meta-pytorch / meta-pytorch/data
Fuse sequential `map` calls.
Open
Nobody has claimed this yet.
- Dominant language
- Python
- Stars
- 1.3k
- Forks
- 179
- Avg merge
- 6d 1h
- Merged PRs (30d)
- 2
Description
🚀 The feature
In https://github.com/pytorch/pytorch/pull/56497 we tried to introduce fusing of sequential map calls into one MapDataPipe, however it doesn't respect input_col and other kwargs.
We can refactor MapperIterDataPipe to fuse such operations automatically.
Motivation, pitch
Fusing map calls will decrease number of iterators and memory counters.
Alternatives
No response
Additional context
No response
Contributor guide
First steps
- Read the whole issue, then the project's contributing guide.
- Comment on the issue to say you are picking it up — it saves two people doing the same work.
- Fork the repository and make your change on a branch.
- Open a pull request that references the issue number.
Research direction
Start by reviewing PR 56497 and the MapperIterDataPipe implementation. Trace how sequential map operations handle input_col and other kwargs, then verify that automatic fusion preserves those behaviors while reducing iterators and memory counters.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- python
- Domain
- data
- Issue type
- Feature
- Difficulty
- 4/5
- Estimated time
- 3-5 days
- Activity status
- Stale
- Clarity
- Mostly clear
- Newbie friendliness
- 38/100