Breaking 02 #145
Merged
Breaking 02 #145
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
* Updated documentation for ball LMOs * Update src/norm_oracles.jl Co-authored-by: Zev Woodstock <[email protected]> Co-authored-by: Mathieu Besançon <[email protected]>
* use modify() to change objective * use type of direction for type of objective function * Update src/moi_oracle.jl * modify objective with direction as ScalarAffineTerm * set coefficients to zero when objective is ScalarAffineTerm * formatting * add boolean to MathOptLMO struct to choose btw modify and set * Update src/moi_oracle.jl * Update src/moi_oracle.jl * Update src/moi_oracle.jl * define objective sense in constructor, add inner constructor method * add comment to docstring * remove ==true * update docstring * fix error in test: direction with coefficients * add use_modify to convert_mathopt methods * Update src/polytope_oracles.jl * Update src/polytope_oracles.jl * Update src/simplex_oracles.jl * Update src/norm_oracles.jl Co-authored-by: Mathieu Besançon <[email protected]>
* added tracking srcand test * added tracking src to FrankWolfe.jl * added counters to trajectory state * added example * remove extra testset * added readme tracking reference * RO multiplication (#303) * Update Project.toml * add details to documentation & example * tracking test * added tracking items to fw_state added cached tracking callback, better docs separated vanilla and lazified frank_wolfe test * Update src/tracking.jl Co-authored-by: Mathieu Besançon <[email protected]>
* break memory_mode into two equations in stochastic_fw * add execute_memory_mode to call @memory_mode * rename function * initialize variables * Update src/utils.jl * Update src/utils.jl * muladd with storage * use muladd_memory_mode in pairwise * use muladd_memory_mode in blended_cg * use muladd_memory_mode in perform_line_search * initialize d outside of loop * remove spaces * add memory_mode to perform_line_search Co-authored-by: Mathieu Besançon <[email protected]>
* Start building cleaner docs * Shorter workflow (test on 1 doc example) * Fix CI without activate.jl by addind TestEnv.activate() * Fix examples CI again * Small tweaks to markdown docs * Add TestEnv to package dependency (temporary fix but very light) * Fix CI once more * Fix CI for examples, hopefully for the last time * I said it was the last time, but I was wrong * Fix CI for good * Rename docs examples * Restructure API reference * Undev FW in docs env * Update docs/src/advanced.md * Update docs/make.jl * Update examples/linear_regression.jl * Update .github/workflows/ci.yml * Update docs/src/algorithms.md * Update docs/src/reference/3_backend.md Co-authored-by: Mathieu Besançon <[email protected]>
Stochastic Frank Wolfe was at the wrong section level
* Update README.md * Update README.md
Fix typo
* Handed correct gamma_max value to perform_line_search in pairwise step * Handed correct gamma_max to away_frank_wolfe Co-authored-by: Hendrych <[email protected]>
* modified tracking callbacks as extension to stop criterion callbacks * renamed grad! state variable to grad, added stop criterion boolean * added stop criterion callbacks, tests * changed grad! state variable in lazified fw to grad * removed stop_criterion varibale, added break instead * changed comparison type to work with all return types * Update src/stop_criterion.jl * Update test/stop_criterion.jl * remove "!" because not modifying anymore * added closure version of tracking callback, changed tests accordingly * new callbacks file * change name into callbacks and added test file * removed trajectory callback * removed trajectory callback * deleted stop_criterion * moved all callbacks to new callbacks file * added default callback behaviour state -> false. f, lmo, grad!, tracking versions * renamed callback files, callback functions to stop criteria * added unified callback * reworked fw algorithm for unified callback * removed print callback from utils, merged into callback doc * removed state formatting, now in each fw_alg separately * added unified callback, refactored fw algorithms: added new entries stop_criterion and traj_data added format_state function at the top of each format modified callback and last callback calls and state recalculation * modified tracking test file according to new unified callback * reverted back to previous stachastic FW version of breaking-02 * changed comment for conflic resolve * second revert to breaking-02 * adapted docs example to new unified callback, added custom callback example * separated make_callback into trajectory and verbose wrappers * added new separated callback wrappers, callback !== nothing check * renaming * reverted default push trajectory to base tuple only * adjusted tracking to unified callback * deleted unnecessary stop criteria, can be added to examples later * adapted docs example to unified callback, added custom callback * deleted unnecessary stop criterion test * edit * added doc * deleted callback test file * EDIT * edits to conform with breaking-02 structure * edit * add last gamma computation * modified docs, changed default interrupt behaviour * conflicts * end of line * modified last, pp cases of print_callback * unified callback for awf * unified callback for stochastic frank wolfe * added format_state function, added batch size to state * added state prefix to afw * renamed tot_time * tot_time edit * changed recompute t to t-1 * changed recompute t to t-1 * adapted blended_cg for unified callback * added === back again * added tt values to state * modified print callback to print dualsteps * state.print_iter -> print_iter * adapted pairwise to unified callback * Merge commit 'ee32c8661ece653a33a2742de49077b7d89d4d9c' * removed print_callback parameters now part of callback * added missing state parameters * added callback === nothing and reversed order in make_trajectory_callback * removed old trajectory callback * added traj_data to FW call parameters * added traj_data * changed format_state * replace vertex_taken by v * added InplaceEmphasis() parameter to perform_line_search * minor improvements * fix condition when traj callback * formatting * simplify callback * update keywords * added active set * corrected added active set and non simplex iter * manual revert to update keywords * changed order * added tot_time, callback condition * added tot_time, callback condition * fix tottime * Update src/tracking.jl * Apply suggestions from code review * changed tt strings to step symbols * added elif to remove duplicate "last" prints * removed assignment from internal FW trajectory * added gamma assignment * also exclude post processing from being pushed onto trajectory * adjusted test to removed post processing trajectory push Co-authored-by: Mathieu Besançon <[email protected]>
* Added docstrings for linesearch methods * reformatted defs.jl * Apply suggestions from code review Co-authored-by: Zev Woodstock <[email protected]> Co-authored-by: Mathieu Besançon <[email protected]>
* Added page for line search methods * more details * syntax * Documenter syntax Co-authored-by: Zev Woodstock <[email protected]> Co-authored-by: Mathieu Besançon <[email protected]>
* adapted doc improvement to breaking-02 * improve structure
Co-authored-by: Hendrych <[email protected]>
* fix tol with epsilon * define epsilon
* Update Project.toml * fix tol with epsilon
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
No description provided.