 Richard Wong
|
c5760d127d
|
Feat: added post_processing based on rules
others:
- added basic data analysis to get histograms of text differences
- added new final delivery model
|
2024-12-18 13:43:56 +09:00 |
|
Richard Wong
|
b01ca4f395
|
Feat: implement hybrid fine-tuning of encoder and decoder networks
separately
|
2024-12-10 23:40:10 +09:00 |
|
Richard Wong
|
737c86bc2e
|
Feat: added de_duplication post-processing method
|
2024-11-28 11:02:22 +09:00 |
|
Richard Wong
|
ff6e11a3c0
|
Feat: added more classification and mapping variations
Feat: added grid-search for threshold in similarity-classifier
Feat: added more abbreviation rules
|
2024-11-25 18:15:28 +09:00 |
|
Richard Wong
|
1f3970459f
|
Chore: re-organized train folders to have standardized naming schemes
Feat: introduced BERT-based binary classification
|
2024-11-20 15:07:47 +09:00 |
|
Richard Wong
|
7699201cb8
|
Feat: implement selection for pattern-mapping
Feat: added error analysis for BERT find-back
Feat: added direct mapping with unit
Feat: added BERT for classification using description only
|
2024-11-11 20:20:43 +09:00 |
|
Richard Wong
|
2b5994cb52
|
Feat: added abbreviation expansion rules
|
2024-11-10 20:28:47 +09:00 |
|
Richard Wong
|
59bbf1f403
|
Feat: implement find-back for analysis in find_closest.py
Feat: implement bert classification
|
2024-11-08 20:50:41 +09:00 |
|
Richard Wong
|
22429ea536
|
Feat: added classification methods
Feat: added mapping to pattern-only method
Chore: re-organized prediction to be within mapping folders
|
2024-11-05 16:49:18 +09:00 |
|
Richard Wong
|
0ad182f2b9
|
Feat: added classification for all data, including non-mdm, as a
training baseline model
|
2024-11-01 13:21:12 +09:00 |
|
Richard Wong
|
18e4a5f7df
|
Chore: moved selection to post_process, mapping to test
|
2024-10-31 16:35:28 +09:00 |
|
Richard Wong
|
16374b9ab8
|
Feat: added train and test directories
|
2024-10-31 15:58:20 +09:00 |