huggingface / huggingface/datasets

`map` with any joblib backend

Open
#5,991 2 comments 0 reactions 0 assignees View on GitHub
enhancement
Dominant language
Python
Stars
22k
Forks
3.4k
Avg merge
5d 7h
Merged PRs (30d)
17

Description

We recently enabled the (experimental) parallel backend switch for data download and extraction but not for `map` yet.

Right now we're using our `iflatmap_unordered` implementation for multiprocessing that uses a shared Queue to gather progress updates from the subprocesses and show a progress bar in the main process.

If a Queue implementation that would work on any joblib backend by leveraging the filesystem that is shared among workers, we can have `iflatmap_unordered` for joblib and therefore a `map` with any joblib backend with a progress bar !

Note that the Queue doesn't need to be that optimized though since we can choose a small frequency for progress updates (like 1 update per second).

Contributor guide

Open the contributing guide

Assessment

This issue has not been assessed yet.

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.