Stuck on an issue?

Lightrun Answers was designed to reduce the constant googling that comes with debugging 3rd party libraries. It collects links to all the places you might be looking at while hunting down a tough bug.

And, if you’re still stuck at the end, we’re happy to hop on a call to see how we can help out.

Module Not found: datasets_modules.datasets.output

See original GitHub issue

Environment info

transformers version: 4.5.0.dev0
Platform: Linux-3.10.0-1160.15.2.el7.x86_64-x86_64-with-glibc2.10
Python version: 3.8.5
PyTorch version (GPU?): 1.8.1 (False)
Tensorflow version (GPU?): not installed (NA)
Using GPU in script?: not sure
Using distributed or parallel set-up in script?: <fill in> ?

Who can help

@patil-suraj

Information

Model I am using (Bert, XLNet …): BART seq2seq

The problem arises when using:

the official example scripts: (give details below)
my own modified scripts: (give details below)

The tasks I am working on is:

an official GLUE/SQUaD task: (give the name)
my own task or dataset: (give details below)

To reproduce

Steps to reproduce the behavior:

“Install from source” method
ran this command, where “data/output.jsonl” is my dataset:

python examples/seq2seq/run_translation.py \
    --model_name_or_path t5-small \
    --do_train \
    --do_eval \
    --source_lang en \
    --target_lang de \
    --source_prefix "Translate English to Logical Forms: " \
    --dataset_name data/output.jsonl \
    --output_dir /tmp/tst-translation \
    --per_device_train_batch_size=4 \
    --per_device_eval_batch_size=4 \
    --overwrite_output_dir \
    --predict_with_generate

Got the following error:

ModuleNotFoundError: No module named 'datasets_modules.datasets.output'

At first it told me that “datasets” was not installed, so I did pip install datasets and that worked fine. Then I got this error and haven’t been able to figure out what it means or how to fix it.

Issue Analytics

State:
Created 2 years ago
Comments:5

Top GitHub Comments

1reaction

ashleylewcommented, Apr 19, 2021

Yeah! So in my command script that I posted here, I used --dataset_name data/output.jsonl \ but that command is for pre-loaded data sets, whereas mine is a custom one. So instead you’ll want to use:

    --train_file train.json \
    --validation_file validation.json \

0reactions

abhishekkubercommented, Apr 19, 2021

@ashleylew Hi I’m getting the same error, can you share the data loading script? I tried doing stuff but it doesn’t work for me. Thanks

Top Results From Across the Web

ImportError: No module named datasets - Stack Overflow

But i am getting error. from datasets import dataset_utils ImportError: No module named datasets. I found this solution How can jupyter access a ......

Modulenotfounderror: no module named datasets ( Solved )

The root cause for the no module named datasets error is that you have not properly installed it in your system. And also...

Source code for datasets.load - Hugging Face

For datasets on the Hugging Face Hub (list all available datasets and ids with ``datasets.list_datasets()``) - if ``path`` is a canonical dataset or...

tf.data: Build TensorFlow input pipelines

Dataset abstraction that represents a sequence of elements, in which each ... please make sure the missing libraries mentioned above are installed properly....

SD (scientific dataset) API (pyhdf.SD) - GitHub Pages

A module of the pyhdf package implementing the SD (scientific dataset) API of the NCSA HDF4 library. Introduction¶. SD is one of the...