html_url stringlengths 48 51 | title stringlengths 5 280 | comments stringlengths 63 51.8k | body stringlengths 0 36.2k ⌀ | comment_length int64 16 1.52k | text stringlengths 159 54.1k | embeddings listlengths 768 768 |
|---|---|---|---|---|---|---|
https://github.com/huggingface/datasets/issues/373 | Segmentation fault when loading local JSON dataset as of #372 | I've seen this sort of thing before -- it might help to delete the directory -- I've also noticed that there is an error with the json Dataloader for any data I've tried to load. I've replaced it with this, which skips over the data feature population step:
```python
import os
import pyarrow.json as paj
imp... | The last issue was closed (#369) once the #372 update was merged. However, I'm still not able to load a SQuAD formatted JSON file. Instead of the previously recorded pyarrow error, I now get a segmentation fault.
```
dataset = nlp.load_dataset('json', data_files={nlp.Split.TRAIN: ["./datasets/train-v2.0.json"]}, f... | 191 | Segmentation fault when loading local JSON dataset as of #372
The last issue was closed (#369) once the #372 update was merged. However, I'm still not able to load a SQuAD formatted JSON file. Instead of the previously recorded pyarrow error, I now get a segmentation fault.
```
dataset = nlp.load_dataset('json',... | [
-0.017292045056819916,
0.289395809173584,
0.06094152852892876,
0.05787059664726257,
0.22360973060131073,
-0.016548238694667816,
0.3200393617153168,
0.6101838946342468,
-0.40100961923599243,
-0.21304242312908173,
-0.2422536462545395,
0.7714078426361084,
0.1327875852584839,
-0.42659667134284... |
https://github.com/huggingface/datasets/issues/373 | Segmentation fault when loading local JSON dataset as of #372 | Yes, deleting the directory solves the error whenever I try to rerun.
By replacing the json-loader, you mean the cached file in my `site-packages` directory? e.g. `/home/XXX/.cache/lib/python3.7/site-packages/nlp/datasets/json/(...)/json.py`
When I was testing this out before the #372 PR was merged I had issues ... | The last issue was closed (#369) once the #372 update was merged. However, I'm still not able to load a SQuAD formatted JSON file. Instead of the previously recorded pyarrow error, I now get a segmentation fault.
```
dataset = nlp.load_dataset('json', data_files={nlp.Split.TRAIN: ["./datasets/train-v2.0.json"]}, f... | 96 | Segmentation fault when loading local JSON dataset as of #372
The last issue was closed (#369) once the #372 update was merged. However, I'm still not able to load a SQuAD formatted JSON file. Instead of the previously recorded pyarrow error, I now get a segmentation fault.
```
dataset = nlp.load_dataset('json',... | [
-0.017292045056819916,
0.289395809173584,
0.06094152852892876,
0.05787059664726257,
0.22360973060131073,
-0.016548238694667816,
0.3200393617153168,
0.6101838946342468,
-0.40100961923599243,
-0.21304242312908173,
-0.2422536462545395,
0.7714078426361084,
0.1327875852584839,
-0.42659667134284... |
https://github.com/huggingface/datasets/issues/373 | Segmentation fault when loading local JSON dataset as of #372 | I see, diving in the JSON file for SQuAD it's a pretty complex structure.
The best solution for you, if you have a dataset really similar to SQuAD would be to copy and modify the SQuAD data processing script. We will probably add soon an option to be able to specify file path to use instead of the automatic URL enco... | The last issue was closed (#369) once the #372 update was merged. However, I'm still not able to load a SQuAD formatted JSON file. Instead of the previously recorded pyarrow error, I now get a segmentation fault.
```
dataset = nlp.load_dataset('json', data_files={nlp.Split.TRAIN: ["./datasets/train-v2.0.json"]}, f... | 117 | Segmentation fault when loading local JSON dataset as of #372
The last issue was closed (#369) once the #372 update was merged. However, I'm still not able to load a SQuAD formatted JSON file. Instead of the previously recorded pyarrow error, I now get a segmentation fault.
```
dataset = nlp.load_dataset('json',... | [
-0.017292045056819916,
0.289395809173584,
0.06094152852892876,
0.05787059664726257,
0.22360973060131073,
-0.016548238694667816,
0.3200393617153168,
0.6101838946342468,
-0.40100961923599243,
-0.21304242312908173,
-0.2422536462545395,
0.7714078426361084,
0.1327875852584839,
-0.42659667134284... |
https://github.com/huggingface/datasets/issues/373 | Segmentation fault when loading local JSON dataset as of #372 | This seems like a more sensible solution! Thanks, @thomwolf. It's been a little daunting to understand what these scripts actually do, due to the level of abstraction and central documentation.
Am I correct in assuming that the `_generate_examples()` function is the actual procedure for how the data is loaded from f... | The last issue was closed (#369) once the #372 update was merged. However, I'm still not able to load a SQuAD formatted JSON file. Instead of the previously recorded pyarrow error, I now get a segmentation fault.
```
dataset = nlp.load_dataset('json', data_files={nlp.Split.TRAIN: ["./datasets/train-v2.0.json"]}, f... | 156 | Segmentation fault when loading local JSON dataset as of #372
The last issue was closed (#369) once the #372 update was merged. However, I'm still not able to load a SQuAD formatted JSON file. Instead of the previously recorded pyarrow error, I now get a segmentation fault.
```
dataset = nlp.load_dataset('json',... | [
-0.017292045056819916,
0.289395809173584,
0.06094152852892876,
0.05787059664726257,
0.22360973060131073,
-0.016548238694667816,
0.3200393617153168,
0.6101838946342468,
-0.40100961923599243,
-0.21304242312908173,
-0.2422536462545395,
0.7714078426361084,
0.1327875852584839,
-0.42659667134284... |
https://github.com/huggingface/datasets/issues/373 | Segmentation fault when loading local JSON dataset as of #372 | Yes `_generate_examples()` is the main entry point. If you change the shape of the returned dictionary you also need to update the `features` in the `_info`.
I'm currently writing the doc so it should be easier soon to use the library and know how to add your datasets.
| The last issue was closed (#369) once the #372 update was merged. However, I'm still not able to load a SQuAD formatted JSON file. Instead of the previously recorded pyarrow error, I now get a segmentation fault.
```
dataset = nlp.load_dataset('json', data_files={nlp.Split.TRAIN: ["./datasets/train-v2.0.json"]}, f... | 48 | Segmentation fault when loading local JSON dataset as of #372
The last issue was closed (#369) once the #372 update was merged. However, I'm still not able to load a SQuAD formatted JSON file. Instead of the previously recorded pyarrow error, I now get a segmentation fault.
```
dataset = nlp.load_dataset('json',... | [
-0.017292045056819916,
0.289395809173584,
0.06094152852892876,
0.05787059664726257,
0.22360973060131073,
-0.016548238694667816,
0.3200393617153168,
0.6101838946342468,
-0.40100961923599243,
-0.21304242312908173,
-0.2422536462545395,
0.7714078426361084,
0.1327875852584839,
-0.42659667134284... |
https://github.com/huggingface/datasets/issues/373 | Segmentation fault when loading local JSON dataset as of #372 | Could you try to update pyarrow to >=0.17.0 @vegarab ?
I don't have any segmentation fault with my version of pyarrow (0.17.1)
I tested with
```python
import nlp
s = nlp.load_dataset("json", data_files="train-v2.0.json", field="data", split="train")
s[0]
# {'title': 'Normans', 'paragraphs': [{'qas': [{'questio... | The last issue was closed (#369) once the #372 update was merged. However, I'm still not able to load a SQuAD formatted JSON file. Instead of the previously recorded pyarrow error, I now get a segmentation fault.
```
dataset = nlp.load_dataset('json', data_files={nlp.Split.TRAIN: ["./datasets/train-v2.0.json"]}, f... | 49 | Segmentation fault when loading local JSON dataset as of #372
The last issue was closed (#369) once the #372 update was merged. However, I'm still not able to load a SQuAD formatted JSON file. Instead of the previously recorded pyarrow error, I now get a segmentation fault.
```
dataset = nlp.load_dataset('json',... | [
-0.017292045056819916,
0.289395809173584,
0.06094152852892876,
0.05787059664726257,
0.22360973060131073,
-0.016548238694667816,
0.3200393617153168,
0.6101838946342468,
-0.40100961923599243,
-0.21304242312908173,
-0.2422536462545395,
0.7714078426361084,
0.1327875852584839,
-0.42659667134284... |
https://github.com/huggingface/datasets/issues/373 | Segmentation fault when loading local JSON dataset as of #372 | Also if you want to have your own dataset script, we now have a new documentation !
See here:
https://huggingface.co/nlp/add_dataset.html | The last issue was closed (#369) once the #372 update was merged. However, I'm still not able to load a SQuAD formatted JSON file. Instead of the previously recorded pyarrow error, I now get a segmentation fault.
```
dataset = nlp.load_dataset('json', data_files={nlp.Split.TRAIN: ["./datasets/train-v2.0.json"]}, f... | 20 | Segmentation fault when loading local JSON dataset as of #372
The last issue was closed (#369) once the #372 update was merged. However, I'm still not able to load a SQuAD formatted JSON file. Instead of the previously recorded pyarrow error, I now get a segmentation fault.
```
dataset = nlp.load_dataset('json',... | [
-0.017292045056819916,
0.289395809173584,
0.06094152852892876,
0.05787059664726257,
0.22360973060131073,
-0.016548238694667816,
0.3200393617153168,
0.6101838946342468,
-0.40100961923599243,
-0.21304242312908173,
-0.2422536462545395,
0.7714078426361084,
0.1327875852584839,
-0.42659667134284... |
https://github.com/huggingface/datasets/issues/373 | Segmentation fault when loading local JSON dataset as of #372 | @lhoestq
For some reason, I am not able to reproduce the segmentation fault, on pyarrow==0.16.0. Using the exact same environment and file.
Anyhow, I discovered that pyarrow>=0.17.0 is required to read in a JSON file where the pandas structs contain lists. Otherwise, pyarrow complains when attempting to cast the s... | The last issue was closed (#369) once the #372 update was merged. However, I'm still not able to load a SQuAD formatted JSON file. Instead of the previously recorded pyarrow error, I now get a segmentation fault.
```
dataset = nlp.load_dataset('json', data_files={nlp.Split.TRAIN: ["./datasets/train-v2.0.json"]}, f... | 219 | Segmentation fault when loading local JSON dataset as of #372
The last issue was closed (#369) once the #372 update was merged. However, I'm still not able to load a SQuAD formatted JSON file. Instead of the previously recorded pyarrow error, I now get a segmentation fault.
```
dataset = nlp.load_dataset('json',... | [
-0.017292045056819916,
0.289395809173584,
0.06094152852892876,
0.05787059664726257,
0.22360973060131073,
-0.016548238694667816,
0.3200393617153168,
0.6101838946342468,
-0.40100961923599243,
-0.21304242312908173,
-0.2422536462545395,
0.7714078426361084,
0.1327875852584839,
-0.42659667134284... |
https://github.com/huggingface/datasets/issues/369 | can't load local dataset: pyarrow.lib.ArrowInvalid: straddling object straddles two block boundaries | I am able to reproduce this with the official SQuAD `train-v2.0.json` file downloaded directly from https://rajpurkar.github.io/SQuAD-explorer/ | Trying to load a local SQuAD-formatted dataset (from a JSON file, about 60MB):
```
dataset = nlp.load_dataset(path='json', data_files={nlp.Split.TRAIN: ["./path/to/file.json"]})
```
causes
```
Traceback (most recent call last):
File "dataloader.py", line 9, in <module>
["./path/to/file.json"]})
File "/... | 16 | can't load local dataset: pyarrow.lib.ArrowInvalid: straddling object straddles two block boundaries
Trying to load a local SQuAD-formatted dataset (from a JSON file, about 60MB):
```
dataset = nlp.load_dataset(path='json', data_files={nlp.Split.TRAIN: ["./path/to/file.json"]})
```
causes
```
Traceback (most re... | [
-0.22482475638389587,
0.20798179507255554,
0.01309872791171074,
0.137778639793396,
0.2781042158603668,
-0.25971946120262146,
0.31692296266555786,
0.46876823902130127,
-0.1691408008337021,
-0.14901141822338104,
0.002377506345510483,
0.42034125328063965,
0.06907229870557785,
-0.0172274410724... |
https://github.com/huggingface/datasets/issues/369 | can't load local dataset: pyarrow.lib.ArrowInvalid: straddling object straddles two block boundaries | I am facing this issue in transformers library 3.0.2 while reading a csv using datasets.
Is this fixed in latest version?
I updated the latest version 4.0.1 but still getting this error. What could cause this error? | Trying to load a local SQuAD-formatted dataset (from a JSON file, about 60MB):
```
dataset = nlp.load_dataset(path='json', data_files={nlp.Split.TRAIN: ["./path/to/file.json"]})
```
causes
```
Traceback (most recent call last):
File "dataloader.py", line 9, in <module>
["./path/to/file.json"]})
File "/... | 37 | can't load local dataset: pyarrow.lib.ArrowInvalid: straddling object straddles two block boundaries
Trying to load a local SQuAD-formatted dataset (from a JSON file, about 60MB):
```
dataset = nlp.load_dataset(path='json', data_files={nlp.Split.TRAIN: ["./path/to/file.json"]})
```
causes
```
Traceback (most re... | [
-0.22482475638389587,
0.20798179507255554,
0.01309872791171074,
0.137778639793396,
0.2781042158603668,
-0.25971946120262146,
0.31692296266555786,
0.46876823902130127,
-0.1691408008337021,
-0.14901141822338104,
0.002377506345510483,
0.42034125328063965,
0.06907229870557785,
-0.0172274410724... |
https://github.com/huggingface/datasets/issues/368 | load_metric can't acquire lock anymore | I found that, in the same process (or the same interactive session), if I do
import nlp
m1 = nlp.load_metric('glue', 'mrpc')
m2 = nlp.load_metric('glue', 'sst2')
I will get the same error `ValueError: Cannot acquire lock, caching file might be used by another process, you should setup a unique 'experiment_id'... | I can't load metric (glue) anymore after an error in a previous run. I even removed the whole cache folder `/home/XXX/.cache/huggingface/`, and the issue persisted. What are the steps to fix this?
Traceback (most recent call last):
File "/home/XXX/miniconda3/envs/ML-DL-py-3.7/lib/python3.7/site-packages/n... | 49 | load_metric can't acquire lock anymore
I can't load metric (glue) anymore after an error in a previous run. I even removed the whole cache folder `/home/XXX/.cache/huggingface/`, and the issue persisted. What are the steps to fix this?
Traceback (most recent call last):
File "/home/XXX/miniconda3/envs/M... | [
0.03870772942900658,
-0.11587371677160263,
0.07235069572925568,
0.31246688961982727,
0.32045966386795044,
0.029449261724948883,
0.09009533375501633,
0.07020638138055801,
0.5023595690727234,
-0.07859200239181519,
-0.1850661337375641,
0.03060711920261383,
-0.048066362738609314,
-0.2614513039... |
https://github.com/huggingface/datasets/issues/365 | How to augment data ? | Using batched map is probably the easiest way at the moment.
What kind of augmentation would you like to do ? | Is there any clean way to augment data ?
For now my work-around is to use batched map, like this :
```python
def aug(samples):
# Simply copy the existing data to have x2 amount of data
for k, v in samples.items():
samples[k].extend(v)
return samples
dataset = dataset.map(aug, batched=T... | 21 | How to augment data ?
Is there any clean way to augment data ?
For now my work-around is to use batched map, like this :
```python
def aug(samples):
# Simply copy the existing data to have x2 amount of data
for k, v in samples.items():
samples[k].extend(v)
return samples
dataset = da... | [
-0.007611233741044998,
-0.07861065864562988,
-0.3029455840587616,
-0.09332121908664703,
0.07471174746751785,
0.27884113788604736,
-0.10457965731620789,
0.28144028782844543,
0.12231405079364777,
0.11616498231887817,
-0.12247373908758163,
0.05612020939588547,
-0.002182275988161564,
0.1659334... |
https://github.com/huggingface/datasets/issues/365 | How to augment data ? | Some samples in the dataset are too long, I want to divide them in several samples. | Is there any clean way to augment data ?
For now my work-around is to use batched map, like this :
```python
def aug(samples):
# Simply copy the existing data to have x2 amount of data
for k, v in samples.items():
samples[k].extend(v)
return samples
dataset = dataset.map(aug, batched=T... | 16 | How to augment data ?
Is there any clean way to augment data ?
For now my work-around is to use batched map, like this :
```python
def aug(samples):
# Simply copy the existing data to have x2 amount of data
for k, v in samples.items():
samples[k].extend(v)
return samples
dataset = da... | [
-0.09215527772903442,
-0.06655604392290115,
-0.29764771461486816,
-0.022698625922203064,
0.023151500150561333,
0.35464316606521606,
-0.014462675899267197,
0.27358660101890564,
0.1672644466161728,
0.15173587203025818,
-0.11430373787879944,
0.0416504442691803,
-0.01273160520941019,
0.1264986... |
https://github.com/huggingface/datasets/issues/365 | How to augment data ? | Using batched map is the way to go then.
We'll make it clearer in the docs that map could be used for augmentation.
Let me know if you think there should be another way to do it. Or feel free to close the issue otherwise. | Is there any clean way to augment data ?
For now my work-around is to use batched map, like this :
```python
def aug(samples):
# Simply copy the existing data to have x2 amount of data
for k, v in samples.items():
samples[k].extend(v)
return samples
dataset = dataset.map(aug, batched=T... | 45 | How to augment data ?
Is there any clean way to augment data ?
For now my work-around is to use batched map, like this :
```python
def aug(samples):
# Simply copy the existing data to have x2 amount of data
for k, v in samples.items():
samples[k].extend(v)
return samples
dataset = da... | [
-0.016944844275712967,
-0.14556512236595154,
-0.30010664463043213,
-0.08540026843547821,
0.12440058588981628,
0.1755865514278412,
-0.06294336915016174,
0.3091181516647339,
0.15044242143630981,
0.16817902028560638,
-0.13062506914138794,
0.1878175288438797,
0.03093229979276657,
0.19953554868... |
https://github.com/huggingface/datasets/issues/365 | How to augment data ? | It just feels awkward to use map to augment data. Also it means it's not possible to augment data in a non-batched way.
But to be honest I have no idea of a good API... | Is there any clean way to augment data ?
For now my work-around is to use batched map, like this :
```python
def aug(samples):
# Simply copy the existing data to have x2 amount of data
for k, v in samples.items():
samples[k].extend(v)
return samples
dataset = dataset.map(aug, batched=T... | 35 | How to augment data ?
Is there any clean way to augment data ?
For now my work-around is to use batched map, like this :
```python
def aug(samples):
# Simply copy the existing data to have x2 amount of data
for k, v in samples.items():
samples[k].extend(v)
return samples
dataset = da... | [
-0.06860491633415222,
-0.03053918480873108,
-0.2990650534629822,
-0.11018501222133636,
0.062444575130939484,
0.3354523777961731,
-0.17495858669281006,
0.27413201332092285,
0.15646104514598846,
0.13222236931324005,
-0.04436758905649185,
0.09031341224908829,
-0.017896469682455063,
0.21457630... |
https://github.com/huggingface/datasets/issues/365 | How to augment data ? | Or for non-batched samples, how about returning a tuple ?
```python
def aug(sample):
# Simply copy the existing data to have x2 amount of data
return sample, sample
dataset = dataset.map(aug)
```
It feels really natural and easy, but :
* it means the behavior with batched data is different
* I ... | Is there any clean way to augment data ?
For now my work-around is to use batched map, like this :
```python
def aug(samples):
# Simply copy the existing data to have x2 amount of data
for k, v in samples.items():
samples[k].extend(v)
return samples
dataset = dataset.map(aug, batched=T... | 60 | How to augment data ?
Is there any clean way to augment data ?
For now my work-around is to use batched map, like this :
```python
def aug(samples):
# Simply copy the existing data to have x2 amount of data
for k, v in samples.items():
samples[k].extend(v)
return samples
dataset = da... | [
-0.06329986453056335,
-0.055646903812885284,
-0.2771684229373932,
-0.03686375170946121,
0.06049259006977081,
0.2008812427520752,
-0.08675959706306458,
0.3342856764793396,
0.23162955045700073,
0.13165463507175446,
-0.043928615748882294,
0.15792135894298553,
-0.08911168575286865,
0.229711785... |
https://github.com/huggingface/datasets/issues/365 | How to augment data ? | As we're working with arrow's columnar format we prefer to play with batches that are dictionaries instead of tuples.
If we have tuple it implies to re-format the data each time we want to write to arrow, which can lower the speed of map for example.
It's also a matter of coherence, as we don't want users to be con... | Is there any clean way to augment data ?
For now my work-around is to use batched map, like this :
```python
def aug(samples):
# Simply copy the existing data to have x2 amount of data
for k, v in samples.items():
samples[k].extend(v)
return samples
dataset = dataset.map(aug, batched=T... | 77 | How to augment data ?
Is there any clean way to augment data ?
For now my work-around is to use batched map, like this :
```python
def aug(samples):
# Simply copy the existing data to have x2 amount of data
for k, v in samples.items():
samples[k].extend(v)
return samples
dataset = da... | [
-0.04496316611766815,
0.03150579333305359,
-0.26236045360565186,
-0.062032848596572876,
0.07125946879386902,
0.17408183217048645,
-0.04355553165078163,
0.285645067691803,
0.29270511865615845,
0.06880292296409607,
-0.04522841423749924,
0.2766706049442291,
-0.06152551621198654,
0.09913749247... |
https://github.com/huggingface/datasets/issues/361 | 🐛 [Metrics] ROUGE is non-deterministic | > Hi, can you give a full self-contained example to reproduce this behavior?
There is a notebook in the post ;) | If I run the ROUGE metric 2 times, with same predictions / references, the scores are slightly different.
Refer to [this Colab notebook](https://colab.research.google.com/drive/1wRssNXgb9ldcp4ulwj-hMJn0ywhDOiDy?usp=sharing) for reproducing the problem.
Example of F-score for ROUGE-1, ROUGE-2, ROUGE-L in 2 differe... | 21 | 🐛 [Metrics] ROUGE is non-deterministic
If I run the ROUGE metric 2 times, with same predictions / references, the scores are slightly different.
Refer to [this Colab notebook](https://colab.research.google.com/drive/1wRssNXgb9ldcp4ulwj-hMJn0ywhDOiDy?usp=sharing) for reproducing the problem.
Example of F-score ... | [
-0.07294240593910217,
-0.4319975674152374,
-0.07466112077236176,
0.24723781645298004,
0.14245866239070892,
-0.29910212755203247,
-0.03466444090008736,
-0.36222320795059204,
-0.04236932098865509,
0.47043144702911377,
0.02695254608988762,
0.3044506311416626,
0.030214812606573105,
0.087276749... |
https://github.com/huggingface/datasets/issues/361 | 🐛 [Metrics] ROUGE is non-deterministic | > If I run the ROUGE metric 2 times, with same predictions / references, the scores are slightly different.
>
> Refer to [this Colab notebook](https://colab.research.google.com/drive/1wRssNXgb9ldcp4ulwj-hMJn0ywhDOiDy?usp=sharing) for reproducing the problem.
>
> Example of F-score for ROUGE-1, ROUGE-2, ROUGE-L in... | If I run the ROUGE metric 2 times, with same predictions / references, the scores are slightly different.
Refer to [this Colab notebook](https://colab.research.google.com/drive/1wRssNXgb9ldcp4ulwj-hMJn0ywhDOiDy?usp=sharing) for reproducing the problem.
Example of F-score for ROUGE-1, ROUGE-2, ROUGE-L in 2 differe... | 112 | 🐛 [Metrics] ROUGE is non-deterministic
If I run the ROUGE metric 2 times, with same predictions / references, the scores are slightly different.
Refer to [this Colab notebook](https://colab.research.google.com/drive/1wRssNXgb9ldcp4ulwj-hMJn0ywhDOiDy?usp=sharing) for reproducing the problem.
Example of F-score ... | [
-0.01896807737648487,
-0.23581278324127197,
-0.08905786275863647,
0.03522185981273651,
0.18351773917675018,
-0.3484433889389038,
-0.00037471577525138855,
-0.2616442143917084,
-0.2713964879512787,
0.46242862939834595,
-0.1386534720659256,
0.36095574498176575,
-0.026859894394874573,
0.079983... |
https://github.com/huggingface/datasets/issues/361 | 🐛 [Metrics] ROUGE is non-deterministic | Now if you re-run the notebook, the two printed results are the same @colanim
```
['0.3356', '0.1466', '0.2318']
['0.3356', '0.1466', '0.2318']
```
However across sessions, the results may change (as numpy's random seed can be different). You can prevent that by setting your seed:
```python
rouge = nlp.load_metr... | If I run the ROUGE metric 2 times, with same predictions / references, the scores are slightly different.
Refer to [this Colab notebook](https://colab.research.google.com/drive/1wRssNXgb9ldcp4ulwj-hMJn0ywhDOiDy?usp=sharing) for reproducing the problem.
Example of F-score for ROUGE-1, ROUGE-2, ROUGE-L in 2 differe... | 50 | 🐛 [Metrics] ROUGE is non-deterministic
If I run the ROUGE metric 2 times, with same predictions / references, the scores are slightly different.
Refer to [this Colab notebook](https://colab.research.google.com/drive/1wRssNXgb9ldcp4ulwj-hMJn0ywhDOiDy?usp=sharing) for reproducing the problem.
Example of F-score ... | [
-0.10732795298099518,
-0.21593067049980164,
-0.06680144369602203,
0.1431296467781067,
0.1828395575284958,
-0.3188179135322571,
-0.13666373491287231,
-0.15665288269519806,
-0.08101814240217209,
0.39420783519744873,
-0.09586959332227707,
0.5732386112213135,
0.0757608711719513,
-0.03530623018... |
https://github.com/huggingface/datasets/issues/360 | [Feature request] Add dataset.ragged_map() function for many-to-many transformations | Actually `map(batched=True)` can already change the size of the dataset.
It can accept examples of length `N` and returns a batch of length `M` (can be null or greater than `N`).
I'll make that explicit in the doc that I'm currently writing. | `dataset.map()` enables one-to-one transformations. Input one example and output one example. This is helpful for tokenizing and cleaning individual lines.
`dataset.filter()` enables one-to-(one-or-none) transformations. Input one example and output either zero/one example. This is helpful for removing portions from t... | 42 | [Feature request] Add dataset.ragged_map() function for many-to-many transformations
`dataset.map()` enables one-to-one transformations. Input one example and output one example. This is helpful for tokenizing and cleaning individual lines.
`dataset.filter()` enables one-to-(one-or-none) transformations. Input one e... | [
-0.5117927193641663,
-0.06494362652301788,
0.04673798754811287,
-0.21277664601802826,
-0.052597470581531525,
-0.09725026786327362,
0.2488027811050415,
0.35040542483329773,
-0.24927513301372528,
0.05863837152719498,
0.193172425031662,
0.4199867844581604,
-0.42083540558815,
-0.11852324008941... |
https://github.com/huggingface/datasets/issues/360 | [Feature request] Add dataset.ragged_map() function for many-to-many transformations | You're two steps ahead of me :) In my testing, it also works if `M` < `N`.
A batched map of different length seems to work if you directly overwrite all of the original keys, but fails if any of the original keys are preserved.
For example,
```python
# Create a dummy dataset
dset = load_dataset("wikitext", "wi... | `dataset.map()` enables one-to-one transformations. Input one example and output one example. This is helpful for tokenizing and cleaning individual lines.
`dataset.filter()` enables one-to-(one-or-none) transformations. Input one example and output either zero/one example. This is helpful for removing portions from t... | 179 | [Feature request] Add dataset.ragged_map() function for many-to-many transformations
`dataset.map()` enables one-to-one transformations. Input one example and output one example. This is helpful for tokenizing and cleaning individual lines.
`dataset.filter()` enables one-to-(one-or-none) transformations. Input one e... | [
-0.5140851140022278,
-0.06388473510742188,
0.04673585668206215,
-0.22718216478824615,
-0.06906390190124512,
-0.05510628968477249,
0.27751094102859497,
0.3193070888519287,
-0.2335917055606842,
0.051775798201560974,
0.18533766269683838,
0.4330143332481384,
-0.38917070627212524,
-0.1567607372... |
https://github.com/huggingface/datasets/issues/359 | ArrowBasedBuilder _prepare_split parse_schema breaks on nested structures | Hi, it depends on what it is in your `dataset_builder.py` file. Can you share it?
If you are just loading `json` files, you can also directly use the `json` script (which will find the schema/features from your JSON structure):
```python
from nlp import load_dataset
ds = load_dataset("json", data_files=rel_data... | I tried using the Json dataloader to load some JSON lines files. but get an exception in the parse_schema function.
```
---------------------------------------------------------------------------
ValueError Traceback (most recent call last)
<ipython-input-23-9aecfbee53bd> in <mo... | 49 | ArrowBasedBuilder _prepare_split parse_schema breaks on nested structures
I tried using the Json dataloader to load some JSON lines files. but get an exception in the parse_schema function.
```
---------------------------------------------------------------------------
ValueError ... | [
-0.32604533433914185,
0.36101141571998596,
-0.06005537882447243,
0.5487004518508911,
0.04070337116718292,
-0.1400742530822754,
0.343112975358963,
0.428407222032547,
0.21433182060718536,
-0.11396986246109009,
-0.04363139718770981,
0.33327051997184753,
0.23643998801708221,
0.0587034896016120... |
https://github.com/huggingface/datasets/issues/359 | ArrowBasedBuilder _prepare_split parse_schema breaks on nested structures | The behavior I'm seeing is from the `json` script.
I hacked this together to overcome the error with the `JSON` dataloader
```
class DatasetBuilder(hf_nlp.ArrowBasedBuilder):
BUILDER_CONFIG_CLASS = BuilderConfig
def _info(self):
return DatasetInfo()
def _split_generators(self, dl_manag... | I tried using the Json dataloader to load some JSON lines files. but get an exception in the parse_schema function.
```
---------------------------------------------------------------------------
ValueError Traceback (most recent call last)
<ipython-input-23-9aecfbee53bd> in <mo... | 254 | ArrowBasedBuilder _prepare_split parse_schema breaks on nested structures
I tried using the Json dataloader to load some JSON lines files. but get an exception in the parse_schema function.
```
---------------------------------------------------------------------------
ValueError ... | [
-0.32604533433914185,
0.36101141571998596,
-0.06005537882447243,
0.5487004518508911,
0.04070337116718292,
-0.1400742530822754,
0.343112975358963,
0.428407222032547,
0.21433182060718536,
-0.11396986246109009,
-0.04363139718770981,
0.33327051997184753,
0.23643998801708221,
0.0587034896016120... |
https://github.com/huggingface/datasets/issues/359 | ArrowBasedBuilder _prepare_split parse_schema breaks on nested structures | Also noticed that if you for example in a loader script
```
from nlp import ArrowBasedBuilder
class MyBuilder(ArrowBasedBuilder):
...
```
and use that in the subclass, it will be on the module's __dict__ and will be selected before the `MyBuilder` subclass, and it will raise `NotImplementedError` on its `_g... | I tried using the Json dataloader to load some JSON lines files. but get an exception in the parse_schema function.
```
---------------------------------------------------------------------------
ValueError Traceback (most recent call last)
<ipython-input-23-9aecfbee53bd> in <mo... | 70 | ArrowBasedBuilder _prepare_split parse_schema breaks on nested structures
I tried using the Json dataloader to load some JSON lines files. but get an exception in the parse_schema function.
```
---------------------------------------------------------------------------
ValueError ... | [
-0.32604533433914185,
0.36101141571998596,
-0.06005537882447243,
0.5487004518508911,
0.04070337116718292,
-0.1400742530822754,
0.343112975358963,
0.428407222032547,
0.21433182060718536,
-0.11396986246109009,
-0.04363139718770981,
0.33327051997184753,
0.23643998801708221,
0.0587034896016120... |
https://github.com/huggingface/datasets/issues/359 | ArrowBasedBuilder _prepare_split parse_schema breaks on nested structures | Indeed this is part of a more general limitation which is the fact that we should generate and update the `features` from the auto-inferred Arrow schema when they are not provided (also happen when a user change the schema using `map()`, the features should be auto-generated and guessed as much as possible to keep the ... | I tried using the Json dataloader to load some JSON lines files. but get an exception in the parse_schema function.
```
---------------------------------------------------------------------------
ValueError Traceback (most recent call last)
<ipython-input-23-9aecfbee53bd> in <mo... | 70 | ArrowBasedBuilder _prepare_split parse_schema breaks on nested structures
I tried using the Json dataloader to load some JSON lines files. but get an exception in the parse_schema function.
```
---------------------------------------------------------------------------
ValueError ... | [
-0.32604533433914185,
0.36101141571998596,
-0.06005537882447243,
0.5487004518508911,
0.04070337116718292,
-0.1400742530822754,
0.343112975358963,
0.428407222032547,
0.21433182060718536,
-0.11396986246109009,
-0.04363139718770981,
0.33327051997184753,
0.23643998801708221,
0.0587034896016120... |
https://github.com/huggingface/datasets/issues/355 | can't load SNLI dataset | I just added the processed files of `snli` on our google storage, so that when you do `load_dataset` it can download the processed files from there :)
We are thinking about having available those processed files for more datasets in the future, because sometimes files aren't available (like for `snli`), or the downl... | `nlp` seems to load `snli` from some URL based on nlp.stanford.edu. This subdomain is frequently down -- including right now, when I'd like to load `snli` in a Colab notebook, but can't.
Is there a plan to move these datasets to huggingface servers for a more stable solution?
Btw, here's the stack trace:
```
... | 66 | can't load SNLI dataset
`nlp` seems to load `snli` from some URL based on nlp.stanford.edu. This subdomain is frequently down -- including right now, when I'd like to load `snli` in a Colab notebook, but can't.
Is there a plan to move these datasets to huggingface servers for a more stable solution?
Btw, here's... | [
0.1304485946893692,
-0.04836539179086685,
0.05979367718100548,
0.4217779040336609,
0.16876092553138733,
-0.21573856472969055,
0.2670188248157501,
0.01613372378051281,
-0.024434279650449753,
-0.10663741827011108,
-0.38732466101646423,
-0.0008524451404809952,
0.1453942060470581,
0.4104742109... |
https://github.com/huggingface/datasets/issues/353 | [Dataset requests] New datasets for Text Classification | - `nlp` has MR! It's called `rotten_tomatoes`
- SST is part of GLUE, or is that just SST-2?
- `nlp` also has `ag_news`, a popular news classification dataset
I'd also like to see:
- the Yahoo Answers topic classification dataset
- the Kaggle Fake News classification dataset | We are missing a few datasets for Text Classification which is an important field.
Namely, it would be really nice to add:
- [x] TREC-6 dataset (see here for instance: https://pytorchnlp.readthedocs.io/en/latest/source/torchnlp.datasets.html#torchnlp.datasets.trec_dataset) **[done]**
- #386
- [x] Yelp-5
- #... | 47 | [Dataset requests] New datasets for Text Classification
We are missing a few datasets for Text Classification which is an important field.
Namely, it would be really nice to add:
- [x] TREC-6 dataset (see here for instance: https://pytorchnlp.readthedocs.io/en/latest/source/torchnlp.datasets.html#torchnlp.dataset... | [
0.04040953889489174,
0.17049479484558105,
-0.17783166468143463,
0.20942765474319458,
0.2170809954404831,
0.2489396631717682,
0.2135149985551834,
0.11218048632144928,
-0.17596904933452606,
-0.03818897902965546,
-0.14788630604743958,
0.11309768259525299,
-0.1243516281247139,
0.05512295663356... |
https://github.com/huggingface/datasets/issues/353 | [Dataset requests] New datasets for Text Classification | Thanks @jxmorris12 for pointing this out.
In glue we only have SST-2 maybe we can add separately SST-1.
| We are missing a few datasets for Text Classification which is an important field.
Namely, it would be really nice to add:
- [x] TREC-6 dataset (see here for instance: https://pytorchnlp.readthedocs.io/en/latest/source/torchnlp.datasets.html#torchnlp.datasets.trec_dataset) **[done]**
- #386
- [x] Yelp-5
- #... | 18 | [Dataset requests] New datasets for Text Classification
We are missing a few datasets for Text Classification which is an important field.
Namely, it would be really nice to add:
- [x] TREC-6 dataset (see here for instance: https://pytorchnlp.readthedocs.io/en/latest/source/torchnlp.datasets.html#torchnlp.dataset... | [
0.028054621070623398,
0.14188173413276672,
-0.16715145111083984,
0.24594809114933014,
0.24599720537662506,
0.2608925998210907,
0.24158169329166412,
0.09973177313804626,
-0.1843867152929306,
-0.01903223991394043,
-0.13721129298210144,
0.08285900205373764,
-0.10888361930847168,
0.07370631396... |
https://github.com/huggingface/datasets/issues/353 | [Dataset requests] New datasets for Text Classification | This is the homepage for the Amazon dataset: https://www.kaggle.com/datafiniti/consumer-reviews-of-amazon-products
Is there an easy way to download kaggle datasets programmatically? If so, I can add this one! | We are missing a few datasets for Text Classification which is an important field.
Namely, it would be really nice to add:
- [x] TREC-6 dataset (see here for instance: https://pytorchnlp.readthedocs.io/en/latest/source/torchnlp.datasets.html#torchnlp.datasets.trec_dataset) **[done]**
- #386
- [x] Yelp-5
- #... | 26 | [Dataset requests] New datasets for Text Classification
We are missing a few datasets for Text Classification which is an important field.
Namely, it would be really nice to add:
- [x] TREC-6 dataset (see here for instance: https://pytorchnlp.readthedocs.io/en/latest/source/torchnlp.datasets.html#torchnlp.dataset... | [
0.02771371230483055,
0.14573022723197937,
-0.18174248933792114,
0.22419831156730652,
0.2346155345439911,
0.26100558042526245,
0.23799993097782135,
0.11230877041816711,
-0.1869863122701645,
0.002322852611541748,
-0.14318500459194183,
0.09494903683662415,
-0.09309766441583633,
0.049053803086... |
https://github.com/huggingface/datasets/issues/353 | [Dataset requests] New datasets for Text Classification | Hi @jxmorris12 for now I think our `dl_manager` does not download from Kaggle.
@thomwolf , @lhoestq | We are missing a few datasets for Text Classification which is an important field.
Namely, it would be really nice to add:
- [x] TREC-6 dataset (see here for instance: https://pytorchnlp.readthedocs.io/en/latest/source/torchnlp.datasets.html#torchnlp.datasets.trec_dataset) **[done]**
- #386
- [x] Yelp-5
- #... | 16 | [Dataset requests] New datasets for Text Classification
We are missing a few datasets for Text Classification which is an important field.
Namely, it would be really nice to add:
- [x] TREC-6 dataset (see here for instance: https://pytorchnlp.readthedocs.io/en/latest/source/torchnlp.datasets.html#torchnlp.dataset... | [
0.027320127934217453,
0.14617666602134705,
-0.17790375649929047,
0.23313388228416443,
0.24200890958309174,
0.255913645029068,
0.24614986777305603,
0.10839653015136719,
-0.18786610662937164,
-0.0035496950149536133,
-0.14122195541858673,
0.0886833518743515,
-0.1026228815317154,
0.07495999336... |
https://github.com/huggingface/datasets/issues/353 | [Dataset requests] New datasets for Text Classification | Great list. Any idea if Amazon Reviews has been added?
- ~40 GB of text (sadly no emoji)
- popular MLM pre-training dataset before bigger datasets like WebText https://arxiv.org/abs/1808.01371
- turns out that binarizing the 1-5 star rating leads to great Pos/Neg/Neutral dataset, T5 paper claims to get very high a... | We are missing a few datasets for Text Classification which is an important field.
Namely, it would be really nice to add:
- [x] TREC-6 dataset (see here for instance: https://pytorchnlp.readthedocs.io/en/latest/source/torchnlp.datasets.html#torchnlp.datasets.trec_dataset) **[done]**
- #386
- [x] Yelp-5
- #... | 92 | [Dataset requests] New datasets for Text Classification
We are missing a few datasets for Text Classification which is an important field.
Namely, it would be really nice to add:
- [x] TREC-6 dataset (see here for instance: https://pytorchnlp.readthedocs.io/en/latest/source/torchnlp.datasets.html#torchnlp.dataset... | [
0.04690634086728096,
0.16295549273490906,
-0.1858590841293335,
0.263063907623291,
0.23614728450775146,
0.280469685792923,
0.20281466841697693,
0.05831862986087799,
-0.15025897324085236,
-0.04523658752441406,
-0.1383245438337326,
0.05618837848305702,
-0.11800996214151382,
0.0506220757961273... |
https://github.com/huggingface/datasets/issues/353 | [Dataset requests] New datasets for Text Classification | On the Amazon Reviews dataset, the original UCSD website has noted these are now updated to include product reviews through 2018 -- actually quite recent compared to many other datasets. Almost certainly the largest NLP dataset out there with labels!
https://jmcauley.ucsd.edu/data/amazon/
Any chance someone has ti... | We are missing a few datasets for Text Classification which is an important field.
Namely, it would be really nice to add:
- [x] TREC-6 dataset (see here for instance: https://pytorchnlp.readthedocs.io/en/latest/source/torchnlp.datasets.html#torchnlp.datasets.trec_dataset) **[done]**
- #386
- [x] Yelp-5
- #... | 56 | [Dataset requests] New datasets for Text Classification
We are missing a few datasets for Text Classification which is an important field.
Namely, it would be really nice to add:
- [x] TREC-6 dataset (see here for instance: https://pytorchnlp.readthedocs.io/en/latest/source/torchnlp.datasets.html#torchnlp.dataset... | [
0.031767334789037704,
0.15160924196243286,
-0.1775275468826294,
0.23074491322040558,
0.2376454472541809,
0.2634916603565216,
0.23497915267944336,
0.09712781012058258,
-0.18781433999538422,
-0.02931693196296692,
-0.15768304467201233,
0.0860939472913742,
-0.10020796209573746,
0.0800861641764... |
https://github.com/huggingface/datasets/issues/353 | [Dataset requests] New datasets for Text Classification | @albertvillanova How up to date is this issue? I see that some of these datasets are now on huggingface but have not been checked off the list | We are missing a few datasets for Text Classification which is an important field.
Namely, it would be really nice to add:
- [x] TREC-6 dataset (see here for instance: https://pytorchnlp.readthedocs.io/en/latest/source/torchnlp.datasets.html#torchnlp.datasets.trec_dataset) **[done]**
- #386
- [x] Yelp-5
- #... | 27 | [Dataset requests] New datasets for Text Classification
We are missing a few datasets for Text Classification which is an important field.
Namely, it would be really nice to add:
- [x] TREC-6 dataset (see here for instance: https://pytorchnlp.readthedocs.io/en/latest/source/torchnlp.datasets.html#torchnlp.dataset... | [
0.01319456472992897,
0.1501375138759613,
-0.18076933920383453,
0.21455153822898865,
0.22666853666305542,
0.23878875374794006,
0.24076393246650696,
0.0955100804567337,
-0.20201937854290009,
0.017560943961143494,
-0.14037388563156128,
0.09364362806081772,
-0.07991866022348404,
0.077525511384... |
https://github.com/huggingface/datasets/issues/347 | 'cp950' codec error from load_dataset('xtreme', 'tydiqa') | This is probably a Windows issue, we need to specify the encoding when `load_dataset()` reads the original CSV file.
Try to find the `open()` statement called by `load_dataset()` and add an `encoding='utf-8'` parameter.
See issues #242 and #307 | 
I guess the error is related to python source encoding issue that my PC is trying to decode the source code with wrong encoding-decoding tools, perhaps :
https://www.python.org/dev/peps/pep-0263/
I gues... | 38 | 'cp950' codec error from load_dataset('xtreme', 'tydiqa')
I guess the error is related to python source encoding issue that my PC is trying to decode the source code with wrong encoding-decoding tools, pe... | [
-0.2610889971256256,
-0.07662422955036163,
-0.05322106182575226,
0.14644591510295868,
0.3858146667480469,
0.010254383087158203,
0.025446821004152298,
0.17635707557201385,
-0.17912451922893524,
0.08384843170642853,
0.17901656031608582,
0.43521299958229065,
0.058860938996076584,
0.3545266389... |
https://github.com/huggingface/datasets/issues/347 | 'cp950' codec error from load_dataset('xtreme', 'tydiqa') | It should be in `xtreme.py:L755`:
```python
if self.config.name == "tydiqa" or self.config.name.startswith("MLQA") or self.config.name == "SQuAD":
with open(filepath) as f:
data = json.load(f)
```
Could you try to add the encoding parameter:
```python
open(filepath, encodin... | 
I guess the error is related to python source encoding issue that my PC is trying to decode the source code with wrong encoding-decoding tools, perhaps :
https://www.python.org/dev/peps/pep-0263/
I gues... | 36 | 'cp950' codec error from load_dataset('xtreme', 'tydiqa')
I guess the error is related to python source encoding issue that my PC is trying to decode the source code with wrong encoding-decoding tools, pe... | [
-0.26959267258644104,
-0.028839677572250366,
-0.020871659740805626,
0.20285934209823608,
0.44842010736465454,
-0.12191356718540192,
-0.14026230573654175,
0.2741994559764862,
-0.30686184763908386,
0.13081809878349304,
0.11824988573789597,
0.5972365736961365,
0.010757948271930218,
0.37257337... |
https://github.com/huggingface/datasets/issues/347 | 'cp950' codec error from load_dataset('xtreme', 'tydiqa') | Hello @jerryIsHere :) Did it work ?
If so we may change the dataset script to force the utf-8 encoding | 
I guess the error is related to python source encoding issue that my PC is trying to decode the source code with wrong encoding-decoding tools, perhaps :
https://www.python.org/dev/peps/pep-0263/
I gues... | 20 | 'cp950' codec error from load_dataset('xtreme', 'tydiqa')
I guess the error is related to python source encoding issue that my PC is trying to decode the source code with wrong encoding-decoding tools, pe... | [
-0.3731941282749176,
-0.0007190480828285217,
-0.05706842988729477,
0.14228716492652893,
0.4632263779640198,
-0.058336690068244934,
0.010565135627985,
0.22932957112789154,
-0.28686612844467163,
0.0433458611369133,
0.1535942107439041,
0.4569697976112366,
-0.0138363903388381,
0.38359206914901... |
https://github.com/huggingface/datasets/issues/347 | 'cp950' codec error from load_dataset('xtreme', 'tydiqa') | @lhoestq sorry for being that late, I found 4 copy of xtreme.py. I did the changes as what has been told to all of them.
The problem is not solved | 
I guess the error is related to python source encoding issue that my PC is trying to decode the source code with wrong encoding-decoding tools, perhaps :
https://www.python.org/dev/peps/pep-0263/
I gues... | 30 | 'cp950' codec error from load_dataset('xtreme', 'tydiqa')
I guess the error is related to python source encoding issue that my PC is trying to decode the source code with wrong encoding-decoding tools, pe... | [
-0.2841985523700714,
-0.17101243138313293,
-0.010297667235136032,
0.24634641408920288,
0.4728096127510071,
-0.05887721851468086,
-0.14265543222427368,
0.21194736659526825,
-0.2602454423904419,
0.1263463795185089,
0.0891541913151741,
0.5027422308921814,
-0.04542113468050957,
0.4290331304073... |
https://github.com/huggingface/datasets/issues/347 | 'cp950' codec error from load_dataset('xtreme', 'tydiqa') | Could you provide a better error message so that we can make sure it comes from the opening of the `tydiqa`'s json files ?
| 
I guess the error is related to python source encoding issue that my PC is trying to decode the source code with wrong encoding-decoding tools, perhaps :
https://www.python.org/dev/peps/pep-0263/
I gues... | 24 | 'cp950' codec error from load_dataset('xtreme', 'tydiqa')
I guess the error is related to python source encoding issue that my PC is trying to decode the source code with wrong encoding-decoding tools, pe... | [
-0.28176912665367126,
-0.07712244987487793,
-0.025954121723771095,
0.18648837506771088,
0.43187904357910156,
-0.10634154826402664,
-0.04907957836985588,
0.21024177968502045,
-0.2888762950897217,
0.07893185317516327,
0.1698380559682846,
0.5185139775276184,
-0.015734095126390457,
0.403936833... |
https://github.com/huggingface/datasets/issues/347 | 'cp950' codec error from load_dataset('xtreme', 'tydiqa') | @lhoestq
The error message is same as before:
Exception has occurred: UnicodeDecodeError
'cp950' codec can't decode byte 0xe2 in position 111: illegal multibyte sequence
File "D:\python\test\test.py", line 3, in <module>
dataset = load_dataset('xtreme', 'tydiqa')
I guess the error is related to python source encoding issue that my PC is trying to decode the source code with wrong encoding-decoding tools, perhaps :
https://www.python.org/dev/peps/pep-0263/
I gues... | 63 | 'cp950' codec error from load_dataset('xtreme', 'tydiqa')
I guess the error is related to python source encoding issue that my PC is trying to decode the source code with wrong encoding-decoding tools, pe... | [
-0.2993043065071106,
-0.0795086920261383,
-0.025460872799158096,
0.21767917275428772,
0.38568681478500366,
-0.07032664120197296,
-0.10065561532974243,
0.18437038362026215,
-0.276649534702301,
0.056092794984579086,
0.13452938199043274,
0.42476361989974976,
-0.08653882145881653,
0.2728086709... |
https://github.com/huggingface/datasets/issues/347 | 'cp950' codec error from load_dataset('xtreme', 'tydiqa') | Hi there !
I encountered the same issue with the IMDB dataset on windows. It threw an error about charmap not being able to decode a symbol during the first time I tried to download it. I checked on a remote linux machine I have, and it can't be reproduced.
I added ```encoding='UTF-8'``` to both lines that have ```op... | 
I guess the error is related to python source encoding issue that my PC is trying to decode the source code with wrong encoding-decoding tools, perhaps :
https://www.python.org/dev/peps/pep-0263/
I gues... | 72 | 'cp950' codec error from load_dataset('xtreme', 'tydiqa')
I guess the error is related to python source encoding issue that my PC is trying to decode the source code with wrong encoding-decoding tools, pe... | [
-0.3488938808441162,
0.058056458830833435,
-0.005317438393831253,
0.21481463313102722,
0.29113948345184326,
0.06939210742712021,
-0.02316361293196678,
0.1682782769203186,
-0.1171187162399292,
-0.00880354642868042,
0.035162076354026794,
0.3814254105091095,
0.059285447001457214,
0.2256461828... |
https://github.com/huggingface/datasets/issues/347 | 'cp950' codec error from load_dataset('xtreme', 'tydiqa') | > Hi there !
> I encountered the same issue with the IMDB dataset on windows. It threw an error about charmap not being able to decode a symbol during the first time I tried to download it. I checked on a remote linux machine I have, and it can't be reproduced.
> I added `encoding='UTF-8'` to both lines that have `op... | 
I guess the error is related to python source encoding issue that my PC is trying to decode the source code with wrong encoding-decoding tools, perhaps :
https://www.python.org/dev/peps/pep-0263/
I gues... | 97 | 'cp950' codec error from load_dataset('xtreme', 'tydiqa')
I guess the error is related to python source encoding issue that my PC is trying to decode the source code with wrong encoding-decoding tools, pe... | [
-0.3819618821144104,
0.032849475741386414,
-0.015480727888643742,
0.2028752863407135,
0.31209009885787964,
0.06895846873521805,
0.005816839635372162,
0.18256886303424835,
-0.08367151767015457,
0.029334619641304016,
0.014070959761738777,
0.4008612036705017,
0.05778166651725769,
0.2399233132... |
https://github.com/huggingface/datasets/issues/347 | 'cp950' codec error from load_dataset('xtreme', 'tydiqa') | > This is probably a Windows issue, we need to specify the encoding when `load_dataset()` reads the original CSV file.
> Try to find the `open()` statement called by `load_dataset()` and add an `encoding='utf-8'` parameter.
> See issues #242 and #307
Sorry for not responding for about a month.
I have just found t... | 
I guess the error is related to python source encoding issue that my PC is trying to decode the source code with wrong encoding-decoding tools, perhaps :
https://www.python.org/dev/peps/pep-0263/
I gues... | 115 | 'cp950' codec error from load_dataset('xtreme', 'tydiqa')
I guess the error is related to python source encoding issue that my PC is trying to decode the source code with wrong encoding-decoding tools, pe... | [
-0.32184672355651855,
0.04104040563106537,
-0.017263025045394897,
0.19213111698627472,
0.38413745164871216,
-0.04183315113186836,
0.053385064005851746,
0.1839183121919632,
-0.24381500482559204,
0.07465638220310211,
0.11224352568387985,
0.4671258330345154,
0.11079854518175125,
0.32633516192... |
https://github.com/huggingface/datasets/issues/347 | 'cp950' codec error from load_dataset('xtreme', 'tydiqa') | Since #481 we shouldn't have other issues with encodings as they need to be set to "utf-8" be default.
Closing this one, but feel free to re-open if you gave other questions | 
I guess the error is related to python source encoding issue that my PC is trying to decode the source code with wrong encoding-decoding tools, perhaps :
https://www.python.org/dev/peps/pep-0263/
I gues... | 32 | 'cp950' codec error from load_dataset('xtreme', 'tydiqa')
I guess the error is related to python source encoding issue that my PC is trying to decode the source code with wrong encoding-decoding tools, pe... | [
-0.29207688570022583,
-0.11546691507101059,
-0.04850517958402634,
0.14296966791152954,
0.39768534898757935,
-0.1386854350566864,
-0.029937509447336197,
0.2073104977607727,
-0.34860384464263916,
0.11078862845897675,
0.17416460812091827,
0.4716801643371582,
0.022073112428188324,
0.3367836475... |
https://github.com/huggingface/datasets/issues/345 | Supporting documents in ELI5 | Hi @saverymax ! For licensing reasons, the original team was unable to release pre-processed CommonCrawl documents. Instead, they provided a script to re-create them from a CommonCrawl dump, but it unfortunately requires access to a medium-large size cluster:
https://github.com/facebookresearch/ELI5#downloading-suppor... | I was attempting to use the ELI5 dataset, when I realized that huggingface does not provide the supporting documents (the source documents from the common crawl). Without the supporting documents, this makes the dataset about as useful for my project as a block of cheese, or some other more apt metaphor. According to ... | 130 | Supporting documents in ELI5
I was attempting to use the ELI5 dataset, when I realized that huggingface does not provide the supporting documents (the source documents from the common crawl). Without the supporting documents, this makes the dataset about as useful for my project as a block of cheese, or some other mo... | [
0.08067645132541656,
-0.14141637086868286,
-0.1308176964521408,
-0.0987008810043335,
-0.3395775556564331,
-0.021614450961351395,
-0.21735234558582306,
0.21030329167842865,
0.05030122771859169,
-0.025913551449775696,
0.12809036672115326,
-0.24119505286216736,
0.10086890310049057,
0.14274106... |
https://github.com/huggingface/datasets/issues/345 | Supporting documents in ELI5 | Hi, thanks for the quick response. The blog post is quite an interesting working example, thanks for sharing it.
Two follow-up points/questions about my original question:
1. Yes, I read that the facebook team could not share the CommonCrawl b/c of licensing reasons. They state "No, we are not allowed to host proce... | I was attempting to use the ELI5 dataset, when I realized that huggingface does not provide the supporting documents (the source documents from the common crawl). Without the supporting documents, this makes the dataset about as useful for my project as a block of cheese, or some other more apt metaphor. According to ... | 256 | Supporting documents in ELI5
I was attempting to use the ELI5 dataset, when I realized that huggingface does not provide the supporting documents (the source documents from the common crawl). Without the supporting documents, this makes the dataset about as useful for my project as a block of cheese, or some other mo... | [
0.161881223320961,
-0.045203134417533875,
-0.038786016404628754,
-0.04772616922855377,
-0.31241023540496826,
0.0002396106719970703,
-0.13539814949035645,
0.13022977113723755,
-0.021920228376984596,
-0.05018483102321625,
0.09411699324846268,
-0.26511600613594055,
0.11306068301200867,
0.1443... |
https://github.com/huggingface/datasets/issues/331 | Loading CNN/Daily Mail dataset produces `nlp.utils.info_utils.NonMatchingSplitsSizesError` | I couldn't reproduce on my side.
It looks like you were not able to generate all the examples, and you have the problem for each split train-test-validation.
Could you try to enable logging, try again and send the logs ?
```python
import logging
logging.basicConfig(level=logging.INFO)
``` | ```
>>> import nlp
>>> nlp.load_dataset('cnn_dailymail', '3.0.0')
Downloading and preparing dataset cnn_dailymail/3.0.0 (download: 558.32 MiB, generated: 1.26 GiB, total: 1.81 GiB) to /u/jm8wx/.cache/huggingface/datasets/cnn_dailymail/3.0.0/3.0.0...
Traceback (most recent call last):
File "<stdin>", line 1, in... | 45 | Loading CNN/Daily Mail dataset produces `nlp.utils.info_utils.NonMatchingSplitsSizesError`
```
>>> import nlp
>>> nlp.load_dataset('cnn_dailymail', '3.0.0')
Downloading and preparing dataset cnn_dailymail/3.0.0 (download: 558.32 MiB, generated: 1.26 GiB, total: 1.81 GiB) to /u/jm8wx/.cache/huggingface/datasets/cnn... | [
-0.14106932282447815,
-0.05698748677968979,
-0.008213063701987267,
0.2086566984653473,
0.0776498094201088,
0.09079600125551224,
0.18330222368240356,
0.5251759886741638,
0.09873126447200775,
0.15423263609409332,
0.01629825122654438,
0.07699623703956604,
-0.3141201138496399,
-0.0437746122479... |
https://github.com/huggingface/datasets/issues/331 | Loading CNN/Daily Mail dataset produces `nlp.utils.info_utils.NonMatchingSplitsSizesError` | here's the log
```
>>> import nlp
import logging
logging.basicConfig(level=logging.INFO)
nlp.load_dataset('cnn_dailymail', '3.0.0')
>>> import logging
>>> logging.basicConfig(level=logging.INFO)
>>> nlp.load_dataset('cnn_dailymail', '3.0.0')
INFO:nlp.load:Checking /u/jm8wx/.cache/huggingface/datasets/720d2e20d... | ```
>>> import nlp
>>> nlp.load_dataset('cnn_dailymail', '3.0.0')
Downloading and preparing dataset cnn_dailymail/3.0.0 (download: 558.32 MiB, generated: 1.26 GiB, total: 1.81 GiB) to /u/jm8wx/.cache/huggingface/datasets/cnn_dailymail/3.0.0/3.0.0...
Traceback (most recent call last):
File "<stdin>", line 1, in... | 223 | Loading CNN/Daily Mail dataset produces `nlp.utils.info_utils.NonMatchingSplitsSizesError`
```
>>> import nlp
>>> nlp.load_dataset('cnn_dailymail', '3.0.0')
Downloading and preparing dataset cnn_dailymail/3.0.0 (download: 558.32 MiB, generated: 1.26 GiB, total: 1.81 GiB) to /u/jm8wx/.cache/huggingface/datasets/cnn... | [
-0.14106932282447815,
-0.05698748677968979,
-0.008213063701987267,
0.2086566984653473,
0.0776498094201088,
0.09079600125551224,
0.18330222368240356,
0.5251759886741638,
0.09873126447200775,
0.15423263609409332,
0.01629825122654438,
0.07699623703956604,
-0.3141201138496399,
-0.0437746122479... |
https://github.com/huggingface/datasets/issues/331 | Loading CNN/Daily Mail dataset produces `nlp.utils.info_utils.NonMatchingSplitsSizesError` | > here's the log
>
> ```
> >>> import nlp
> import logging
> logging.basicConfig(level=logging.INFO)
> nlp.load_dataset('cnn_dailymail', '3.0.0')
> >>> import logging
> >>> logging.basicConfig(level=logging.INFO)
> >>> nlp.load_dataset('cnn_dailymail', '3.0.0')
> INFO:nlp.load:Checking /u/jm8wx/.cache/huggin... | ```
>>> import nlp
>>> nlp.load_dataset('cnn_dailymail', '3.0.0')
Downloading and preparing dataset cnn_dailymail/3.0.0 (download: 558.32 MiB, generated: 1.26 GiB, total: 1.81 GiB) to /u/jm8wx/.cache/huggingface/datasets/cnn_dailymail/3.0.0/3.0.0...
Traceback (most recent call last):
File "<stdin>", line 1, in... | 376 | Loading CNN/Daily Mail dataset produces `nlp.utils.info_utils.NonMatchingSplitsSizesError`
```
>>> import nlp
>>> nlp.load_dataset('cnn_dailymail', '3.0.0')
Downloading and preparing dataset cnn_dailymail/3.0.0 (download: 558.32 MiB, generated: 1.26 GiB, total: 1.81 GiB) to /u/jm8wx/.cache/huggingface/datasets/cnn... | [
-0.14106932282447815,
-0.05698748677968979,
-0.008213063701987267,
0.2086566984653473,
0.0776498094201088,
0.09079600125551224,
0.18330222368240356,
0.5251759886741638,
0.09873126447200775,
0.15423263609409332,
0.01629825122654438,
0.07699623703956604,
-0.3141201138496399,
-0.0437746122479... |
https://github.com/huggingface/datasets/issues/331 | Loading CNN/Daily Mail dataset produces `nlp.utils.info_utils.NonMatchingSplitsSizesError` | In general if some examples are missing after processing (hence causing the `NonMatchingSplitsSizesError `), it is often due to either
1) corrupted cached files
2) decoding errors
I just checked the dataset script for code that could lead to decoding errors but I couldn't find any. Before we try to dive more into ... | ```
>>> import nlp
>>> nlp.load_dataset('cnn_dailymail', '3.0.0')
Downloading and preparing dataset cnn_dailymail/3.0.0 (download: 558.32 MiB, generated: 1.26 GiB, total: 1.81 GiB) to /u/jm8wx/.cache/huggingface/datasets/cnn_dailymail/3.0.0/3.0.0...
Traceback (most recent call last):
File "<stdin>", line 1, in... | 74 | Loading CNN/Daily Mail dataset produces `nlp.utils.info_utils.NonMatchingSplitsSizesError`
```
>>> import nlp
>>> nlp.load_dataset('cnn_dailymail', '3.0.0')
Downloading and preparing dataset cnn_dailymail/3.0.0 (download: 558.32 MiB, generated: 1.26 GiB, total: 1.81 GiB) to /u/jm8wx/.cache/huggingface/datasets/cnn... | [
-0.14106932282447815,
-0.05698748677968979,
-0.008213063701987267,
0.2086566984653473,
0.0776498094201088,
0.09079600125551224,
0.18330222368240356,
0.5251759886741638,
0.09873126447200775,
0.15423263609409332,
0.01629825122654438,
0.07699623703956604,
-0.3141201138496399,
-0.0437746122479... |
https://github.com/huggingface/datasets/issues/331 | Loading CNN/Daily Mail dataset produces `nlp.utils.info_utils.NonMatchingSplitsSizesError` | Yes thanks for the support! I cleared out my cache folder and everything works fine now | ```
>>> import nlp
>>> nlp.load_dataset('cnn_dailymail', '3.0.0')
Downloading and preparing dataset cnn_dailymail/3.0.0 (download: 558.32 MiB, generated: 1.26 GiB, total: 1.81 GiB) to /u/jm8wx/.cache/huggingface/datasets/cnn_dailymail/3.0.0/3.0.0...
Traceback (most recent call last):
File "<stdin>", line 1, in... | 16 | Loading CNN/Daily Mail dataset produces `nlp.utils.info_utils.NonMatchingSplitsSizesError`
```
>>> import nlp
>>> nlp.load_dataset('cnn_dailymail', '3.0.0')
Downloading and preparing dataset cnn_dailymail/3.0.0 (download: 558.32 MiB, generated: 1.26 GiB, total: 1.81 GiB) to /u/jm8wx/.cache/huggingface/datasets/cnn... | [
-0.14106932282447815,
-0.05698748677968979,
-0.008213063701987267,
0.2086566984653473,
0.0776498094201088,
0.09079600125551224,
0.18330222368240356,
0.5251759886741638,
0.09873126447200775,
0.15423263609409332,
0.01629825122654438,
0.07699623703956604,
-0.3141201138496399,
-0.0437746122479... |
https://github.com/huggingface/datasets/issues/329 | [Bug] FileLock dependency incompatible with filesystem | Environment is Ubuntu 18.04, Python 3.7.5, nlp==0.3.0, filelock=3.0.12.
The external volume is Amazon FSx for Lustre, and it by default creates files with limited permissions. My working theory is that FileLock creates a lockfile that isn't writable, and thus there's no way to acquire it by removing the .lock file. ... | I'm downloading a dataset successfully with
`load_dataset("wikitext", "wikitext-2-raw-v1")`
But when I attempt to cache it on an external volume, it hangs indefinitely:
`load_dataset("wikitext", "wikitext-2-raw-v1", cache_dir="/fsx") # /fsx is an external volume mount`
The filesystem when hanging looks like thi... | 118 | [Bug] FileLock dependency incompatible with filesystem
I'm downloading a dataset successfully with
`load_dataset("wikitext", "wikitext-2-raw-v1")`
But when I attempt to cache it on an external volume, it hangs indefinitely:
`load_dataset("wikitext", "wikitext-2-raw-v1", cache_dir="/fsx") # /fsx is an external vo... | [
0.03299279883503914,
0.07716108858585358,
-0.02503032237291336,
0.027964189648628235,
0.08898356556892395,
0.15139448642730713,
0.36469778418540955,
0.06515342742204666,
0.709405779838562,
-0.09789606928825378,
0.09096771478652954,
-0.0051192305982112885,
0.02144586853682995,
-0.3159702420... |
https://github.com/huggingface/datasets/issues/329 | [Bug] FileLock dependency incompatible with filesystem | Looks like the `flock` syscall does not work on Lustre filesystems by default: https://github.com/benediktschmitt/py-filelock/issues/67.
I added the `-o flock` option when mounting the filesystem, as [described here](https://docs.aws.amazon.com/fsx/latest/LustreGuide/getting-started-step2.html), which fixed the issu... | I'm downloading a dataset successfully with
`load_dataset("wikitext", "wikitext-2-raw-v1")`
But when I attempt to cache it on an external volume, it hangs indefinitely:
`load_dataset("wikitext", "wikitext-2-raw-v1", cache_dir="/fsx") # /fsx is an external volume mount`
The filesystem when hanging looks like thi... | 31 | [Bug] FileLock dependency incompatible with filesystem
I'm downloading a dataset successfully with
`load_dataset("wikitext", "wikitext-2-raw-v1")`
But when I attempt to cache it on an external volume, it hangs indefinitely:
`load_dataset("wikitext", "wikitext-2-raw-v1", cache_dir="/fsx") # /fsx is an external vo... | [
-0.022068779915571213,
-0.07033123075962067,
-0.02411775104701519,
0.047782376408576965,
0.06480738520622253,
0.05381765216588974,
0.27190303802490234,
-0.09048175811767578,
0.7102741003036499,
-0.06344860792160034,
-0.023035697638988495,
-0.030101286247372627,
0.140387162566185,
-0.404071... |
https://github.com/huggingface/datasets/issues/329 | [Bug] FileLock dependency incompatible with filesystem | I'm wondering if this can be revisited. In some managed environments the same person using HF cannot change the file-system mount flags, (and the organization may be unwilling to change these flags due to other concerns) but can ensure that there won't be concurrent writes, for example because HF is offline and the mod... | I'm downloading a dataset successfully with
`load_dataset("wikitext", "wikitext-2-raw-v1")`
But when I attempt to cache it on an external volume, it hangs indefinitely:
`load_dataset("wikitext", "wikitext-2-raw-v1", cache_dir="/fsx") # /fsx is an external volume mount`
The filesystem when hanging looks like thi... | 102 | [Bug] FileLock dependency incompatible with filesystem
I'm downloading a dataset successfully with
`load_dataset("wikitext", "wikitext-2-raw-v1")`
But when I attempt to cache it on an external volume, it hangs indefinitely:
`load_dataset("wikitext", "wikitext-2-raw-v1", cache_dir="/fsx") # /fsx is an external vo... | [
-0.02784925512969494,
0.08454528450965881,
0.03664127364754677,
0.019807685166597366,
0.14161589741706848,
0.15671434998512268,
0.387428879737854,
0.18339408934116364,
0.6804934740066528,
-0.07296523451805115,
-0.10137461125850677,
-0.12369554489850998,
0.0754348486661911,
-0.2549614012241... |
https://github.com/huggingface/datasets/issues/329 | [Bug] FileLock dependency incompatible with filesystem | > I'm wondering if this can be revisited. In some managed environments the same person using HF cannot change the file-system mount flags, (and the organization may be unwilling to change these flags due to other concerns) but can ensure that there won't be concurrent writes, for example because HF is offline and the m... | I'm downloading a dataset successfully with
`load_dataset("wikitext", "wikitext-2-raw-v1")`
But when I attempt to cache it on an external volume, it hangs indefinitely:
`load_dataset("wikitext", "wikitext-2-raw-v1", cache_dir="/fsx") # /fsx is an external volume mount`
The filesystem when hanging looks like thi... | 71 | [Bug] FileLock dependency incompatible with filesystem
I'm downloading a dataset successfully with
`load_dataset("wikitext", "wikitext-2-raw-v1")`
But when I attempt to cache it on an external volume, it hangs indefinitely:
`load_dataset("wikitext", "wikitext-2-raw-v1", cache_dir="/fsx") # /fsx is an external vo... | [
-0.019822867587208748,
0.10659496486186981,
-0.0011312365531921387,
0.029618140310049057,
0.07107679545879364,
0.13281425833702087,
0.4453546404838562,
0.07700248062610626,
0.7312152981758118,
-0.024141281843185425,
-0.0960533544421196,
-0.1840338408946991,
0.05087317153811455,
-0.23052029... |
https://github.com/huggingface/datasets/issues/329 | [Bug] FileLock dependency incompatible with filesystem | The machines I use have a shared FS which has the filelock problem as well as a local one that does not. Using some env vars (HF_HOME, which controls both models and datasets, and HF_DATASETS_OFFLINE) for both transformers and datasets library one can influence where these downloads happen, and whether the locks get t... | I'm downloading a dataset successfully with
`load_dataset("wikitext", "wikitext-2-raw-v1")`
But when I attempt to cache it on an external volume, it hangs indefinitely:
`load_dataset("wikitext", "wikitext-2-raw-v1", cache_dir="/fsx") # /fsx is an external volume mount`
The filesystem when hanging looks like thi... | 96 | [Bug] FileLock dependency incompatible with filesystem
I'm downloading a dataset successfully with
`load_dataset("wikitext", "wikitext-2-raw-v1")`
But when I attempt to cache it on an external volume, it hangs indefinitely:
`load_dataset("wikitext", "wikitext-2-raw-v1", cache_dir="/fsx") # /fsx is an external vo... | [
-0.011772923171520233,
-0.11260530352592468,
0.036588866263628006,
0.1368192434310913,
0.15281780064105988,
0.14302775263786316,
0.4243270754814148,
0.011899184435606003,
0.6362172365188599,
-0.07410603761672974,
0.012147091329097748,
-0.16300678253173828,
0.03955768048763275,
-0.213929265... |
https://github.com/huggingface/datasets/issues/329 | [Bug] FileLock dependency incompatible with filesystem | Thanks @orm011 . These filesystems are such a pain. I'll dig around, looks like setting `cache_dir` to a non-lustre filesystem works for `transformers` but not `datasets`. | I'm downloading a dataset successfully with
`load_dataset("wikitext", "wikitext-2-raw-v1")`
But when I attempt to cache it on an external volume, it hangs indefinitely:
`load_dataset("wikitext", "wikitext-2-raw-v1", cache_dir="/fsx") # /fsx is an external volume mount`
The filesystem when hanging looks like thi... | 26 | [Bug] FileLock dependency incompatible with filesystem
I'm downloading a dataset successfully with
`load_dataset("wikitext", "wikitext-2-raw-v1")`
But when I attempt to cache it on an external volume, it hangs indefinitely:
`load_dataset("wikitext", "wikitext-2-raw-v1", cache_dir="/fsx") # /fsx is an external vo... | [
-0.011807598173618317,
0.06888744235038757,
-0.030581938102841377,
0.08284123241901398,
0.15879729390144348,
0.22022587060928345,
0.35836729407310486,
0.02480284497141838,
0.6490548253059387,
-0.09143224358558655,
0.015682140365242958,
-0.029650678858160973,
-0.03373141586780548,
-0.276616... |
https://github.com/huggingface/datasets/issues/329 | [Bug] FileLock dependency incompatible with filesystem | Note I `export HF_HOME=` in the shell prior to running python (I do not use the `cache_dir` argument, I think I ran into similar issues with it, nor `HF_DATASETS_CACHE` , though maybe that works, or maybe you can set it in python prior to importing the library ), and I change no other variables. Then `datasets.load_d... | I'm downloading a dataset successfully with
`load_dataset("wikitext", "wikitext-2-raw-v1")`
But when I attempt to cache it on an external volume, it hangs indefinitely:
`load_dataset("wikitext", "wikitext-2-raw-v1", cache_dir="/fsx") # /fsx is an external volume mount`
The filesystem when hanging looks like thi... | 81 | [Bug] FileLock dependency incompatible with filesystem
I'm downloading a dataset successfully with
`load_dataset("wikitext", "wikitext-2-raw-v1")`
But when I attempt to cache it on an external volume, it hangs indefinitely:
`load_dataset("wikitext", "wikitext-2-raw-v1", cache_dir="/fsx") # /fsx is an external vo... | [
-0.03621460497379303,
0.09330640733242035,
0.005952887237071991,
0.018699772655963898,
0.11465820670127869,
0.24775221943855286,
0.3859681189060211,
0.09870736300945282,
0.7011548280715942,
-0.08528017997741699,
0.0002230547834187746,
-0.06770867109298706,
0.03757619485259056,
-0.280363798... |
https://github.com/huggingface/datasets/issues/329 | [Bug] FileLock dependency incompatible with filesystem | I am using a shared cluster with a lustre system that I can't change. I am unable to download or load datsets onto the filesystem because of file lock. @thomwolf can this issue be reopened? | I'm downloading a dataset successfully with
`load_dataset("wikitext", "wikitext-2-raw-v1")`
But when I attempt to cache it on an external volume, it hangs indefinitely:
`load_dataset("wikitext", "wikitext-2-raw-v1", cache_dir="/fsx") # /fsx is an external volume mount`
The filesystem when hanging looks like thi... | 35 | [Bug] FileLock dependency incompatible with filesystem
I'm downloading a dataset successfully with
`load_dataset("wikitext", "wikitext-2-raw-v1")`
But when I attempt to cache it on an external volume, it hangs indefinitely:
`load_dataset("wikitext", "wikitext-2-raw-v1", cache_dir="/fsx") # /fsx is an external vo... | [
-0.03016287088394165,
-0.021740294992923737,
-0.0299295112490654,
0.1396872103214264,
0.046884503215551376,
0.20129162073135376,
0.33957213163375854,
0.014964722096920013,
0.7149495482444763,
-0.0786730945110321,
0.07439523190259933,
-0.027142448350787163,
-0.029201705008745193,
-0.2727197... |
https://github.com/huggingface/datasets/issues/328 | Fork dataset | To be able to generate the Arrow dataset you need to either use our csv or json utilities `load_dataset("json", data_files=my_json_files)` OR write your own custom dataset script (you can find some inspiration from the [squad](https://github.com/huggingface/nlp/blob/master/datasets/squad/squad.py) script for example). ... | We have a multi-task learning model training I'm trying to convert to using the Arrow-based nlp dataset.
We're currently training a custom TensorFlow model but the nlp paradigm should be a bridge for us to be able to use the wealth of pre-trained models in Transformers.
Our preprocessing flow parses raw text and... | 72 | Fork dataset
We have a multi-task learning model training I'm trying to convert to using the Arrow-based nlp dataset.
We're currently training a custom TensorFlow model but the nlp paradigm should be a bridge for us to be able to use the wealth of pre-trained models in Transformers.
Our preprocessing flow pars... | [
-0.24039007723331451,
-0.15193112194538116,
-0.0046680644154548645,
0.13880124688148499,
-0.19918325543403625,
0.166174978017807,
0.08553393930196762,
0.3236580491065979,
0.19925619661808014,
-0.14080895483493805,
0.0522022545337677,
0.6768646836280823,
-0.4008253216743469,
0.1611249148845... |
https://github.com/huggingface/datasets/issues/328 | Fork dataset | Thanks for the helpful advice, @lhoestq -- I wasn't quite able to get the json recipe working -
```
~/.virtualenvs/inv-text2struct/lib/python3.6/site-packages/pyarrow/ipc.py in __init__(self, source)
60
61 def __init__(self, source):
---> 62 self._open(source)
63
64
~/.vir... | We have a multi-task learning model training I'm trying to convert to using the Arrow-based nlp dataset.
We're currently training a custom TensorFlow model but the nlp paradigm should be a bridge for us to be able to use the wealth of pre-trained models in Transformers.
Our preprocessing flow parses raw text and... | 87 | Fork dataset
We have a multi-task learning model training I'm trying to convert to using the Arrow-based nlp dataset.
We're currently training a custom TensorFlow model but the nlp paradigm should be a bridge for us to be able to use the wealth of pre-trained models in Transformers.
Our preprocessing flow pars... | [
-0.2709978222846985,
-0.17086929082870483,
-0.014252671040594578,
0.24242204427719116,
-0.09015285968780518,
0.03737718611955643,
0.07227937877178192,
0.3559104800224304,
0.021982531994581223,
-0.15691523253917694,
0.05683229863643646,
0.6875868439674377,
-0.31585821509361267,
0.0238235350... |
https://github.com/huggingface/datasets/issues/328 | Fork dataset | Thanks this answers my question. I think the issue I was having using the json loader were due to using gzipped jsonl files.
The error I get now is :
```
Using custom data configuration test
---------------------------------------------------------------------------
ValueError ... | We have a multi-task learning model training I'm trying to convert to using the Arrow-based nlp dataset.
We're currently training a custom TensorFlow model but the nlp paradigm should be a bridge for us to be able to use the wealth of pre-trained models in Transformers.
Our preprocessing flow parses raw text and... | 324 | Fork dataset
We have a multi-task learning model training I'm trying to convert to using the Arrow-based nlp dataset.
We're currently training a custom TensorFlow model but the nlp paradigm should be a bridge for us to be able to use the wealth of pre-trained models in Transformers.
Our preprocessing flow pars... | [
-0.2616364061832428,
-0.20614203810691833,
0.03750167414546013,
0.3549155592918396,
-0.16824215650558472,
0.1174597516655922,
0.14335043728351593,
0.47420835494995117,
0.30133935809135437,
-0.025156378746032715,
-0.04132690280675888,
0.7794691324234009,
-0.4380645453929901,
-0.000689070671... |
https://github.com/huggingface/datasets/issues/328 | Fork dataset | I'll close this -- It's still unclear how to go about troubleshooting the json example as I mentioned above. If I decide it's worth the trouble, I'll create another issue, or wait for a better support for using nlp for making custom data-loaders. | We have a multi-task learning model training I'm trying to convert to using the Arrow-based nlp dataset.
We're currently training a custom TensorFlow model but the nlp paradigm should be a bridge for us to be able to use the wealth of pre-trained models in Transformers.
Our preprocessing flow parses raw text and... | 43 | Fork dataset
We have a multi-task learning model training I'm trying to convert to using the Arrow-based nlp dataset.
We're currently training a custom TensorFlow model but the nlp paradigm should be a bridge for us to be able to use the wealth of pre-trained models in Transformers.
Our preprocessing flow pars... | [
-0.22559243440628052,
-0.2139756977558136,
0.013712126761674881,
0.18746061623096466,
-0.28556931018829346,
0.10272865742444992,
0.08597820997238159,
0.3090709149837494,
0.26297426223754883,
-0.13470236957073212,
0.07713841646909714,
0.6757094860076904,
-0.37672412395477295,
0.130016028881... |
https://github.com/huggingface/datasets/issues/326 | Large dataset in Squad2-format | I'm pretty sure you can get some inspiration from the squad_v2 script. It looks like the dataset is quite big so it will take some time for the users to generate it, but it should be reasonable.
Also you are saying that you are still making the dataset grow in size right ?
It's probably good practice to let the use... | At the moment we are building an large question answering dataset and think about sharing it with the huggingface community.
Caused the computing power we splitted it into multiple tiles, but they are all in the same format.
Right now the most important facts about are this:
- Contexts: 1.047.671
- questions: 1.677... | 121 | Large dataset in Squad2-format
At the moment we are building an large question answering dataset and think about sharing it with the huggingface community.
Caused the computing power we splitted it into multiple tiles, but they are all in the same format.
Right now the most important facts about are this:
- Contex... | [
-0.09977766126394272,
-0.24149999022483826,
-0.11181317269802094,
0.18103185296058655,
0.1970852017402649,
-0.015226934105157852,
0.023252900689840317,
0.4885348975658417,
-0.09128059446811676,
0.18163174390792847,
-0.2126610428094864,
0.05594272166490555,
-0.24527452886104584,
0.329041808... |
https://github.com/huggingface/datasets/issues/326 | Large dataset in Squad2-format | It would also be good if there is any possibility for versioning, I think this way is much better than the dynamic way.
If you mean that part to put the tiles into one is the generation it would take up to 15-20 minutes on home computer hardware.
Are there any compression or optimization algorithms while generating the... | At the moment we are building an large question answering dataset and think about sharing it with the huggingface community.
Caused the computing power we splitted it into multiple tiles, but they are all in the same format.
Right now the most important facts about are this:
- Contexts: 1.047.671
- questions: 1.677... | 93 | Large dataset in Squad2-format
At the moment we are building an large question answering dataset and think about sharing it with the huggingface community.
Caused the computing power we splitted it into multiple tiles, but they are all in the same format.
Right now the most important facts about are this:
- Contex... | [
-0.07837636023759842,
-0.33690345287323,
-0.1021033227443695,
0.3794311285018921,
0.2913864254951477,
-0.09671299904584885,
-0.1217113733291626,
0.5122985243797302,
-0.027225544676184654,
0.19465230405330658,
-0.2856156527996063,
0.06141316518187523,
-0.2502670884132385,
0.2593445181846618... |
https://github.com/huggingface/datasets/issues/326 | Large dataset in Squad2-format | 15-20 minutes is fine !
Also there's no RAM limitations as we save to disk every 1000 elements while generating the dataset by default.
After generation, the dataset is ready to use with (again) no RAM limitations as we do memory-mapping. | At the moment we are building an large question answering dataset and think about sharing it with the huggingface community.
Caused the computing power we splitted it into multiple tiles, but they are all in the same format.
Right now the most important facts about are this:
- Contexts: 1.047.671
- questions: 1.677... | 41 | Large dataset in Squad2-format
At the moment we are building an large question answering dataset and think about sharing it with the huggingface community.
Caused the computing power we splitted it into multiple tiles, but they are all in the same format.
Right now the most important facts about are this:
- Contex... | [
-0.14090381562709808,
-0.28052520751953125,
-0.10055729746818542,
0.2925081253051758,
0.23439905047416687,
-0.0667673647403717,
-0.006627250462770462,
0.48271888494491577,
-0.004641016945242882,
0.10125841200351715,
-0.28785791993141174,
0.017062775790691376,
-0.24039669334888458,
0.270600... |
https://github.com/huggingface/datasets/issues/326 | Large dataset in Squad2-format | Wow, that sounds pretty cool.
Actually I have the problem of running out of memory while tokenization on our local machine.
That wouldn't happen again, would it ? | At the moment we are building an large question answering dataset and think about sharing it with the huggingface community.
Caused the computing power we splitted it into multiple tiles, but they are all in the same format.
Right now the most important facts about are this:
- Contexts: 1.047.671
- questions: 1.677... | 28 | Large dataset in Squad2-format
At the moment we are building an large question answering dataset and think about sharing it with the huggingface community.
Caused the computing power we splitted it into multiple tiles, but they are all in the same format.
Right now the most important facts about are this:
- Contex... | [
-0.004863277077674866,
-0.23592394590377808,
-0.06830370426177979,
0.29932230710983276,
0.2564297318458557,
-0.11595623195171356,
0.019137907773256302,
0.46602293848991394,
-0.09819957613945007,
0.1532813012599945,
-0.25992703437805176,
0.029207782819867134,
-0.24323512613773346,
0.2509178... |
https://github.com/huggingface/datasets/issues/326 | Large dataset in Squad2-format | You can do the tokenization step using `my_tokenized_dataset = my_dataset.map(my_tokenize_function)` that writes the tokenized texts on disk as well. And then `my_tokenized_dataset` will be a memory-mapped dataset too, so you should be fine :) | At the moment we are building an large question answering dataset and think about sharing it with the huggingface community.
Caused the computing power we splitted it into multiple tiles, but they are all in the same format.
Right now the most important facts about are this:
- Contexts: 1.047.671
- questions: 1.677... | 34 | Large dataset in Squad2-format
At the moment we are building an large question answering dataset and think about sharing it with the huggingface community.
Caused the computing power we splitted it into multiple tiles, but they are all in the same format.
Right now the most important facts about are this:
- Contex... | [
-0.0358472503721714,
-0.24777738749980927,
-0.05869962275028229,
0.24638879299163818,
0.2725088596343994,
-0.08004845678806305,
0.03257644176483154,
0.41994839906692505,
-0.12915077805519104,
0.04069605469703674,
-0.26981914043426514,
0.10341095179319382,
-0.2411339432001114,
0.21754868328... |
https://github.com/huggingface/datasets/issues/326 | Large dataset in Squad2-format | In your training loop, loading the tokenized texts is going to be fast and pretty much negligible compared to a forward pass. You shouldn't expect any slow down. | At the moment we are building an large question answering dataset and think about sharing it with the huggingface community.
Caused the computing power we splitted it into multiple tiles, but they are all in the same format.
Right now the most important facts about are this:
- Contexts: 1.047.671
- questions: 1.677... | 28 | Large dataset in Squad2-format
At the moment we are building an large question answering dataset and think about sharing it with the huggingface community.
Caused the computing power we splitted it into multiple tiles, but they are all in the same format.
Right now the most important facts about are this:
- Contex... | [
-0.09984652698040009,
-0.34005001187324524,
-0.07225452363491058,
0.274052232503891,
0.23301471769809723,
-0.13411931693553925,
0.05346614867448807,
0.4866882562637329,
-0.08115601539611816,
0.03897847980260849,
-0.2371414601802826,
0.07050888240337372,
-0.256649374961853,
0.30955618619918... |
https://github.com/huggingface/datasets/issues/324 | Error when calculating glue score | The glue metric for cola is a metric for classification. It expects label ids as integers as inputs. | I was trying glue score along with other metrics here. But glue gives me this error;
```
import nlp
glue_metric = nlp.load_metric('glue',name="cola")
glue_score = glue_metric.compute(predictions, references)
```
```
---------------------------------------------------------------------------
--------------... | 18 | Error when calculating glue score
I was trying glue score along with other metrics here. But glue gives me this error;
```
import nlp
glue_metric = nlp.load_metric('glue',name="cola")
glue_score = glue_metric.compute(predictions, references)
```
```
-------------------------------------------------------... | [
-0.16945281624794006,
-0.23958125710487366,
-0.03047320619225502,
0.16115909814834595,
0.2367406189441681,
-0.049165286123752594,
0.10730020701885223,
0.3623579144477844,
0.433027446269989,
0.05162825435400009,
-0.3168230354785919,
0.27634739875793457,
-0.07980765402317047,
-0.024109499529... |
https://github.com/huggingface/datasets/issues/324 | Error when calculating glue score | I want to evaluate a sentence pair whether they are semantically equivalent, so I used MRPC and it gives the same error, does that mean we have to encode the sentences and parse as input?
using BertTokenizer;
```
encoded_reference=tokenizer.encode(reference, add_special_tokens=False)
encoded_prediction=tokenizer.... | I was trying glue score along with other metrics here. But glue gives me this error;
```
import nlp
glue_metric = nlp.load_metric('glue',name="cola")
glue_score = glue_metric.compute(predictions, references)
```
```
---------------------------------------------------------------------------
--------------... | 297 | Error when calculating glue score
I was trying glue score along with other metrics here. But glue gives me this error;
```
import nlp
glue_metric = nlp.load_metric('glue',name="cola")
glue_score = glue_metric.compute(predictions, references)
```
```
-------------------------------------------------------... | [
-0.16945281624794006,
-0.23958125710487366,
-0.03047320619225502,
0.16115909814834595,
0.2367406189441681,
-0.049165286123752594,
0.10730020701885223,
0.3623579144477844,
0.433027446269989,
0.05162825435400009,
-0.3168230354785919,
0.27634739875793457,
-0.07980765402317047,
-0.024109499529... |
https://github.com/huggingface/datasets/issues/324 | Error when calculating glue score | MRPC is also a binary classification task, so its metric is a binary classification metric.
To evaluate if pairs of sentences are semantically equivalent, maybe you could take a look at models that compute if one sentence entails the other or not (typically the kinds of model that could work well on the MRPC task). | I was trying glue score along with other metrics here. But glue gives me this error;
```
import nlp
glue_metric = nlp.load_metric('glue',name="cola")
glue_score = glue_metric.compute(predictions, references)
```
```
---------------------------------------------------------------------------
--------------... | 55 | Error when calculating glue score
I was trying glue score along with other metrics here. But glue gives me this error;
```
import nlp
glue_metric = nlp.load_metric('glue',name="cola")
glue_score = glue_metric.compute(predictions, references)
```
```
-------------------------------------------------------... | [
-0.16945281624794006,
-0.23958125710487366,
-0.03047320619225502,
0.16115909814834595,
0.2367406189441681,
-0.049165286123752594,
0.10730020701885223,
0.3623579144477844,
0.433027446269989,
0.05162825435400009,
-0.3168230354785919,
0.27634739875793457,
-0.07980765402317047,
-0.024109499529... |
https://github.com/huggingface/datasets/issues/321 | ERROR:root:mwparserfromhell | It looks like it comes from `mwparserfromhell`.
Would it be possible to get the bad `section` that causes this issue ? The `section` string is from `datasets/wikipedia.py:L548` ? You could just add a `try` statement and print the section if the line `section_text.append(section.strip_code().strip())` crashes.
It ... | Hi,
I am trying to download some wikipedia data but I got this error for spanish "es" (but there are maybe some others languages which have the same error I haven't tried all of them ).
`ERROR:root:mwparserfromhell ParseError: This is a bug and should be reported. Info: C tokenizer exited with non-empty token sta... | 65 | ERROR:root:mwparserfromhell
Hi,
I am trying to download some wikipedia data but I got this error for spanish "es" (but there are maybe some others languages which have the same error I haven't tried all of them ).
`ERROR:root:mwparserfromhell ParseError: This is a bug and should be reported. Info: C tokenizer e... | [
-0.2354934811592102,
-0.13825522363185883,
0.055253949016332626,
0.5212735533714294,
0.06802239269018173,
-0.0856214240193367,
-0.134895458817482,
0.30712220072746277,
0.20676663517951965,
0.28884899616241455,
0.3477610945701599,
-0.07156240195035934,
0.05988486856222153,
-0.47903141379356... |
https://github.com/huggingface/datasets/issues/321 | ERROR:root:mwparserfromhell | Hi,
Thank you for you answer.
I have try to print the bad section using `try` and `except`, but it is a bit weird as the error seems to appear 3 times for instance, but the two first error does not print anything (as if the function did not go in the `except` part).
For the third one, I got that (I haven't displa... | Hi,
I am trying to download some wikipedia data but I got this error for spanish "es" (but there are maybe some others languages which have the same error I haven't tried all of them ).
`ERROR:root:mwparserfromhell ParseError: This is a bug and should be reported. Info: C tokenizer exited with non-empty token sta... | 234 | ERROR:root:mwparserfromhell
Hi,
I am trying to download some wikipedia data but I got this error for spanish "es" (but there are maybe some others languages which have the same error I haven't tried all of them ).
`ERROR:root:mwparserfromhell ParseError: This is a bug and should be reported. Info: C tokenizer e... | [
-0.2193867564201355,
-0.22226984798908234,
0.0420609749853611,
0.36096668243408203,
0.0011144783347845078,
-0.05777616798877716,
-0.2417544573545456,
0.20955367386341095,
0.30749455094337463,
0.19320741295814514,
0.3478332757949829,
0.010102486237883568,
0.004232372157275677,
-0.4349838793... |
https://github.com/huggingface/datasets/issues/321 | ERROR:root:mwparserfromhell | The error appears several times as Apache Beam retries to process examples up to 4 times irc.
I just tried to run this text into `mwparserfromhell` but it worked without the issue.
I used this code (from the `wikipedia.py` script):
```python
import mwparserfromhell as parser
import re
import six
raw_conten... | Hi,
I am trying to download some wikipedia data but I got this error for spanish "es" (but there are maybe some others languages which have the same error I haven't tried all of them ).
`ERROR:root:mwparserfromhell ParseError: This is a bug and should be reported. Info: C tokenizer exited with non-empty token sta... | 264 | ERROR:root:mwparserfromhell
Hi,
I am trying to download some wikipedia data but I got this error for spanish "es" (but there are maybe some others languages which have the same error I haven't tried all of them ).
`ERROR:root:mwparserfromhell ParseError: This is a bug and should be reported. Info: C tokenizer e... | [
-0.2439156472682953,
-0.19704169034957886,
0.06708210706710815,
0.3134256899356842,
0.027157023549079895,
-0.03122444450855255,
-0.11951199173927307,
0.21098530292510986,
0.3312293589115143,
0.14182540774345398,
0.32294970750808716,
0.01704266108572483,
0.051209188997745514,
-0.49286463856... |
https://github.com/huggingface/datasets/issues/321 | ERROR:root:mwparserfromhell | Not sure why we're having this issue. Maybe could you get also the file that's causing that ? | Hi,
I am trying to download some wikipedia data but I got this error for spanish "es" (but there are maybe some others languages which have the same error I haven't tried all of them ).
`ERROR:root:mwparserfromhell ParseError: This is a bug and should be reported. Info: C tokenizer exited with non-empty token sta... | 18 | ERROR:root:mwparserfromhell
Hi,
I am trying to download some wikipedia data but I got this error for spanish "es" (but there are maybe some others languages which have the same error I haven't tried all of them ).
`ERROR:root:mwparserfromhell ParseError: This is a bug and should be reported. Info: C tokenizer e... | [
-0.24387389421463013,
-0.15484470129013062,
0.09030304849147797,
0.4575003385543823,
0.12444977462291718,
-0.054339099675416946,
-0.11729739606380463,
0.20664753019809723,
0.33644378185272217,
0.30087152123451233,
0.29528728127479553,
-0.002494564512744546,
0.05222313106060028,
-0.38132801... |
https://github.com/huggingface/datasets/issues/321 | ERROR:root:mwparserfromhell | thanks for your answer.
How can I know which file is causing the issue ?
I am trying to load the spanish wikipedia data. | Hi,
I am trying to download some wikipedia data but I got this error for spanish "es" (but there are maybe some others languages which have the same error I haven't tried all of them ).
`ERROR:root:mwparserfromhell ParseError: This is a bug and should be reported. Info: C tokenizer exited with non-empty token sta... | 24 | ERROR:root:mwparserfromhell
Hi,
I am trying to download some wikipedia data but I got this error for spanish "es" (but there are maybe some others languages which have the same error I haven't tried all of them ).
`ERROR:root:mwparserfromhell ParseError: This is a bug and should be reported. Info: C tokenizer e... | [
-0.38053974509239197,
-0.340488076210022,
0.05695919319987297,
0.6597745418548584,
-0.02394382655620575,
-0.1370640993118286,
-0.22147642076015472,
0.35511940717697144,
0.32051554322242737,
0.2352277934551239,
0.3936889171600342,
-0.04771595448255539,
0.02794928103685379,
-0.45238190889358... |
https://github.com/huggingface/datasets/issues/321 | ERROR:root:mwparserfromhell | Because of the way Apache Beam works we indeed don't have access to the file name at this point in the code.
We'll have to use some tricks I think :p
You can append `filepath` to `title` in `wikipedia.py:L512` for example. [[EDIT: it's L494 my bad]]
Then just do `try:...except:` on the call of `_parse_and_clean_... | Hi,
I am trying to download some wikipedia data but I got this error for spanish "es" (but there are maybe some others languages which have the same error I haven't tried all of them ).
`ERROR:root:mwparserfromhell ParseError: This is a bug and should be reported. Info: C tokenizer exited with non-empty token sta... | 95 | ERROR:root:mwparserfromhell
Hi,
I am trying to download some wikipedia data but I got this error for spanish "es" (but there are maybe some others languages which have the same error I haven't tried all of them ).
`ERROR:root:mwparserfromhell ParseError: This is a bug and should be reported. Info: C tokenizer e... | [
-0.12096287310123444,
-0.07050880789756775,
0.056422214955091476,
0.389180064201355,
0.10341434925794601,
0.017031557857990265,
-0.12144532799720764,
0.3558145761489868,
0.25141623616218567,
0.2640950679779053,
0.21410585939884186,
-0.06851916760206223,
0.04064324125647545,
-0.449892550706... |
https://github.com/huggingface/datasets/issues/321 | ERROR:root:mwparserfromhell | Thanks for your help.
I tried to print the "title" of the document inside the` except (mwparserfromhell.parser.ParserError) as e`,the title displayed was : "Campeonato Mundial de futsal de la AMF 2015". (Wikipedia ES) Is it what you were looking for ? | Hi,
I am trying to download some wikipedia data but I got this error for spanish "es" (but there are maybe some others languages which have the same error I haven't tried all of them ).
`ERROR:root:mwparserfromhell ParseError: This is a bug and should be reported. Info: C tokenizer exited with non-empty token sta... | 41 | ERROR:root:mwparserfromhell
Hi,
I am trying to download some wikipedia data but I got this error for spanish "es" (but there are maybe some others languages which have the same error I haven't tried all of them ).
`ERROR:root:mwparserfromhell ParseError: This is a bug and should be reported. Info: C tokenizer e... | [
-0.09824054688215256,
-0.14601217210292816,
0.05791466310620308,
0.45700687170028687,
0.03996896743774414,
0.08334863930940628,
-0.24785323441028595,
0.19130896031856537,
0.3329883813858032,
0.294177383184433,
0.25431105494499207,
0.10800562053918839,
0.09529786556959152,
-0.34020701050758... |
https://github.com/huggingface/datasets/issues/321 | ERROR:root:mwparserfromhell | Thanks a lot @Shiro-LK !
I was able to reproduce the issue. It comes from [this table on wikipedia](https://es.wikipedia.org/wiki/Campeonato_Mundial_de_futsal_de_la_AMF_2015#Clasificados) that can't be parsed.
The file in which the problem occurs comes from the wikipedia dumps, and it can be downloaded [here](htt... | Hi,
I am trying to download some wikipedia data but I got this error for spanish "es" (but there are maybe some others languages which have the same error I haven't tried all of them ).
`ERROR:root:mwparserfromhell ParseError: This is a bug and should be reported. Info: C tokenizer exited with non-empty token sta... | 248 | ERROR:root:mwparserfromhell
Hi,
I am trying to download some wikipedia data but I got this error for spanish "es" (but there are maybe some others languages which have the same error I haven't tried all of them ).
`ERROR:root:mwparserfromhell ParseError: This is a bug and should be reported. Info: C tokenizer e... | [
-0.10873883962631226,
-0.26876991987228394,
0.05885482206940651,
0.37449944019317627,
0.06803174316883087,
0.05851078778505325,
-0.09784887731075287,
0.40723302960395813,
0.2762869894504547,
0.2615221440792084,
0.16486422717571259,
0.16331365704536438,
0.06970024853944778,
-0.4300827980041... |
https://github.com/huggingface/datasets/issues/320 | Blog Authorship Corpus, Non Matching Splits Sizes Error, nlp viewer | This dataset seems to have a decoding problem that results in inconsistencies in the number of generated examples.
See #215.
That's why we end up with a `NonMatchingSplitsSizesError `. | Selecting `blog_authorship_corpus` in the nlp viewer throws the following error:
```
NonMatchingSplitsSizesError: [{'expected': SplitInfo(name='train', num_bytes=610252351, num_examples=532812, dataset_name='blog_authorship_corpus'), 'recorded': SplitInfo(name='train', num_bytes=614706451, num_examples=535568, dat... | 29 | Blog Authorship Corpus, Non Matching Splits Sizes Error, nlp viewer
Selecting `blog_authorship_corpus` in the nlp viewer throws the following error:
```
NonMatchingSplitsSizesError: [{'expected': SplitInfo(name='train', num_bytes=610252351, num_examples=532812, dataset_name='blog_authorship_corpus'), 'recorded':... | [
-0.23413603007793427,
0.17466294765472412,
0.05287710204720497,
0.22421227395534515,
-0.054631300270557404,
-0.027442969381809235,
0.037543538957834244,
0.5022202730178833,
-0.2198076993227005,
0.19498035311698914,
-0.08825415372848511,
-0.07330672442913055,
-0.06501486152410507,
-0.021802... |
https://github.com/huggingface/datasets/issues/319 | Nested sequences with dicts | Oh yes, this is a backward compatibility feature with tensorflow_dataset in which a `Sequence` or `dict` is converted in a `dict` of `lists`, unfortunately it is not very intuitive, see here: https://github.com/huggingface/nlp/blob/master/src/nlp/features.py#L409
To avoid this behavior, you can just define the list ... | Am pretty much finished [adding a dataset](https://github.com/ghomasHudson/nlp/blob/DocRED/datasets/docred/docred.py) for [DocRED](https://github.com/thunlp/DocRED), but am getting an error when trying to add a nested `nlp.features.sequence(nlp.features.sequence({key:value,...}))`.
The original data is in this form... | 88 | Nested sequences with dicts
Am pretty much finished [adding a dataset](https://github.com/ghomasHudson/nlp/blob/DocRED/datasets/docred/docred.py) for [DocRED](https://github.com/thunlp/DocRED), but am getting an error when trying to add a nested `nlp.features.sequence(nlp.features.sequence({key:value,...}))`.
The... | [
0.11176292598247528,
0.0667421817779541,
-0.0714067667722702,
-0.02101646363735199,
-0.10916800796985626,
-0.041018303483724594,
0.19878682494163513,
0.169729083776474,
-0.08717702329158783,
-0.40041184425354004,
0.28796878457069397,
0.3965296745300293,
-0.08983413130044937,
0.203804865479... |
https://github.com/huggingface/datasets/issues/317 | Adding a dataset with multiple subtasks | For one dataset you can have different configurations that each have their own `nlp.Features`.
We imagine having one configuration per subtask for example.
They are loaded with `nlp.load_dataset("my_dataset", "my_config")`.
For example the `glue` dataset has many configurations. It is a bit different from your cas... | I intent to add the datasets of the MT Quality Estimation shared tasks to `nlp`. However, they have different subtasks -- such as word-level, sentence-level and document-level quality estimation, each of which having different language pairs, and some of the data reused in different subtasks.
For example, in [QE 201... | 66 | Adding a dataset with multiple subtasks
I intent to add the datasets of the MT Quality Estimation shared tasks to `nlp`. However, they have different subtasks -- such as word-level, sentence-level and document-level quality estimation, each of which having different language pairs, and some of the data reused in diff... | [
-0.07099547237157822,
-0.1814526915550232,
-0.043702878057956696,
0.38491493463516235,
-0.21697475016117096,
0.09465894848108292,
0.35820919275283813,
-0.13353407382965088,
0.12713974714279175,
-0.1552470475435257,
-0.2816210687160492,
0.1489303708076477,
-0.15718218684196472,
0.6809746026... |
https://github.com/huggingface/datasets/issues/315 | [Question] Best way to batch a large dataset? | Update: I think I've found a solution.
```python
output_types = {"input_ids": tf.int64, "token_type_ids": tf.int64, "attention_mask": tf.int64}
def train_dataset_gen():
for i in range(len(train_dataset)):
yield train_dataset[i]
tf_dataset = tf.data.Dataset.from_generator(train_dataset_gen, output_ty... | I'm training on large datasets such as Wikipedia and BookCorpus. Following the instructions in [the tutorial notebook](https://colab.research.google.com/github/huggingface/nlp/blob/master/notebooks/Overview.ipynb), I see the following recommended for TensorFlow:
```python
train_tf_dataset = train_tf_dataset.filter(... | 48 | [Question] Best way to batch a large dataset?
I'm training on large datasets such as Wikipedia and BookCorpus. Following the instructions in [the tutorial notebook](https://colab.research.google.com/github/huggingface/nlp/blob/master/notebooks/Overview.ipynb), I see the following recommended for TensorFlow:
```pyt... | [
-0.30983999371528625,
-0.3549017310142517,
0.040070291608572006,
-0.13144156336784363,
0.04347967356443405,
0.17069199681282043,
0.5454060435295105,
0.39459624886512756,
0.18538275361061096,
0.07585114240646362,
-0.010929248295724392,
0.09057066589593887,
-0.0458250567317009,
0.16368539631... |
https://github.com/huggingface/datasets/issues/315 | [Question] Best way to batch a large dataset? | Yes this is the current best solution. We should probably show it in the tutorial notebook.
Note that this solution unfortunately doesn't allow to train on TPUs (yet). See #193 | I'm training on large datasets such as Wikipedia and BookCorpus. Following the instructions in [the tutorial notebook](https://colab.research.google.com/github/huggingface/nlp/blob/master/notebooks/Overview.ipynb), I see the following recommended for TensorFlow:
```python
train_tf_dataset = train_tf_dataset.filter(... | 30 | [Question] Best way to batch a large dataset?
I'm training on large datasets such as Wikipedia and BookCorpus. Following the instructions in [the tutorial notebook](https://colab.research.google.com/github/huggingface/nlp/blob/master/notebooks/Overview.ipynb), I see the following recommended for TensorFlow:
```pyt... | [
-0.30983999371528625,
-0.3549017310142517,
0.040070291608572006,
-0.13144156336784363,
0.04347967356443405,
0.17069199681282043,
0.5454060435295105,
0.39459624886512756,
0.18538275361061096,
0.07585114240646362,
-0.010929248295724392,
0.09057066589593887,
-0.0458250567317009,
0.16368539631... |
https://github.com/huggingface/datasets/issues/315 | [Question] Best way to batch a large dataset? | This approach still seems quite slow. When using TFRecords with a similar training loop, I get ~3.0-3.5 it/s on multi-node, multi-GPU training. I notice a pretty severe performance regression when scaling, with observed performance numbers. Since the allreduce step takes less than 100ms/it and I've achieved 80% scaling... | I'm training on large datasets such as Wikipedia and BookCorpus. Following the instructions in [the tutorial notebook](https://colab.research.google.com/github/huggingface/nlp/blob/master/notebooks/Overview.ipynb), I see the following recommended for TensorFlow:
```python
train_tf_dataset = train_tf_dataset.filter(... | 146 | [Question] Best way to batch a large dataset?
I'm training on large datasets such as Wikipedia and BookCorpus. Following the instructions in [the tutorial notebook](https://colab.research.google.com/github/huggingface/nlp/blob/master/notebooks/Overview.ipynb), I see the following recommended for TensorFlow:
```pyt... | [
-0.30983999371528625,
-0.3549017310142517,
0.040070291608572006,
-0.13144156336784363,
0.04347967356443405,
0.17069199681282043,
0.5454060435295105,
0.39459624886512756,
0.18538275361061096,
0.07585114240646362,
-0.010929248295724392,
0.09057066589593887,
-0.0458250567317009,
0.16368539631... |
https://github.com/huggingface/datasets/issues/315 | [Question] Best way to batch a large dataset? | An interesting alternative to investigate here would be to use the tf.io library which has some support for Arrow to TF conversion: https://www.tensorflow.org/io/api_docs/python/tfio/arrow/ArrowDataset
There are quite a few types supported, including lists so if the unsupported columns are dropped then we could mayb... | I'm training on large datasets such as Wikipedia and BookCorpus. Following the instructions in [the tutorial notebook](https://colab.research.google.com/github/huggingface/nlp/blob/master/notebooks/Overview.ipynb), I see the following recommended for TensorFlow:
```python
train_tf_dataset = train_tf_dataset.filter(... | 77 | [Question] Best way to batch a large dataset?
I'm training on large datasets such as Wikipedia and BookCorpus. Following the instructions in [the tutorial notebook](https://colab.research.google.com/github/huggingface/nlp/blob/master/notebooks/Overview.ipynb), I see the following recommended for TensorFlow:
```pyt... | [
-0.30983999371528625,
-0.3549017310142517,
0.040070291608572006,
-0.13144156336784363,
0.04347967356443405,
0.17069199681282043,
0.5454060435295105,
0.39459624886512756,
0.18538275361061096,
0.07585114240646362,
-0.010929248295724392,
0.09057066589593887,
-0.0458250567317009,
0.16368539631... |
https://github.com/huggingface/datasets/issues/315 | [Question] Best way to batch a large dataset? | Interesting. There's no support for strings, but it does enable int and floats so that would work for tokenized inputs.
ArrowStreamDataset requires loading from a "record batch iterator", which can be instantiated from in-memory arrays as described here: https://arrow.apache.org/docs/python/ipc.html.
But the nl... | I'm training on large datasets such as Wikipedia and BookCorpus. Following the instructions in [the tutorial notebook](https://colab.research.google.com/github/huggingface/nlp/blob/master/notebooks/Overview.ipynb), I see the following recommended for TensorFlow:
```python
train_tf_dataset = train_tf_dataset.filter(... | 86 | [Question] Best way to batch a large dataset?
I'm training on large datasets such as Wikipedia and BookCorpus. Following the instructions in [the tutorial notebook](https://colab.research.google.com/github/huggingface/nlp/blob/master/notebooks/Overview.ipynb), I see the following recommended for TensorFlow:
```pyt... | [
-0.30983999371528625,
-0.3549017310142517,
0.040070291608572006,
-0.13144156336784363,
0.04347967356443405,
0.17069199681282043,
0.5454060435295105,
0.39459624886512756,
0.18538275361061096,
0.07585114240646362,
-0.010929248295724392,
0.09057066589593887,
-0.0458250567317009,
0.16368539631... |
https://github.com/huggingface/datasets/issues/315 | [Question] Best way to batch a large dataset? | Also note that since #322 it is now possible to do
```python
ids = [1, 10, 42, 100]
batch = dataset[ids]
```
From my experience it is quite fast but it can take lots of memory for large batches (haven't played that much with it).
Let me know if you think there could be a better way to implement it. (current code ... | I'm training on large datasets such as Wikipedia and BookCorpus. Following the instructions in [the tutorial notebook](https://colab.research.google.com/github/huggingface/nlp/blob/master/notebooks/Overview.ipynb), I see the following recommended for TensorFlow:
```python
train_tf_dataset = train_tf_dataset.filter(... | 64 | [Question] Best way to batch a large dataset?
I'm training on large datasets such as Wikipedia and BookCorpus. Following the instructions in [the tutorial notebook](https://colab.research.google.com/github/huggingface/nlp/blob/master/notebooks/Overview.ipynb), I see the following recommended for TensorFlow:
```pyt... | [
-0.30983999371528625,
-0.3549017310142517,
0.040070291608572006,
-0.13144156336784363,
0.04347967356443405,
0.17069199681282043,
0.5454060435295105,
0.39459624886512756,
0.18538275361061096,
0.07585114240646362,
-0.010929248295724392,
0.09057066589593887,
-0.0458250567317009,
0.16368539631... |
https://github.com/huggingface/datasets/issues/315 | [Question] Best way to batch a large dataset? | Thanks @lhoestq! That format is much better to work with.
I put together a benchmarking script. This doesn't measure the CPU-to-GPU efficiency, nor how it scales with multi-GPU multi-node training where many processes are making the same demands on the same dataset. But it does show some interesting results:
```p... | I'm training on large datasets such as Wikipedia and BookCorpus. Following the instructions in [the tutorial notebook](https://colab.research.google.com/github/huggingface/nlp/blob/master/notebooks/Overview.ipynb), I see the following recommended for TensorFlow:
```python
train_tf_dataset = train_tf_dataset.filter(... | 285 | [Question] Best way to batch a large dataset?
I'm training on large datasets such as Wikipedia and BookCorpus. Following the instructions in [the tutorial notebook](https://colab.research.google.com/github/huggingface/nlp/blob/master/notebooks/Overview.ipynb), I see the following recommended for TensorFlow:
```pyt... | [
-0.30983999371528625,
-0.3549017310142517,
0.040070291608572006,
-0.13144156336784363,
0.04347967356443405,
0.17069199681282043,
0.5454060435295105,
0.39459624886512756,
0.18538275361061096,
0.07585114240646362,
-0.010929248295724392,
0.09057066589593887,
-0.0458250567317009,
0.16368539631... |
https://github.com/huggingface/datasets/issues/315 | [Question] Best way to batch a large dataset? | Hey @jarednielsen
Thanks for this very interesting analysis!! IMHO to read text data one should use `tf.data.TextLineDataset`. It would be interesting to compare what you have done with simply load with a `TextLineDataset` and see if there is a difference.
A good example can be found here https://www.tensorflow.... | I'm training on large datasets such as Wikipedia and BookCorpus. Following the instructions in [the tutorial notebook](https://colab.research.google.com/github/huggingface/nlp/blob/master/notebooks/Overview.ipynb), I see the following recommended for TensorFlow:
```python
train_tf_dataset = train_tf_dataset.filter(... | 48 | [Question] Best way to batch a large dataset?
I'm training on large datasets such as Wikipedia and BookCorpus. Following the instructions in [the tutorial notebook](https://colab.research.google.com/github/huggingface/nlp/blob/master/notebooks/Overview.ipynb), I see the following recommended for TensorFlow:
```pyt... | [
-0.30983999371528625,
-0.3549017310142517,
0.040070291608572006,
-0.13144156336784363,
0.04347967356443405,
0.17069199681282043,
0.5454060435295105,
0.39459624886512756,
0.18538275361061096,
0.07585114240646362,
-0.010929248295724392,
0.09057066589593887,
-0.0458250567317009,
0.16368539631... |
https://github.com/huggingface/datasets/issues/315 | [Question] Best way to batch a large dataset? | Thanks! I'm not actually loading in raw text data, that was just the synthetic data I created for this benchmark. A more realistic use case would be a dataset of tokenized examples, which would be a dict of lists of integers. TensorFlow's TextLineDataset greedily loads the dataset into the graph itself, which can lead ... | I'm training on large datasets such as Wikipedia and BookCorpus. Following the instructions in [the tutorial notebook](https://colab.research.google.com/github/huggingface/nlp/blob/master/notebooks/Overview.ipynb), I see the following recommended for TensorFlow:
```python
train_tf_dataset = train_tf_dataset.filter(... | 98 | [Question] Best way to batch a large dataset?
I'm training on large datasets such as Wikipedia and BookCorpus. Following the instructions in [the tutorial notebook](https://colab.research.google.com/github/huggingface/nlp/blob/master/notebooks/Overview.ipynb), I see the following recommended for TensorFlow:
```pyt... | [
-0.30983999371528625,
-0.3549017310142517,
0.040070291608572006,
-0.13144156336784363,
0.04347967356443405,
0.17069199681282043,
0.5454060435295105,
0.39459624886512756,
0.18538275361061096,
0.07585114240646362,
-0.010929248295724392,
0.09057066589593887,
-0.0458250567317009,
0.16368539631... |
https://github.com/huggingface/datasets/issues/315 | [Question] Best way to batch a large dataset? | Sorry, I think I badly expressed myself, my bad. What I suggested is to compare with the usual loading textual data in pure TF with `TextLineDataset` with `nlp`. I know it is not recommended with very large datasets to use it, but I was curious to see how it behaves compared to a processing with `nlp` on smaller datase... | I'm training on large datasets such as Wikipedia and BookCorpus. Following the instructions in [the tutorial notebook](https://colab.research.google.com/github/huggingface/nlp/blob/master/notebooks/Overview.ipynb), I see the following recommended for TensorFlow:
```python
train_tf_dataset = train_tf_dataset.filter(... | 68 | [Question] Best way to batch a large dataset?
I'm training on large datasets such as Wikipedia and BookCorpus. Following the instructions in [the tutorial notebook](https://colab.research.google.com/github/huggingface/nlp/blob/master/notebooks/Overview.ipynb), I see the following recommended for TensorFlow:
```pyt... | [
-0.30983999371528625,
-0.3549017310142517,
0.040070291608572006,
-0.13144156336784363,
0.04347967356443405,
0.17069199681282043,
0.5454060435295105,
0.39459624886512756,
0.18538275361061096,
0.07585114240646362,
-0.010929248295724392,
0.09057066589593887,
-0.0458250567317009,
0.16368539631... |
https://github.com/huggingface/datasets/issues/312 | [Feature request] Add `shard()` method to dataset | Hi Jared,
Interesting, thanks for raising this question. You can also do that after loading with `dataset.select()` or `dataset.filter()` which let you keep only a specific subset of rows in a dataset.
What is your use-case for sharding? | Currently, to shard a dataset into 10 pieces on different ranks, you can run
```python
rank = 3 # for example
size = 10
dataset = nlp.load_dataset('wikitext', 'wikitext-2-raw-v1', split=f"train[{rank*10}%:{(rank+1)*10}%]")
```
However, this breaks down if you have a number of ranks that doesn't divide cleanly... | 38 | [Feature request] Add `shard()` method to dataset
Currently, to shard a dataset into 10 pieces on different ranks, you can run
```python
rank = 3 # for example
size = 10
dataset = nlp.load_dataset('wikitext', 'wikitext-2-raw-v1', split=f"train[{rank*10}%:{(rank+1)*10}%]")
```
However, this breaks down if yo... | [
-0.09021862596273422,
0.03354489803314209,
-0.13852675259113312,
-0.010493777692317963,
-0.12532028555870056,
-0.039162348955869675,
0.4583161175251007,
0.41311776638031006,
-0.11074130237102509,
0.18661761283874512,
0.19244475662708282,
0.2701655328273773,
-0.20456524193286896,
0.33234158... |
https://github.com/huggingface/datasets/issues/312 | [Feature request] Add `shard()` method to dataset | Thanks for the pointer to those functions! It's still a little more verbose since you have to manually calculate which ids each rank would keep, but definitely works.
My use case is multi-node, multi-GPU training and avoiding global batches of duplicate elements. I'm using horovod. You can shuffle indices, or set ra... | Currently, to shard a dataset into 10 pieces on different ranks, you can run
```python
rank = 3 # for example
size = 10
dataset = nlp.load_dataset('wikitext', 'wikitext-2-raw-v1', split=f"train[{rank*10}%:{(rank+1)*10}%]")
```
However, this breaks down if you have a number of ranks that doesn't divide cleanly... | 71 | [Feature request] Add `shard()` method to dataset
Currently, to shard a dataset into 10 pieces on different ranks, you can run
```python
rank = 3 # for example
size = 10
dataset = nlp.load_dataset('wikitext', 'wikitext-2-raw-v1', split=f"train[{rank*10}%:{(rank+1)*10}%]")
```
However, this breaks down if yo... | [
-0.09039951860904694,
0.016252130270004272,
-0.12798944115638733,
-0.07298393547534943,
-0.058893002569675446,
-0.07297815382480621,
0.4991563558578491,
0.40996676683425903,
-0.05547717213630676,
0.2099149227142334,
0.23607973754405975,
0.3211589455604553,
-0.3143123686313629,
0.4003582000... |
https://github.com/huggingface/datasets/issues/302 | Question - Sign Language Datasets | Even more complicating -
As I see it, datasets can have "addons".
For example, the WebNLG dataset is a dataset for data-to-text. However, a work of mine and other works enriched this dataset with text plans / underlying text structures. In that case, I see a need to load the dataset "WebNLG" with "plans" addon.
... | An emerging field in NLP is SLP - sign language processing.
I was wondering about adding datasets here, specifically because it's shaping up to be large and easily usable.
The metrics for sign language to text translation are the same.
So, what do you think about (me, or others) adding datasets here?
An exa... | 131 | Question - Sign Language Datasets
An emerging field in NLP is SLP - sign language processing.
I was wondering about adding datasets here, specifically because it's shaping up to be large and easily usable.
The metrics for sign language to text translation are the same.
So, what do you think about (me, or other... | [
-0.327766478061676,
0.45256537199020386,
-0.07328386604785919,
0.012339726090431213,
-0.07884039729833603,
-0.011533364653587341,
0.22011063992977142,
0.26100069284439087,
0.20139634609222412,
-0.3687557578086853,
0.08788055926561356,
0.29479268193244934,
-0.3939642906188965,
0.28343367576... |
https://github.com/huggingface/datasets/issues/302 | Question - Sign Language Datasets | This is a really cool idea !
The example for data objects you gave for the RWTH-PHOENIX-Weather 2014 T dataset can totally fit inside the library.
For your point about formats like `ilex`, `eaf`, or `srt`, it is possible to use any library in your dataset script.
However most user probably won't need these librari... | An emerging field in NLP is SLP - sign language processing.
I was wondering about adding datasets here, specifically because it's shaping up to be large and easily usable.
The metrics for sign language to text translation are the same.
So, what do you think about (me, or others) adding datasets here?
An exa... | 139 | Question - Sign Language Datasets
An emerging field in NLP is SLP - sign language processing.
I was wondering about adding datasets here, specifically because it's shaping up to be large and easily usable.
The metrics for sign language to text translation are the same.
So, what do you think about (me, or other... | [
-0.23924104869365692,
0.43220704793930054,
-0.06781695038080215,
0.026453379541635513,
-0.0018658638000488281,
-0.026517171412706375,
0.21930284798145294,
0.335012823343277,
0.2512378394603729,
-0.29820746183395386,
0.07764798402786255,
0.22359903156757355,
-0.3672277629375458,
0.393398433... |
https://github.com/huggingface/datasets/issues/302 | Question - Sign Language Datasets | Thanks, Quentin, I think a `requirements.txt` per dataset will be a good thing.
I will work on adding this dataset next week, and once we sort all of the kinks, I'll add more. | An emerging field in NLP is SLP - sign language processing.
I was wondering about adding datasets here, specifically because it's shaping up to be large and easily usable.
The metrics for sign language to text translation are the same.
So, what do you think about (me, or others) adding datasets here?
An exa... | 33 | Question - Sign Language Datasets
An emerging field in NLP is SLP - sign language processing.
I was wondering about adding datasets here, specifically because it's shaping up to be large and easily usable.
The metrics for sign language to text translation are the same.
So, what do you think about (me, or other... | [
-0.19740070402622223,
0.402252733707428,
-0.0791092962026596,
-0.005023062229156494,
-0.07678969204425812,
0.04361025243997574,
0.11855506896972656,
0.25038862228393555,
0.22359159588813782,
-0.3281882405281067,
0.17706722021102905,
0.22711651027202606,
-0.44375598430633545,
0.264460325241... |
https://github.com/huggingface/datasets/issues/301 | Setting cache_dir gives error on wikipedia download | Whoops didn't mean to close this one.
I did some changes, could you try to run it from the master branch ? | First of all thank you for a super handy library! I'd like to download large files to a specific drive so I set `cache_dir=my_path`. This works fine with e.g. imdb and squad. But on wikipedia I get an error:
```
nlp.load_dataset('wikipedia', '20200501.de', split = 'train', cache_dir=my_path)
```
```
OSError ... | 22 | Setting cache_dir gives error on wikipedia download
First of all thank you for a super handy library! I'd like to download large files to a specific drive so I set `cache_dir=my_path`. This works fine with e.g. imdb and squad. But on wikipedia I get an error:
```
nlp.load_dataset('wikipedia', '20200501.de', split =... | [
-0.01641315221786499,
0.2276870310306549,
-0.008926328271627426,
0.10926683247089386,
0.12234675884246826,
0.19993731379508972,
0.25578585267066956,
0.33174675703048706,
0.44854623079299927,
-0.10235477983951569,
-0.19036279618740082,
0.1178651824593544,
0.06666570901870728,
-0.51885932683... |
https://github.com/huggingface/datasets/issues/297 | Error in Demo for Specific Datasets | Thanks for reporting these errors :)
I can actually see two issues here.
First, datasets like `natural_questions` require apache_beam to be processed. Right now the import is not at the right place so we have this error message. However, even the imports are fixed, the nlp viewer doesn't actually have the resourc... | Selecting `natural_questions` or `newsroom` dataset in the online demo results in an error similar to the following.
| 165 | Error in Demo for Specific Datasets
Selecting `natural_questions` or `newsroom` dataset in the online demo results in an error similar to the following.
Thanks for reporting these errors :)
I can a... | [
-0.19810688495635986,
0.40337812900543213,
-0.03457416221499443,
0.09161671996116638,
-0.06098029762506485,
-0.040471259504556656,
0.26432257890701294,
0.39570868015289307,
-0.02065020240843296,
0.11497977375984192,
0.036538660526275635,
0.12681786715984344,
-0.14337976276874542,
0.1707724... |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.