facebookresearch/fairseq

Invalid suffix of raw dataset when preprocessing without language

Open

#1,426 opened on Nov 25, 2019

View on GitHub
 (2 comments) (0 reactions) (0 assignees)Python (29,107 stars) (6,224 forks)batch import
bughelp wanted

Description

When preprocessing using --dataset-impl raw and no source and target languages are specified, the datasets are stored under train.None-None due to this line:

https://github.com/pytorch/fairseq/blob/5349052aae4ec1350822c894fbb6be350dff61a0/preprocess.py#L218

Is this expected behavior or can we remove this suffix?

Contributor guide

Tech stack
python
Domain
datamachine learning
Issue type
bug
DifficultyEstimated implementation difficulty for a new contributor, from 1 for very small changes to 5 for expert-level work.
3
Estimated timeA rough time range for an experienced contributor to investigate, implement, test, and prepare a pull request.
under 1 hour
Activity statusHow available the issue appears right now: fresh, active, stale, blocked, or waiting on maintainer input.
needs maintainer response
ClarityHow clearly the issue explains the expected change, acceptance criteria, and next step.
clear
Prerequisites
Pythonfairseq preprocessing
Newbie friendlinessA 1-100 score estimating how approachable this issue is for first-time contributors.
50
Research direction
Investigate line 218 of preprocess.py (https://github.com/pytorch/fairseq/blob/5349052aae4ec1350822c894fbb6be350dff61a0/preprocess.py#L218). When source lang and target lang are not provided, the lang variables become None, resulting in filenames like 'train.None None'. Determine if this suffix is necessary for raw dataset files. If not, remove the suffix or handle the None case. Review any existing tests or documentation for expected behavior.