Register
Login
Resources
Docs Blog Datasets Glossary Case Studies Tutorials & Webinars
Product
Data Engine LLMs Platform Enterprise
Pricing Explore
Connect to our Discord channel
avatar
Emmanuel M. Smith

mraxilus

I mainly use DAGsHub for mirroring purposes. My personal projects are hosted on GitLab (mraxilus).

avatar

Emmanuel M. Smith

mraxilus

I mainly use DAGsHub for mirroring purposes. My personal projects are hosted on GitLab (mraxilus).


mraxilus synced commits to master at mraxilus/thesis-doctorate from mirror

2 years ago

mraxilus synced commits to master at mraxilus/thesis-doctorate from mirror

  • 67dba604a3 feat: finish running entire pipeline I.e. the air quality prophet stages.

2 years ago

mraxilus synced and deleted reference refs/pipelines/533685439 at mraxilus/thesis-doctorate from mirror

2 years ago

mraxilus synced new reference refs/pipelines/533685439 to mraxilus/thesis-doctorate from mirror

2 years ago

mraxilus synced commits to master at mraxilus/thesis-doctorate from mirror

  • ba164f790c feat: run Prophet pipeline for water data

2 years ago

mraxilus synced commits to master at mraxilus/thesis-doctorate from mirror

  • 3ee837d693 feat: rerun entire pipeline less Prophet air/water

2 years ago

mraxilus synced commits to master at mraxilus/thesis-doctorate from mirror

  • e5be80d65e fix: ensure Prophet predictions don't include timestep 0
  • 2d50202703 feat: refactor entire pipeline to remove data redundency across cores Drastically simplified code in all stages of the pipeline, which shuold also give some minor performance improvements. I swapped from `multiprocessing` to `ray` to reduce the memory duplication. Also, simplified the logging so that it would be more clear what progress has been made within a stage (more specifically modelling).
  • View comparison for these 2 commits »

2 years ago

mraxilus synced commits to master at mraxilus/thesis-doctorate from mirror

2 years ago

mraxilus synced commits to master at mraxilus/thesis-doctorate from mirror

2 years ago

mraxilus synced commits to master at mraxilus/thesis-doctorate from mirror

  • ef9a3dc81e refactor: run black on notebook
  • c8a8319a13 feat: add more validation checks for pipeline data Ideally I would actually create tets to verify things such as whether or not the caches adhere to the desired schema, but these spot checks will have to do for now.
  • 5938d5f3dc feat: add root relative squared error metric
  • View comparison for these 3 commits »

2 years ago

mraxilus synced commits to master at mraxilus/thesis-doctorate from mirror

2 years ago

mraxilus synced commits to master at mraxilus/thesis-doctorate from mirror

  • 67dba9e2e9 ci: fix pipeline from failing on dag generation Stage names have changed which broke it.
  • 4bed5d2d9e feat: add addional moments for error surface generation Added standard deviation and skew in addition to mean.
  • 084b8e5cee revert: remove numexpr as it doesnt work well with new nullable dtypes
  • 3049cf531e feat: use numexpr for error calucations for performance improvements
  • 222162e920 feat: run stages up to estalbish and all of naive
  • View comparison for these 6 commits »

2 years ago

mraxilus synced and deleted reference refs/pipelines/523127577 at mraxilus/thesis-doctorate from mirror

2 years ago

mraxilus synced new reference refs/pipelines/523127577 to mraxilus/thesis-doctorate from mirror

2 years ago

mraxilus synced commits to master at mraxilus/thesis-doctorate from mirror

2 years ago

mraxilus synced commits to master at mraxilus/thesis-doctorate from mirror

  • 838939fa55 feat: produce complete results for air quality dataset

2 years ago

mraxilus synced commits to master at mraxilus/thesis-doctorate from mirror

  • a6ab245f85 fix: remove unnecesary file checks from pre-commit

2 years ago

mraxilus synced commits to master at mraxilus/thesis-doctorate from mirror

  • 6fb8f28bbe ci: ensure dag plot produces entire graphs It broke last commit as the stage names were misspecified.
  • a7a42aee55 feat: run pipeline for naive air
  • 67618ca5e2 ci: move over to aws from gcs GCS and GitLab runners don't seem to work well together. I can run the pipeline successfully, locally, but it always fails with `Forbidden` on the shared runners. Just using s3 instead, but still keeping the main backup in a GCS bucket.
  • View comparison for these 3 commits »

2 years ago

mraxilus synced commits to master at mraxilus/thesis-doctorate from mirror

  • 6b356e342b feat: add amazon as backup dvc remote
  • d94e7fb5df feat: produce results for synthetic and water data
  • 6cc3be582b ci: attempt to resolve pipeline failure Remove verbosity flags. Added visualization of the pipelines. I got it working on a local GitLab runner instance, so it should work when I push. I'm unsure if the `status` command is the fix culprit, or if I had to change the cache type to copy and relink (or both).
  • f2c5c676fa ci: add dvc status command for troubleshooting
  • e58578f144 ci: try using copy as cache type to resolve dvc errors
  • View comparison for these 7 commits »

2 years ago

mraxilus synced commits to master at mraxilus/thesis-doctorate from mirror

  • 3b26d8fa26 ci: ensure stages are properly defined
  • e7b3736d91 ci: attempt to get more debugging info from dvc
  • 9ece4ee368 ci: fix pipeline by ingoring notebooks Notebooks will still be checked locally. Once I have more time I can refactor all the existing notebooks and start checking them within the CI pipeline.
  • 43a34e33c0 ci: fix prettier command line option
  • 4e6fcf5ef0 ci: add pre-commit validation stage Ensure that pre-commit was run locally. It required that I ignore some things that I don't want to deal with at the moment. I added TODOs so that I would remember to unignore them in the future.
  • View comparison for these 7 commits »

2 years ago