GBSG/SUPPORT data tasks and curve evaluation for the merged SurvivalMDN

Abierto
  • ¿Algo no está bien?
#1,357 1 comentario 0 reacciones 0 asignados Ver en GitHub

Los mantenedores suelen responder en 2 días

Nadie ha tomado este issue todavía.

Evaluación

Dificultad
5/5
Tiempo estimado
Más de una semana
Aptitud para principiantes
25/100
Tipo de issue
Nueva funcionalidad
Claridad
Bastante claro
Estado de actividad
Activo
Stack tecnológico
python
Área
machine-learning

Línea de trabajo

Start with the contributor guide, the merged SurvivalMDN work in #1279, and the collaborator's fork PR to understand what is already underway. The proposed work spans GBSG and Support2Dataset tasks, survival-curve metrics, examples, tests, and API documentation; first get maintainer guidance on scope and PR boundaries before treating it as a bounded contribution.

Escrito por el modelo de indexación a partir del texto del issue.

Descripción

Hi PyHealth team,

I'm working with @enjaybaze on a reusable survival workflow around the SurvivalMDN merged in #1279. Following your contributor guide, I'd like to check the scope with you before opening an upstream PR.

Our proposed contribution adds:

  • A GBSG dataset loader and time-to-event tasks for GBSG and the existing Support2Dataset. The tasks emit separate duration/event tensors for the merged model and standard Trainer.
  • Helpers for the authors' ten released splits of both datasets, with documented time conventions and training-only preprocessing.
  • Reusable survival-curve metrics, including time-dependent concordance, IBS, IBLL and censoring-aware D-calibration. The paper-reference and aligned-grid evaluation conventions are reported separately.
  • Runnable training and checkpoint-reload examples, an optional Cox/DeepSurv benchmark, synthetic tests and API documentation.

The upstream SurvivalMDN model, its existing tests and Trainer are unchanged. Our current review branch is on my collaborator's fork: https://github.com/enjaybaze/PyHealth/pull/2. Its core, NLP and contribution CI checks pass. Additional CPU/CUDA pilots cover both datasets, checkpoint reload and saved-result reuse.

Would this scope be useful for PyHealth? Would you prefer one PR, or a smaller first PR for GBSG and the tasks with the general metrics and benchmark reviewed separately? I'd also appreciate any naming or integration guidance before we submit.

Thanks,
Aritra

Lenguaje dominante
Python
Estrellas
1.7k
Forks
803
Merge medio
4 d 18 h
PR fusionados (30 d)
58

Preparar el entorno

Primeros pasos

  1. Lee el issue completo y luego la guía de contribución del proyecto.
  2. Comenta en el issue que vas a ocuparte — evita que dos personas hagan lo mismo.
  3. Haz un fork del repositorio y trabaja en una rama.
  4. Abre un pull request que haga referencia al número del issue.

Más de sunlabuiuc/PyHealth

Todos los issues de sunlabuiuc/PyHealth

Issues similares

Más issues de Python

Recibe los nuevos issues en tu correo

Un resumen breve de issues de GitHub para principiantes.