Matthieu Gomez

Do you have a functional NVIDIA driver? What does `CUDA.functional()` return?

I also think one should be able to specify standard errors in `fit` (rather than using a `vcov` function on the result of `fit` as in R) for several reasons:...

Exactly, computing errors as part of the `fit` does not require code duplication. I actually also define a vcov method in my package. The only difference with your package is...

Just a quote from a user that supports specifying standard errors in `fit` "I don’t know if GLM.jl supports somehow directly providing the adjusted SEs, in which case this request...

I think get similar timings for FixedEffectModels and `feols` on a mac with ~~only one thread~~ four threads after using `drop_singletons = false` and `tol = 1e-6` (which correspond to...

I have pushed a PR for `feols` to change the default values of `tol` and `drop_singletons` https://github.com/lrberge/fixest/pull/128

If you have access to a GPU, you could try the new experimental feature for GPU. This makes the package an order of magnitude faster. It is an experimental feature,...

I am not sure this is an error. The definition of r2 is ambiguous in the absence of an intercept FixedEffectModels reports the same thing as Stata. See also: https://github.com/FixedEffects/FixedEffectModels.jl/issues/173

Yes. That would be really appreciated! It would be nice to check that the values of fixedeffects, r2, and other statistics match the ones given from Stata `reghdfe`. I've written...

I talk about fe(::Symbol) in the Readme. Isn’t it enough? Just wondering whether people actually use the function with a Term argument.