99, in _get_data_files_patterns(pattern_resolver)
97 try:
98 for pattern in patterns:
---> 99 data_files = pattern_resolver(pattern)
100 if len(data_files) > 0:
101 non_empty_splits.append(split)
File /opt/conda/lib/python3.10/site-packages/datasets/data_files.py:303, in _resolve_single_pattern_in_dataset_repository(dataset_info, pattern, allowed_extensions)
301 data_files_ignore = FILES_TO_IGNORE
302 fs = HfFileSystem(repo_info=dataset_info)
--> 303 glob_iter = [PurePath(filepath) for filepath in fs.glob(PurePath(pattern).as_posix()) if fs.isfile(filepath)]
304 matched_paths = [
305 filepath
306 for filepath in glob_iter
307 if filepath.name not in data_files_ignore and not filepath.name.startswith(".")
308 ]
309 if allowed_extensions is not None:
File /opt/conda/lib/python3.10/site-packages/fsspec/spec.py:606, in AbstractFileSystem.glob(self, path, maxdepth, **kwargs)
602 depth = None
604 allpaths = self.find(root, maxdepth=depth, withdirs=True, detail=True, **kwargs)
--> 606 pattern = glob_translate(path + ("/" if ends_with_sep else ""))
607 pattern = re.compile(pattern)
609 out = {
610 p: info
611 for p, info in sorted(allpaths.items())
(...)
618 )
619 }
File /opt/conda/lib/python3.10/site-packages/fsspec/utils.py:734, in glob_translate(pat)
732 continue
733 elif "**" in part:
--> 734 raise ValueError(
735 "Invalid pattern: '**' can only be an entire path component"
736 )
737 if part:
738 results.extend(_translate(part, f"{not_sep}*", not_sep))
ValueError: Invalid pattern: '**' can only be an entire path component
```
```
After loading this dataset
### Steps to reproduce the bug
```
from datasets import load_dataset
dataset = load_dataset("llm-blender/mix-instruct")
```
### Expected behavior
The dataset should load with desired split.
### Environment info
- `datasets` version: 2.17.1
- Platform: Linux-5.15.133+-x86_64-with-glibc2.31
- Python version: 3.10.13
- `huggingface_hub` version: 0.20.3
- PyArrow version: 15.0.0
- Pandas version: 2.2.0
- `fsspec` version: 2023.10.0
|
[] |
{
"url": null,
"html_url": null,
"diff_url": null,
"patch_url": null,
"merged_at": null
}
| false
|
[
"FWIW, I run `load_dataset(\"llm-blender/mix-instruct\")` and it ran successfully.\r\nCan you clear your cache and try again?\r\n\r\n\r\n### Environment Info\r\n\r\n- `datasets` version: 2.17.0\r\n- Platform: Linux-6.2.6-76060206-generic-x86_64-with-glibc2.35\r\n- Python version: 3.9.13\r\n- `huggingface_hub` version: 0.20.3\r\n- PyArrow version: 15.0.0\r\n- Pandas version: 1.5.3\r\n- `fsspec` version: 2023.10.0",
"It is working on the Kaggle GPU instance but gives this same error when running on the CPU instance. Still to run it on Kaggle you require to install the latest versions of datasets and transformers.",
"This error means that `fsspec>=2023.12.0` is installed, which is incompatible with the current releases (the next `datasets` release will be the first to support it). In the meantime, downgrading `fsspec` (`pip install fsspec<=2023.12.0`) should fix the issue.",
"@mariosasko Thanks I got it to work with installing that version of fsspec."
] |
6,696
|
Make JSON builder support an array of strings
|
Support JSON file with an array of strings.
Fix #6695.
|
[] |
{
"url": "https://api.github.com/repos/huggingface/datasets/pulls/6696",
"html_url": "https://github.com/huggingface/datasets/pull/6696",
"diff_url": "https://github.com/huggingface/datasets/pull/6696.diff",
"patch_url": "https://github.com/huggingface/datasets/pull/6696.patch",
"merged_at": "2024-02-28T06:39:12"
}
| true
|
[
"The docs for this PR live [here](https://moon-ci-docs.huggingface.co/docs/datasets/pr_6696). All of your documentation changes will be reflected on that endpoint. The docs are available until 30 days after the last update.",
"<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) | 0.005057 / 0.011353 (-0.006296) | 0.003665 / 0.011008 (-0.007343) | 0.063217 / 0.038508 (0.024709) | 0.028789 / 0.023109 (0.005679) | 0.233597 / 0.275898 (-0.042301) | 0.254792 / 0.323480 (-0.068687) | 0.003065 / 0.007986 (-0.004921) | 0.002686 / 0.004328 (-0.001642) | 0.050182 / 0.004250 (0.045932) | 0.042204 / 0.037052 (0.005151) | 0.254262 / 0.258489 (-0.004227) | 0.277099 / 0.293841 (-0.016742) | 0.027564 / 0.128546 (-0.100982) | 0.010768 / 0.075646 (-0.064878) | 0.207302 / 0.419271 (-0.211969) | 0.035737 / 0.043533 (-0.007796) | 0.242388 / 0.255139 (-0.012751) | 0.259833 / 0.283200 (-0.023367) | 0.019833 / 0.141683 (-0.121850) | 1.135928 / 1.452155 (-0.316227) | 1.162851 / 1.492716 (-0.329865) |\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) | 0.089209 / 0.018006 (0.071202) | 0.300493 / 0.000490 (0.300003) | 0.000216 / 0.000200 (0.000016) | 0.000049 / 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) | 0.017968 / 0.037411 (-0.019444) | 0.061773 / 0.014526 (0.047247) | 0.073835 / 0.176557 (-0.102722) | 0.118592 / 0.737135 (-0.618544) | 0.073606 / 0.296338 (-0.222732) |\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) | 0.287858 / 0.215209 (0.072649) | 2.822917 / 2.077655 (0.745262) | 1.485259 / 1.504120 (-0.018861) | 1.355922 / 1.541195 (-0.185273) | 1.364008 / 1.468490 (-0.104482) | 0.557713 / 4.584777 (-4.027064) | 2.378972 / 3.745712 (-1.366741) | 2.737218 / 5.269862 (-2.532643) | 1.718317 / 4.565676 (-2.847359) | 0.062362 / 0.424275 (-0.361913) | 0.004992 / 0.007607 (-0.002615) | 0.350765 / 0.226044 (0.124721) | 3.387579 / 2.268929 (1.118650) | 1.860408 / 55.444624 (-53.584216) | 1.569355 / 6.876477 (-5.307122) | 1.593013 / 2.142072 (-0.549059) | 0.639325 / 4.805227 (-4.165902) | 0.121769 / 6.500664 (-6.378895) | 0.042148 / 0.075469 (-0.033322) |\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) | 0.990594 / 1.841788 (-0.851194) | 11.460904 / 8.074308 (3.386596) | 9.438691 / 10.191392 (-0.752701) | 0.141884 / 0.680424 (-0.538540) | 0.013725 / 0.534201 (-0.520476) | 0.288847 / 0.579283 (-0.290436) | 0.278815 / 0.434364 (-0.155549) | 0.337108 / 0.540337 (-0.203229) | 0.441659 / 1.386936 (-0.945277) |\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) | 0.005265 / 0.011353 (-0.006088) | 0.003734 / 0.011008 (-0.007274) | 0.049365 / 0.038508 (0.010857) | 0.030483 / 0.023109 (0.007373) | 0.275085 / 0.275898 (-0.000813) | 0.296004 / 0.323480 (-0.027475) | 0.004964 / 0.007986 (-0.003022) | 0.002542 / 0.004328 (-0.001787) | 0.048734 / 0.004250 (0.044483) | 0.044098 / 0.037052 (0.007046) | 0.292517 / 0.258489 (0.034028) | 0.319992 / 0.293841 (0.026151) | 0.029552 / 0.128546 (-0.098994) | 0.010669 / 0.075646 (-0.064977) | 0.058887 / 0.419271 (-0.360385) | 0.051163 / 0.043533 (0.007630) | 0.277266 / 0.255139 (0.022127) | 0.295347 / 0.283200 (0.012147) | 0.018403 / 0.141683 (-0.123280) | 1.151979 / 1.452155 (-0.300176) | 1.204583 / 1.492716 (-0.288134) |\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) | 0.091157 / 0.018006 (0.073151) | 0.300109 / 0.000490 (0.299619) | 0.000211 / 0.000200 (0.000011) | 0.000052 / 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) | 0.021521 / 0.037411 (-0.015890) | 0.074954 / 0.014526 (0.060428) | 0.087010 / 0.176557 (-0.089546) | 0.125853 / 0.737135 (-0.611282) | 0.087877 / 0.296338 (-0.208461) |\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) | 0.297890 / 0.215209 (0.082681) | 2.912159 / 2.077655 (0.834504) | 1.619311 / 1.504120 (0.115192) | 1.501726 / 1.541195 (-0.039468) | 1.494143 / 1.468490 (0.025652) | 0.566744 / 4.584777 (-4.018033) | 2.497594 / 3.745712 (-1.248118) | 2.631403 / 5.269862 (-2.638459) | 1.727896 / 4.565676 (-2.837780) | 0.065937 / 0.424275 (-0.358339) | 0.005023 / 0.007607 (-0.002585) | 0.345747 / 0.226044 (0.119702) | 3.417615 / 2.268929 (1.148686) | 1.949970 / 55.444624 (-53.494654) | 1.680019 / 6.876477 (-5.196457) | 1.789879 / 2.142072 (-0.352193) | 0.648053 / 4.805227 (-4.157174) | 0.117408 / 6.500664 (-6.383256) | 0.040681 / 0.075469 (-0.034788) |\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) | 1.012535 / 1.841788 (-0.829252) | 11.935819 / 8.074308 (3.861511) | 10.241452 / 10.191392 (0.050060) | 0.130956 / 0.680424 (-0.549468) | 0.015396 / 0.534201 (-0.518805) | 0.289166 / 0.579283 (-0.290117) | 0.274149 / 0.434364 (-0.160215) | 0.325844 / 0.540337 (-0.214493) | 0.424919 / 1.386936 (-0.962017) |\n\n</details>\n</details>\n\n![](https://cml.dev/watermark.png#cb834d9c63ab8cb14725ae8e4fc2da8672892a6d \"CML watermark\")\n"
] |
6,695
|
Support JSON file with an array of strings
|
Support loading a dataset from a JSON file with an array of strings.
See: https://huggingface.co/datasets/CausalLM/Refined-Anime-Text/discussions/1
|
[
{
"id": 1935892871,
"node_id": "MDU6TGFiZWwxOTM1ODkyODcx",
"url": "https://api.github.com/repos/huggingface/datasets/labels/enhancement",
"name": "enhancement",
"color": "a2eeef",
"default": true,
"description": "New feature or request"
}
] |
{
"url": null,
"html_url": null,
"diff_url": null,
"patch_url": null,
"merged_at": null
}
| false
|
[
"https://huggingface.co/datasets/CausalLM/Refined-Anime-Text/discussions/1 has been fixed, but how can we check if there are other datasets with the same error, in datasets-server's database? I don't know how to get the list of erroneous cache entries, since we only copied `Error code: JobManagerCrashedError`, but not the traceback in `details`... Do you remember the error message, or the underlying exception, we had?"
] |
6,694
|
__add__ for Dataset, IterableDataset
|
It's too cumbersome to write this command every time we perform a dataset merging operation. ```pythonfrom datasets import concatenate_datasets``` We have added a simple `__add__` magic method to each class using `concatenate_datasets.`
```python
from datasets import load_dataset
bookcorpus = load_dataset("bookcorpus", split="train")
wiki = load_dataset("wikimedia/wikipedia", "20231101.ab", split="train")
wiki = wiki.remove_columns([col for col in wiki.column_names if col != "text"]) # only keep the 'text' column
bookcorpus + wiki
#Dataset({
# features: ['text'],
# num_rows: 74004228
#})
#Dataset({
# features: ['text'],
# num_rows: 6152
#})
#Dataset({
# features: ['text'],
# num_rows: 74010380
#})
```
|
[] |
{
"url": "https://api.github.com/repos/huggingface/datasets/pulls/6694",
"html_url": "https://github.com/huggingface/datasets/pull/6694",
"diff_url": "https://github.com/huggingface/datasets/pull/6694.diff",
"patch_url": "https://github.com/huggingface/datasets/pull/6694.patch",
"merged_at": null
}
| true
|
[
"Hi! You can find a reason why we are against this feature in https://github.com/huggingface/datasets/issues/3449. \r\n\r\n> It's too cumbersome to write this command every time we perform a dataset merging operation\r\n\r\nExplicit is better than implicit, so this isn't a good enough reason. \r\n\r\nThanks for the effort nonetheless :)!"
] |
6,693
|
Update the print message for chunked_dataset in process.mdx
|
Update documentation to align with `Dataset.__repr__` change after #423
|
[] |
{
"url": "https://api.github.com/repos/huggingface/datasets/pulls/6693",
"html_url": "https://github.com/huggingface/datasets/pull/6693",
"diff_url": "https://github.com/huggingface/datasets/pull/6693.diff",
"patch_url": "https://github.com/huggingface/datasets/pull/6693.patch",
"merged_at": "2024-02-25T19:51:02"
}
| true
|
[
"The docs for this PR live [here](https://moon-ci-docs.huggingface.co/docs/datasets/pr_6693). All of your documentation changes will be reflected on that endpoint. The docs are available until 30 days after the last update.",
"<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) | 0.005069 / 0.011353 (-0.006284) | 0.003682 / 0.011008 (-0.007326) | 0.063733 / 0.038508 (0.025225) | 0.030377 / 0.023109 (0.007268) | 0.242962 / 0.275898 (-0.032936) | 0.262865 / 0.323480 (-0.060615) | 0.004760 / 0.007986 (-0.003225) | 0.002772 / 0.004328 (-0.001557) | 0.049094 / 0.004250 (0.044843) | 0.041093 / 0.037052 (0.004041) | 0.260423 / 0.258489 (0.001934) | 0.283908 / 0.293841 (-0.009933) | 0.027409 / 0.128546 (-0.101138) | 0.010548 / 0.075646 (-0.065098) | 0.208637 / 0.419271 (-0.210634) | 0.035386 / 0.043533 (-0.008147) | 0.242352 / 0.255139 (-0.012787) | 0.264201 / 0.283200 (-0.018999) | 0.017822 / 0.141683 (-0.123860) | 1.140792 / 1.452155 (-0.311363) | 1.166782 / 1.492716 (-0.325934) |\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) | 0.094727 / 0.018006 (0.076720) | 0.308548 / 0.000490 (0.308059) | 0.000213 / 0.000200 (0.000013) | 0.000052 / 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) | 0.018106 / 0.037411 (-0.019305) | 0.062057 / 0.014526 (0.047531) | 0.073821 / 0.176557 (-0.102735) | 0.121269 / 0.737135 (-0.615867) | 0.074062 / 0.296338 (-0.222277) |\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) | 0.282978 / 0.215209 (0.067768) | 2.788626 / 2.077655 (0.710971) | 1.479756 / 1.504120 (-0.024364) | 1.360620 / 1.541195 (-0.180575) | 1.363996 / 1.468490 (-0.104494) | 0.571646 / 4.584777 (-4.013131) | 2.430630 / 3.745712 (-1.315083) | 2.783909 / 5.269862 (-2.485953) | 1.744617 / 4.565676 (-2.821060) | 0.062771 / 0.424275 (-0.361504) | 0.004978 / 0.007607 (-0.002629) | 0.347929 / 0.226044 (0.121884) | 3.368837 / 2.268929 (1.099908) | 1.855635 / 55.444624 (-53.588990) | 1.581555 / 6.876477 (-5.294922) | 1.589888 / 2.142072 (-0.552184) | 0.655821 / 4.805227 (-4.149406) | 0.118990 / 6.500664 (-6.381674) | 0.042191 / 0.075469 (-0.033278) |\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) | 0.991099 / 1.841788 (-0.850688) | 11.627919 / 8.074308 (3.553611) | 9.554180 / 10.191392 (-0.637212) | 0.140541 / 0.680424 (-0.539882) | 0.014264 / 0.534201 (-0.519937) | 0.288465 / 0.579283 (-0.290818) | 0.266400 / 0.434364 (-0.167964) | 0.324400 / 0.540337 (-0.215938) | 0.423158 / 1.386936 (-0.963778) |\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) | 0.005588 / 0.011353 (-0.005765) | 0.003784 / 0.011008 (-0.007224) | 0.049961 / 0.038508 (0.011453) | 0.031215 / 0.023109 (0.008105) | 0.280859 / 0.275898 (0.004961) | 0.306416 / 0.323480 (-0.017063) | 0.004310 / 0.007986 (-0.003676) | 0.002884 / 0.004328 (-0.001445) | 0.049662 / 0.004250 (0.045412) | 0.046611 / 0.037052 (0.009559) | 0.293353 / 0.258489 (0.034864) | 0.327839 / 0.293841 (0.033998) | 0.050784 / 0.128546 (-0.077763) | 0.010890 / 0.075646 (-0.064757) | 0.059612 / 0.419271 (-0.359659) | 0.033175 / 0.043533 (-0.010358) | 0.281085 / 0.255139 (0.025946) | 0.302746 / 0.283200 (0.019547) | 0.019201 / 0.141683 (-0.122481) | 1.126722 / 1.452155 (-0.325433) | 1.225678 / 1.492716 (-0.267038) |\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) | 0.094335 / 0.018006 (0.076329) | 0.304774 / 0.000490 (0.304285) | 0.000207 / 0.000200 (0.000007) | 0.000050 / 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) | 0.021648 / 0.037411 (-0.015763) | 0.077920 / 0.014526 (0.063394) | 0.087125 / 0.176557 (-0.089432) | 0.125481 / 0.737135 (-0.611654) | 0.089415 / 0.296338 (-0.206924) |\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) | 0.304955 / 0.215209 (0.089746) | 2.992587 / 2.077655 (0.914932) | 1.654609 / 1.504120 (0.150490) | 1.509114 / 1.541195 (-0.032081) | 1.530906 / 1.468490 (0.062416) | 0.572092 / 4.584777 (-4.012685) | 2.477902 / 3.745712 (-1.267810) | 2.731363 / 5.269862 (-2.538498) | 1.750000 / 4.565676 (-2.815677) | 0.063662 / 0.424275 (-0.360613) | 0.005008 / 0.007607 (-0.002600) | 0.353066 / 0.226044 (0.127022) | 3.528309 / 2.268929 (1.259380) | 2.009238 / 55.444624 (-53.435387) | 1.717792 / 6.876477 (-5.158685) | 1.861699 / 2.142072 (-0.280373) | 0.667392 / 4.805227 (-4.137835) | 0.119197 / 6.500664 (-6.381467) | 0.041131 / 0.075469 (-0.034338) |\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) | 1.032182 / 1.841788 (-0.809605) | 12.042613 / 8.074308 (3.968305) | 10.256293 / 10.191392 (0.064901) | 0.141180 / 0.680424 (-0.539244) | 0.015005 / 0.534201 (-0.519196) | 0.290081 / 0.579283 (-0.289202) | 0.281081 / 0.434364 (-0.153283) | 0.331425 / 0.540337 (-0.208912) | 0.418674 / 1.386936 (-0.968262) |\n\n</details>\n</details>\n\n![](https://cml.dev/watermark.png#ad5b221c01a183a66cbf52a6d708f94e0cff0b53 \"CML watermark\")\n"
] |
6,692
|
Enhancement: Enable loading TSV files in load_dataset()
|
Fix #6691
|
[] |
{
"url": "https://api.github.com/repos/huggingface/datasets/pulls/6692",
"html_url": "https://github.com/huggingface/datasets/pull/6692",
"diff_url": "https://github.com/huggingface/datasets/pull/6692.diff",
"patch_url": "https://github.com/huggingface/datasets/pull/6692.patch",
"merged_at": null
}
| true
|
[
"Hi @harsh1504660,\r\n\r\nThanks for your work, but this functionality already exists. See my comment in the corresponding issue: https://github.com/huggingface/datasets/issues/6691#issuecomment-1963449923\r\n\r\nNext time you would like to contribute, I would suggest you take on an issue that is previously validated by one of the maintainers. Thanks anyway."
] |
6,691
|
load_dataset() does not support tsv
|
### Feature request
the load_dataset() for local functions support file types like csv, json etc but not of type tsv (tab separated values).
### Motivation
cant easily load files of type tsv, have to convert them to another type like csv then load
### Your contribution
Can try by raising a PR with a little help, currently went through the code but didn't fully understand
|
[
{
"id": 1935892871,
"node_id": "MDU6TGFiZWwxOTM1ODkyODcx",
"url": "https://api.github.com/repos/huggingface/datasets/labels/enhancement",
"name": "enhancement",
"color": "a2eeef",
"default": true,
"description": "New feature or request"
}
] |
{
"url": null,
"html_url": null,
"diff_url": null,
"patch_url": null,
"merged_at": null
}
| false
|
[
"#self-assign",
"Hi @dipsivenkatesh,\r\n\r\nPlease note that this functionality is already implemented. Our CSV builder uses `pandas.read_csv` under the hood, and you can pass the parameter `delimiter=\"\\t\"` to read TSV files.\r\n\r\nSee the list of CSV config parameters in our docs: https://huggingface.co/docs/datasets/package_reference/loading_methods#datasets.packaged_modules.csv.CsvConfig"
] |
6,690
|
Add function to convert a script-dataset to Parquet
|
Add function to convert a script-dataset to Parquet and push it to the Hub, analogously to the Space: "Convert a Hugging Face dataset to Parquet"
|
[
{
"id": 1935892871,
"node_id": "MDU6TGFiZWwxOTM1ODkyODcx",
"url": "https://api.github.com/repos/huggingface/datasets/labels/enhancement",
"name": "enhancement",
"color": "a2eeef",
"default": true,
"description": "New feature or request"
}
] |
{
"url": null,
"html_url": null,
"diff_url": null,
"patch_url": null,
"merged_at": null
}
| false
|
[] |
6,689
|
.load_dataset() method defaults to zstandard
|
### Describe the bug
Regardless of what method I use, datasets defaults to zstandard for unpacking my datasets.
This is poor behavior, because not only is zstandard not a dependency in the huggingface package (and therefore, your dataset loading will be interrupted while it asks you to install the package), but it happens on datasets that are uploaded in json format too, meaning the dataset loader will attempt to convert the data to a zstandard compatible format, and THEN try to unpackage it.
My 4tb drive runs out of room when using zstandard on slimpajama. It loads fine on 1.5tb when using json, however I lack the understanding of the "magic numbers" system used to select the unpackaging algorithm, so I can't push a change myself.
Commenting out this line, in "/datasets/utils/extract.py" fixes the issue, and causes SlimPajama to properly extract using rational amounts of storage, however it completely disables zstandard, which is probably undesirable behavior. Someone with an understanding of the "magic numbers" system should probably take a pass over this issue.
```
class Extractor:
# Put zip file to the last, b/c it is possible wrongly detected as zip (I guess it means: as tar or gzip)
extractors: Dict[str, Type[BaseExtractor]] = {
"tar": TarExtractor,
"gzip": GzipExtractor,
"zip": ZipExtractor,
"xz": XzExtractor,
#"zstd": ZstdExtractor, # This line needs to go, in order for datasets to work w/o non-dependent packages
"rar": RarExtractor,
"bz2": Bzip2Extractor,
"7z": SevenZipExtractor, # <Added version="2.4.0"/>
"lz4": Lz4Extractor, # <Added version="2.4.0"/>
}
```
### Steps to reproduce the bug
'''
from datasaets import load_dataset
load_dataset(path="/cerebras/SlimPajama-627B")
'''
This alone should trigger the error on any system that does not have zstandard pip installed.
### Expected behavior
This repository (which is encoded in json format, not zstandard) should check whether zstandard is installed before defaulting to it. Additionally, using zstandard should not use more than 3x the required space that other extraction mechanisms use.
### Environment info
- `datasets` version: 2.17.1
- Platform: Linux-6.5.0-18-generic-x86_64-with-glibc2.35
- Python version: 3.12.0
- `huggingface_hub` version: 0.20.3
- PyArrow version: 15.0.0
- Pandas version: 2.2.0
- `fsspec` version: 2023.10.0
|
[] |
{
"url": null,
"html_url": null,
"diff_url": null,
"patch_url": null,
"merged_at": null
}
| false
|
[
"The dataset is made of JSON files compressed using zstandard, as you can see here: https://huggingface.co/datasets/cerebras/SlimPajama-627B/tree/main/test/chunk1\r\n\r\nThat's why it asks for zstandard to be installed.\r\n\r\nThough I'm intrigued that you manage to load the dataset without zstandard installed. Maybe `pyarrow` that we use to load JSON data under the hood got support for zstandard at one point.",
"> The dataset is made of JSON files compressed using zstandard, as you can see here: https://huggingface.co/datasets/cerebras/SlimPajama-627B/tree/main/test/chunk1\r\n> \r\n> That's why it asks for zstandard to be installed.\r\n> \r\n> Though I'm intrigued that you manage to load the dataset without zstandard installed. Maybe `pyarrow` that we use to load JSON data under the hood got support for zstandard at one point.\r\n\r\nQuestion, then.\r\n\r\nWhen I loaded this dataset back in October, it downloaded all the files, and then loaded into memory just fine.\r\n\r\nNOW, it has to sit there and unpack all these zstd files (3.6TB worth). Further, when they're in my harddrive, they're regular json files. It's only when looking at the LFS, or when the loading script runs, that I get asked to install zstd.\r\n\r\nMy question is, **is this normal?** As far as I can tell, there's no reason the dataset or the loading methods should have changed between then and now. Was my old behavior flawed, and the new behavior correct?\r\n\r\nI mean, I got it working eventually, but it was pulling teeth, and it still doesn't load right, as I had to unpack each chunk separately, so there's no clean mapping between the chunks and the broader dataset.",
"The `ZstdExtractor` has been added 3 years ago and we haven't touched it since then. Same for the JSON loader.\r\n\r\n`zstandard` is required as soon as you try to load a file with the `.zstd` extension or if a file starts with the Zstandard magic number `b\"\\x28\\xb5\\x2f\\xfd\"` (used to recognize Zstandard files).\r\n\r\nNote that the extraction only has to happen once - if you reload the dataset it will be reloaded from your cache directly.\r\n\r\nNot sure what happened between October and now unfortunately",
"Understood, thank you for clarifying that for me.\r\n\r\nI'll look into how best to collate my stack of batches w/o creating duplicate arrow tables for each one."
] |
6,688
|
Tensor type (e.g. from `return_tensors`) ignored in map
|
### Describe the bug
I don't know if it is a bug or an expected behavior, but the tensor type seems to be ignored after applying map. For example, mapping over to tokenize text with a transformers' tokenizer always returns lists and it ignore the `return_tensors` argument.
If this is an expected behaviour (e.g., for caching/Arrow compatibility/etc.) it should be clearly documented. For example, current documentation (see [here](https://huggingface.co/docs/datasets/v2.17.1/en/nlp_process#map)) clearly state to "set `return_tensors="np"` when you tokenize your text" to have Numpy arrays.
### Steps to reproduce the bug
```py
# %%%
import datasets
import numpy as np
import tensorflow as tf
import torch
from transformers import AutoTokenizer
# %%
ds = datasets.load_dataset("cnn_dailymail", "1.0.0", split="train[:1%]")
tokenizer = AutoTokenizer.from_pretrained("bert-base-cased")
#%%
for return_tensors in [None, "np", "pt", "tf", "jax"]:
print(f"********** no map, return_tensors={return_tensors} **********")
_ds = tokenizer(ds["article"], return_tensors=return_tensors, truncation=True, padding=True)
print('Type <input_ids>:', type(_ds["input_ids"]))
# %%
for return_tensors in [None, "np", "pt", "tf", "jax"]:
print(f"********** map, return_tensors={return_tensors} **********")
_ds = ds.map(
lambda examples: tokenizer(examples["article"], return_tensors=return_tensors, truncation=True, padding=True),
batched=True,
remove_columns=["article"],
)
print('Type <input_ids>:', type(_ds[0]["input_ids"]))
```
### Expected behavior
The output from the script above. I would expect the second half to be the same.
```
********** no map, return_tensors=None **********
Type <input_ids>: <class 'list'>
********** no map, return_tensors=np **********
Type <input_ids>: <class 'numpy.ndarray'>
********** no map, return_tensors=pt **********
Type <input_ids>: <class 'torch.Tensor'>
********** no map, return_tensors=tf **********
Type <input_ids>: <class 'tensorflow.python.framework.ops.EagerTensor'>
********** no map, return_tensors=jax **********
Type <input_ids>: <class 'jaxlib.xla_extension.ArrayImpl'>
********** map, return_tensors=None **********
Type <input_ids>: <class 'list'>
********** map, return_tensors=np **********
Type <input_ids>: <class 'list'>
********** map, return_tensors=pt **********
Type <input_ids>: <class 'list'>
********** map, return_tensors=tf **********
Type <input_ids>: <class 'list'>
********** map, return_tensors=jax **********
Type <input_ids>: <class 'list'>
```
### Environment info
- `datasets` version: 2.17.1
- Platform: Redacted (linux)
- Python version: 3.10.12
- `huggingface_hub` version: 0.20.3
- PyArrow version: 15.0.0
- Pandas version: 2.1.3
- `fsspec` version: 2023.10.0
|
[] |
{
"url": null,
"html_url": null,
"diff_url": null,
"patch_url": null,
"merged_at": null
}
| false
|
[
"Hi, this is expected behavior since all the tensors are converted to Arrow data (the storage type behind a Dataset).\r\n\r\nTo get pytorch tensors back, you can set the dataset format to \"torch\":\r\n\r\n```python\r\nds = ds.with_format(\"torch\")\r\n```",
"Thanks. Just one additional question. During the pipeline `<framework> -> arrow -> <framework>`, does `.with_format` zero-copies the tensors or is it a deep copy? And is this behavior framework-dependent?\r\n\r\nThanks again.",
"We do zero-copy Arrow <-> NumPy <-> PyTorch when the output dtype matches the original dtype, but for other frameworks it depends. For example JAX doesn't allow zero-copy NumPy -> JAX at all IIRC.\r\n\r\nCurrently tokenized data are formatted using a copy though, since tokens are stored as int32 and returned as int64 torch tensors."
] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.