facebookresearch/fairseq

Data loading in multi-lingual translation

Open

#2,410 opened on Aug 2, 2020

View on GitHub
 (8 comments) (0 reactions) (0 assignees)Python (6,224 forks)batch import
enhancementhelp wantedneeds triage

Repository metrics

Stars
 (29,107 stars)
PR merge metrics
 (No merged PRs in 30d)

Description

Hello, I use the multi-lingual translation task and find some issues:

  1. When I use the round_robin_dataset and multi_corpus_sampled_dataset for more than 20 language pairs, the data loading will take much more time in the functions of filter_by_size and batch_by_size. I find that in multi_corpus_sampled_dataset, the computation cost in calling the function of num_tokens is O(N^2). Could you provide some optimization for these two datasets?

  2. The data loading will fail (RuntimeError: received 0 items of ancdata) when I use more language pairs, such as 28 language pairs. I find there is a daemon thread, which supports the buffer mechanism. When I close the buffer mechanism, the data loading can succeed. I do not figure out why this happens. Could you fix this issue? Thank you very much.

Contributor guide

Research direction
Inspect the multi corpus sampled dataset implementation for O(N^2) complexity in num tokens calls and examine the buffer daemon thread causing RuntimeError when using many language pairs. Consider optimizing the computation and diagnosing the buffer mechanism failure.
Tech stack
python
Domain
backend
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.
Half day
Activity statusHow available the issue appears right now: fresh, active, stale, blocked, or waiting on maintainer input.
Active
ClarityHow clearly the issue explains the expected change, acceptance criteria, and next step.
Clear
Prerequisites
Python
Newbie friendlinessA 1-100 score estimating how approachable this issue is for first-time contributors.
70