NVIDIA-Merlin / NVIDIA-Merlin/Transformers4Rec
Improve performance of ranking metrics calculation
Open
Nobody has claimed this yet.
area/pytorch
enhancement
- Dominant language
- Python
- Stars
- 1.3k
- Forks
- 165
- Avg merge
- 1m
- Merged PRs (30d)
- 2
Description
- The current implementation of ranking metrics computes the top k scores within each metric class using the method extract_topk which is computationally expensive.
- A possible solution is to Include the extraction of the top-k scores within the
NextItemPredictionTaskclass and provide the ranked scores to the ranking metrics.
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 in transformers4rec/torch/ranking_metric.py at extract_topk, then inspect NextItemPredictionTask. Trace how ranking metrics currently compute top-k scores and how ranked scores could be provided by the task. Done means the proposed top-k extraction path is implemented across these components and the ranking metrics receive the ranked scores.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- python, pytorch
- Domain
- machine-learning, performance
- Issue type
- Refactor
- Difficulty
- 4/5
- Estimated time
- 3-5 days
- Activity status
- Stale
- Clarity
- Mostly clear
- Newbie friendliness
- 35/100