v0.3.4
This release includes a variety of bugfixes and new options for running evaluations.
What's Changed
- Bump project version to 0.3.3 by @caufieldjh in #218
- Name params in call to complete by @caufieldjh in #225
- Fixes for the BC5CDR evaluation by @caufieldjh in #226
- Catch requests errors when call to NodeNormalizer endpoint fails by @caufieldjh in #234
- Reduce redundancy in named entity list in evals by @caufieldjh in #236
- Update documentation by @caufieldjh in #220
- Repair clinical note generation and basic completion in CLI by @caufieldjh in #239
- Parameter mis-usage fixes for CLI, OpenAI client, SPIRES engine, and gene set utils by @caufieldjh in #241
- Add medical action extraction template by @caufieldjh in #208
- Bugfixes for Wikipedia extract and search commands by @caufieldjh in #242
- Add a template specifically for extracting human phenotypes by @caufieldjh in #214
- Add parameter to disable chunking in SPIRES evals by @caufieldjh in #245
- fix typo by @andrewsu in #247
- Add
model
parameter to evaluation engine by @caufieldjh in #249 - Fix #250 by @caufieldjh in #251
- Add NER eval for BC5CDR by @caufieldjh in #252
- Fix #253 by @caufieldjh in #254
- Add
gpt-4-1106-preview
by @caufieldjh in #258 - Add two more items to troubleshooting doc by @caufieldjh in #263
- Add MAXO annotation evaluation by @caufieldjh in #259
- Update dependencies for OpenAI API client and Pydantic compatibility and GPT4All by @caufieldjh in #271
New Contributors
Full Changelog: v0.3.3...v0.3.4