Skip to content

Latest commit

 

History

History
31 lines (23 loc) · 885 Bytes

README.md

File metadata and controls

31 lines (23 loc) · 885 Bytes

Uploading models to Hugging Face Hub

Upload

Upload all pretrained hf_ehr models and tokenizers to the Hugging Face via:

python hf_upload.py

This script will:

  1. Create a new Hugging Face config
  2. Create a new Hugging Face model
  3. Create a new Hugging Face tokenizer
  4. Save the model and tokenizer to a local directory
  5. Upload the model and tokenizer to Hugging Face Hub

Usage

An example of how to use a pretrained hf_ehr model+tokenizer from Hugging Face to run inference on a patient can be found in hf_test.py. Simply run:

python hf_test.py

For every model that we upload to Hugging Face, this script will:

  1. Download the model+tokenizer from Hugging Face Hub
  2. Tokenize a patient
  3. Run inference on the patient
  4. Print the patient representation

If the model is not working, it will be printed out as part of the bad_models list.