ODIN

Repository for ODIN project

Notes for users:

• DO NOT PUSH ANY HOSPITAL DATA TO GIT, EVER
• THIS ALSO MEANS THAT YOU HAVE TO CLEAR YOU NOTEBOOK CELLS
• Put jsons/yaml etc. with configuration information in the asset folder

We have three models:

• base model: pipeline([TFIDF vectorizer (vanilla settings) for the text->PCA(0.90), standardize the rest], SVM/XGBoost)
• ResNet1D
• H-Transformer 1D

from sklearn.pipeline import Pipeline

il_tubo = Pipeline([
    ('Vectorizer', tfidf_vectorizer(**vkwargs)),
    ('PCA', PCA(**pkwargs)), 
    ('Clf', SVM(**clfkwargs))
])

Difficulty is putting the cumulative variance restriction IN il_tubo during inference(il_tubo.predict).