Releases: cumc-dbmi/cehrbert
Releases · cumc-dbmi/cehrbert
v1.2.3
check whether patient_id exists in the meds_reader database because s…
v1.2.2
upgraded cehrbert_data from v0.0.3 to v0.0.4 (#65)
v1.2.1
What's Changed
- created a helper function to convert the existing huggingface dataset to an iterable dataset by @ChaoPang in #64
Full Changelog: v1.2.0...v1.2.1
v1.2.0
What's Changed
- Fix the data and model relative paths by @ChaoPang in #58
- Added units to meds to cehrbert data transformation and added unit in normalizing numeric values by @ChaoPang in #60
- Fixed the relative paths in the finetuning runner by @ChaoPang in #61
- The zero division bug for calculating the unit probabilities for numeric concepts by @ChaoPang in #62
- used the existing subject_split in meds_reader to split the cohort automatically instead of doing it manually by @ChaoPang in #63
Full Changelog: v1.1.9...v1.2.0
v1.1.9
Implemented TruncatedOnlineStatistics to collect truncated stats in a…
v1.1.8
v1.1.7
What's Changed
- Implement a custom do_predict because the HF trainer predict has memory leakage for large test sets by @ChaoPang in #55
Full Changelog: v1.1.6...v1.1.7
v1.1.6: Remove eval compute metrics (#54)
What's Changed
- Stopped using the transformers (4.39.3) trainer compute_metrics to calculate custom metrics as large validation sets will crash by @ChaoPang in #54
Full Changelog: v1.1.4...v1.1.6
v1.1.4
v1.1.2
What's Changed
- Create pylint.yml by @ChaoPang in #46
- Apply pre-commit code formatters by @ChaoPang in #47
- updated the image references in README to reference the ones on github by @ChaoPang in #48
- removed fastparquet from the dependency list by @ChaoPang in #49
- removed the abosolute model and tokenizer paths from pretraining and finetuning by @ChaoPang in #50
- Updated meds_reader to 0.1.9 by @ChaoPang in #51
Full Changelog: https://github.com/cumc-dbmi/cehrbert/commits/v1.1.2