alibaba / alibaba/EasyNLP

Buffered shuffle on odps dataset

Open
#119 0 comments 0 reactions 0 assignees View on GitHub
Dominant language
Python
Stars
2.2k
Forks
257
PR merge metrics
No merged PRs in 30d

Description

https://discuss.pytorch.org/t/how-to-shuffle-an-iterable-dataset/64130/5

- torch 1.8.1
[torch.utils.data.BufferedShuffleDataset](https://pytorch.org/docs/1.8.1/data.html#torch.utils.data.BufferedShuffleDataset)

- latest
[torch.utils.data.datapipes.iter.combinatorics.ShuffleIterDataPipe](https://github.com/pytorch/pytorch/blob/615dd2508846d03bc1f4e21898cc2694535011fe/torch/utils/data/datapipes/iter/combinatorics.py#L55)

Contributor guide

No contributing guide indexed for this repository

Research direction

Start by reading the linked PyTorch BufferedShuffleDataset documentation and the referenced ShuffleIterDataPipe implementation, then compare them with EasyNLP's ODPS dataset entry point. The issue does not name a repository file or test, so identify the affected dataset path and define completion criteria before changing code.

Written by the indexing model from the issue text.

Assessment

Tech stack
python, pytorch
Domain
data-engineering
Issue type
Feature
Difficulty
5/5
Estimated time
Over a week
Activity status
Stale
Clarity
Needs clarification
Newbie friendliness
20/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.