Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

0.3.5 #345

Merged
merged 23 commits into from
Jul 16, 2024
Merged

0.3.5 #345

merged 23 commits into from
Jul 16, 2024

Conversation

sigmafelix
Copy link
Collaborator

  • Base learner targets (draft; should be changed in accordance with Big Data Considerations #325)
  • Download targets
  • Post-checkout hooks
  • Utility functions for the new targets

@sigmafelix
download part dev in progress
dba0b4a
@sigmafelix
Bug fix: cv utility functions
ceb8633 
- attach_xy: Retaining all columns
- generate_cv_index: Retaining all columns
@sigmafelix
update
e8dcc9d 
- following amadeus changes (process_blackmarble)
- New GPU-enabled R torch library binding
@sigmafelix
set_args_calc
5a2129a 
- path setting is changed to if-else
- pipeline error failing mode to "abridge"
@sigmafelix
path fix + inject_geos fix
12c67f6 
- removing irrelevant arguments passed to terra::rast
@sigmafelix
impute_all update
29bda22
@sigmafelix
0.3.2
8e9f6ce 
- Internal functions are not exported
- attach_xy logic fix: join to the "grand" data rather than the leaned one
- fit_base_brulee and fit_base_xgb: added device to manually distribute GPU workload
@sigmafelix
0.3.2 update
955e975 
- Branching base learner fitting
@sigmafelix
typo fix
0fcdba4
@sigmafelix
wrapper: prepare_cvindex
1f943e5 
- shared interface for branching cv set generation
@sigmafelix
0.3.3
fc39f7a 
- Pipeline base learner: 3 CV strategies and hyperparameter tuning targets were added
- fit_base_* function get `return_best` and `tune_bayes_iter` for `workflow` compatibility
- prepare_cvindex assigns fold ids using function names when spatialsample functions are used
- restore_rset_full: as speed-up and disk saving measures
  - rset objects are generated based on essential coordinates only; this function restores full data for subsequent steps
- TODO: size issues persist in hyperparamter tuning and identifying the best model as the entire workflow should be saved. rsample always saves training/test data.
- TODO: duplicates in features; identify where the duplicates come
@sigmafelix
0.3.4
0e556fd 
- Data size reduction for memory / storage management: added trim_resamples argument in fit_base_*
- make_subdata: bootstrapping (currently 30%)
- restore_fit_best: restore full data with CV rsample rset objects, extract the best tuned results, then fit the data
  - Dealing with nested list in tibbles from tidymodels workflow/hyperparameter tuning
  - TODO/Q: Do we save fitted model object or just keep predictions?
@sigmafelix
0.3.4 update
5e9757f 
- set_args_download and feature_raw_download are written
- targets_download.R is revised to reflect the structure of the two functions
@sigmafelix
0.3.5
15b7fff 
- Base learner: lightgbm
- Added dependencies: bonsai, lightgbm
- Explicit definition of hyperparameter search controls for each
@sigmafelix
set_args_download update + write permission test
136720b 
- LICENSE file gets 554
@sigmafelix
post-checkout hook
3f1b2da 
- set_args_download update
- _targets.R update to generate arglist_download
@sigmafelix
another try
8a33177
@sigmafelix
hook update
68ed2a6
@sigmafelix
Instruction to setup hooks
c7f6757 
- README.md update
- setup_hook.sh is capable of immediately activating permission change
@sigmafelix
Update the list of users with write permission
716ede2
@sigmafelix
Lint and typo fix
24e7376 
- targets_download.R: duplicate target names
- Roxygen2 documentation typo fix in fit_base_lightgbm
@kyle-messier
Copy link
Collaborator

@sigmafelix It looks like main is one ahead of pipeline-compact and needs to be merged that way. It's probably a minor README change. After that, I'll review ASAP. Thanks! 🚀

@sigmafelix
Copy link
Collaborator Author

@kyle-messier I merged main into pipeline-compact. Most functions are still in nocov status, but I will work on improving actual coverage soon.

@sigmafelix sigmafelix merged commit a5d13bb into main Jul 16, 2024
4 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Reviewers

@kyle-messier kyle-messier kyle-messier approved these changes

Assignees
No one assigned
Labels
None yet
Projects
None yet
Milestone
No milestone
Development

Successfully merging this pull request may close these issues.

None yet
2 participants
@sigmafelix @kyle-messier