microsoft / microsoft/SynapseML
Support VW native training format
Nobody has claimed this yet.
- Dominant language
- Scala
- Stars
- 5.2k
- Forks
- 868
- Avg merge
- 22h 9m
- Merged PRs (30d)
- 45
Description
This is not related to a current issue, however it would be beneficial to support VW native training format, since this allows us to have namespace-prefixed coefficients in our human-readable output models (for specific model types, such as LR or FTRL).
Currently training is limited to libsvm format.
According to @eisber this will require creating a new input format parser.
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 locating the current libsvm training input path and the input format parser mentioned in the issue, then review how LR and FTRL models produce human-readable output. Compare those entry points with the VW native format requirements; done means training accepts VW input and produces namespace-prefixed coefficients for the supported model types.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- scala, spark
- Domain
- machine-learning
- Issue type
- Feature
- Difficulty
- 4/5
- Estimated time
- 3-5 days
- Activity status
- Stale
- Clarity
- Mostly clear
- Newbie friendliness
- 35/100