Downloads · 30 days
0
kaers-bert
kimsiun/kaers-bert
kaers-bert is a machine learning model from kimsiun. Use it for the machine learning task on the model card, and read the license before you ship it in a product. The card lists the license as mit.
The Publicly Available KAERS BERT Embeddings paper contains the unique KAERS-BERT model: initialized with KoBERT (skt/kobert-base-v1) & trained on adverse events (ADEs) narratives reported through KAERS (Korean Advers…
Access
Public
Updated Oct 24, 2022
Repo size
369 MB
Likes
0
Public
Repo makeup
Click a slice to open those files.
.bin369 MB · 100%
At a glance
- License
- mit
- Access
- Public
- Created
- Oct 24, 2022
- Updated
- Oct 24, 2022
- SHA
- 7a493268
Model card
From the Hugging Face model README
KAERS-BERT - KoBERT + KAERS BERT Model
The Publicly Available KAERS BERT Embeddings paper contains the unique KAERS-BERT model: initialized with KoBERT (skt/kobert-base-v1) & trained on adverse events (ADEs) narratives reported through KAERS (Korean Adverse Event Reporting System).
This model card describes the KoBERT model.
Pretraining Data
The KAERS-BERT model was trained on 1.2 million ADE narratives reported through KAERS between January 1, 2015 and December 31, 2019. The ADE narratives used for pertaining were mainly written in Korean.
Model Pretraining
Note Preprocessing
We only used ADE narratives reported as 'disease history in detail', 'adverse event in detail', and 'laboratory test in detail' for model pertaining, because ADE narratives of '(original) reporter's opinion' were highly redundant.
- License
- mit
- Created
- Oct 24, 2022
- Updated
- Oct 24, 2022