curator
Synthetic data curation for post-training and structured data extraction
GraphCanon updated today · GitHub synced today
Decision brief
Synthetic data curation for post-training and structured data extraction
Good fit when
- Ideal for enhancing the performance of existing machine learning models through fine-tuning in natural language processing contexts
- Best suited when you require synthetic datasets to test or train models under varied conditions not present in real-world data
Avoid when
- Not recommended if your needs extend beyond NLP and you do not work with structured text data
- May not be the best choice for simple data generation tasks that do not benefit from complex synthetic dataset creation processes
Observed Jul 17, 2026 · Source: enrich:decision_facts
Verify the decision
Maintenance and security
Full trust report- Maintenance
- Active (16d since push)
- As of today
- Provenance
- Not a fork · Organization account
- As of today
- Security (OSV)
- No lockfile
- As of 1mo
Public GitHub metadata and optional OSV scans. Signals, not a guarantee. Trust methodology.
Install
pip install curatorSimilar tools
Same-category neighbours. No typed graph edges are catalogued for this tool yet.
Evidence and technical details
Sourced facts, taxonomy, compatibility claims, README excerpt, and machine-readable endpoints.
Overview
A Python-based tool focused on generating synthetic datasets for improving machine learning models through fine-tuning and instruction tuning, especially useful in natural language processing contexts.
Capability facts
- CLI
- CLI entrypoint
Source: pyproject.toml:[project.scripts] · Aug 24, 2026
- Languages
- python
Source: github.language+pyproject.toml · Aug 24, 2026
Categories
Compatibility
Sourced claims from the README excerpt - not unsourced marketing copy.
Source: README excerpt (regex_v1, Aug 24, 2026)
response = trainer.sample("Explain recursion in Python")Source link
Tags
README
🛠️ Installation
pip install bespokelabs-curator
serverlessly, so this provisions an on-demand deployment (takes a few minutes).
response = trainer.sample("Explain recursion in Python") print(response)
trainer.close() # tear down the deployment when done
> **Note:** Running real Fireworks training requires an account with training quota
> (Tier 2 / credits). Without the SDK or an API key, `FireworksTrainer` runs in mock
> mode so examples and tests work offline. Subclass `FireworksTrainer` and override
> `format_example()` to handle custom data layouts, exactly as with `TinkerTrainer`.
See the [Fireworks examples](examples/fireworks/) for basic and custom-trainer pipelines.
For agents
This page has a .md twin and JSON over the API.