search

WadhwaniAI / covid-modelling

Repo for modelling the spread of novel Covid-19
MIT License
5 stars 0 forks source link

Parallel uncertainty #3

Closed jerome-ai closed 3 years ago

jerome-ai commented 4 years ago

The purpose of this pull request is to:

  1. Give an idea of how I would go about parallelizing this workflow (see run.sh for the general pipeline)
  2. Allow you to replicate the RuntimeWarning I'm seeing (see the WARNING logs from estimation.py)
  3. Provide an alternative (and potentially a simplification) of overall Pandas processing (see get-c19i.py)

I think there are subtle bugs in a few places, so merge only if you want. This is more meant to provide my perspective.

sansiddh commented 3 years ago

Hi Jerome.

Is it possible if we can close this PR?

Avtansh and Mihir are working out of the avtansh branch now. The uncertainty branch is stale and doesn't have a "working" MCMC code

sansiddh commented 3 years ago

@avtansh1317 and @mihirk-wadhwaniai you guys can have a look at @jerome-ai 's branch for a parallel MCMC implementation. I remember you were facing some issues parallelising on odin right?

jerome-ai commented 3 years ago

Is it possible if we can close this PR?

Sure!