Repository navigation

All pull requests

Labels (14)Milestones (0)New pull request

Open

(23)

Closed

(92)

Optimize large-N multi-way clustered VCOV and fix 64-bit matrix indexing
#700tholden opened Nov 5, 2026Contributor
Add mem.clean = aggressive option to drop non-cluster data.table columns in-place
#699tholden opened Nov 5, 2026Contributor
Skip cpp_xbeta and allow in-place demeaning of X when lean = TRUE
#698tholden opened Nov 5, 2026Contributor
Avoid redundant full-matrix copies of X_all and X_demean in multi-LHS feols
#697tholden opened Nov 5, 2026Contributor
Bound intermediate memory in cpp_demean via column batching
#696tholden opened Nov 5, 2026Contributor
Fix 32-bit integer overflow in demeaning coefficient counts and offsets
#695tholden opened Nov 5, 2026Contributor
Port check_conv to parallel C++ (cpp_check_conv) and handle zero-slope clusters
#691tholden opened Oct 28, 2026Contributor
Fix numerical instability in demeaning with constant or near-constant varying slopes
#690tholden opened Oct 27, 2026Contributor
Remove unused std::string 'res'
#688MichaelChirico opened Oct 27, 2026
Fix double I() wrapping of user-supplied I(x^2)
#685kylebutts opened Sep 25, 2026Collaborator
fix: sparse_model_matrix improvements
#684kylebutts opened Sep 24, 2026Collaborator
Enable two-level nested parallelism in demeaning using OpenMP tasks and taskloops
#683tholden opened Sep 21, 2026Contributor
Bound total demeaning iterations across warmup and post-warmup by iterMax
#682tholden opened Sep 19, 2026Contributor
Avoid O(N^2) element-shifting bottleneck during fixed-effect removal
#681tholden opened Sep 18, 2026Contributor
1
Add glm.continueOnWarn option to feglm/fepois to continue outer iterations on inner demeaning warning
#680tholden opened Sep 16, 2026Contributor
6
Regenerate the documentation with roxygen2 8.1.0
#678raffaelemancuso opened Sep 9, 2026
Add a .gitattributes enforcing LF line endings
#677raffaelemancuso opened Sep 9, 2026
Support weighted estimation in femlm/fenegbin/feNmlm
#676raffaelemancuso opened Sep 9, 2026
Fix subscript out of bounds in predict() for stepwise models with i()
#673SAY-5 opened Jul 18, 2026
Add fml_extract proposal
#649kylebutts opened May 1, 2026Collaborator
Proof-of-Concept within demeaner backend
#647s3alfisc opened Apr 29, 2026Contributor
10
Fix crash in IV estimation when demeaning does not converge
#639adamaltmejd opened Mar 19, 2026
7
Add Anderson rubin inference
#621kv9898 opened Nov 28, 2025
11

Navigation Menu

Repository navigation

All pull requests

Optimize large-N multi-way clustered VCOV and fix 64-bit matrix indexing

Add mem.clean = aggressive option to drop non-cluster data.table columns in-place

Skip cpp_xbeta and allow in-place demeaning of X when lean = TRUE

Avoid redundant full-matrix copies of X_all and X_demean in multi-LHS feols

Bound intermediate memory in cpp_demean via column batching

Fix 32-bit integer overflow in demeaning coefficient counts and offsets

Port check_conv to parallel C++ (cpp_check_conv) and handle zero-slope clusters

Fix numerical instability in demeaning with constant or near-constant varying slopes

Remove unused std::string 'res'

Fix double I() wrapping of user-supplied I(x^2)

fix: `sparse_model_matrix` improvements

Enable two-level nested parallelism in demeaning using OpenMP tasks and taskloops

Bound total demeaning iterations across warmup and post-warmup by iterMax

Avoid O(N^2) element-shifting bottleneck during fixed-effect removal

Add glm.continueOnWarn option to feglm/fepois to continue outer iterations on inner demeaning warning

Regenerate the documentation with roxygen2 8.1.0

Add a .gitattributes enforcing LF line endings

Support weighted estimation in femlm/fenegbin/feNmlm

Fix subscript out of bounds in predict() for stepwise models with i()

Add `fml_extract` proposal

Proof-of-Concept `within` demeaner backend

Fix crash in IV estimation when demeaning does not converge

Add Anderson rubin inference

Pull requests